JP2003123074A - Moving object detector, moving object detection method and moving object detection program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a moving object detector capable of accurately detecting a moving object and continuously recognizing even the moving object which sometimes stops movement as the moving object without immediately recognizing it as a background. SOLUTION: A background difference processing (S20) of comparing an input image and a background image and detecting pixels of different luminance values and an inter-input-image difference processing (S21) of comparing the input image and the previous input image and detecting the pixels of the different luminance values are performed, an object detection area is extracted on the basis of a detected result of the background difference processing and a moving object area is recognized on the basis of the detected result of the inter-input-image difference processing for the extracted object detection area (S22). Then, a counter for counting a count value for each pixel and a storage part for counting a cumulative value of the luminance value for each pixel are provided, the count value is increased and also the cumulative value is increased for the pixels out of the moving object area and data of the background image are updated for the pixels for which the count value is a prescribed value.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving object detection apparatus and method involving background image updating, and more particularly to a moving object detection apparatus based on background subtraction method involving background image updating using digital image processing. It is about the method.

[0002]

2. Description of the Related Art Generally, as a method of detecting a moving object (which may be a moving object or a moving object) using image processing, an interframe difference method, a method of detecting corresponding points between consecutive images, The background subtraction method and the like are known.

Here, the inter-frame difference method is a method of using a difference between frames, and is a method of taking a difference between a plurality of images input at a certain time interval. Although this inter-frame difference method has a simple configuration, the area obtained as a result of the difference does not represent the object itself, so it is difficult to detect the area of the moving object from only the result.

Further, a method of detecting the corresponding points between the continuous images is a method of dividing the images into small areas and calculating the corresponding points by taking the correlation of the small areas between the images. The calculation amount is large, the processing takes a lot of time, and it is necessary to add a special device to shorten the processing time.

Further, the background subtraction method subtracts a background image from an input image, that is, an input image containing an object to be extracted (that is, a moving object) and a background image, thereby extracting an object that does not exist in the background. This is the method of detection. This method can detect the target object satisfactorily if there is a correct background image, but it is necessary to acquire a background image that does not include the target object in advance, and if the background changes, correct results will not be obtained. , It is necessary to update the background one by one.

Here, as a method for detecting a moving object in an image using the above background subtraction method, Japanese Patent Laid-Open No. 62-1140 has been proposed.
There is a method described in No. 64. FIG. 13 shows this JP-A-62.
-114064 is a flow chart for explaining the method described in No. 114064, and first, by the simple addition method,
The background image is updated, and the moving object is detected by calculating the difference between the background image and the input image. Here, in the above simple addition method, when the input image is f (t) and the old background image is g (t−1), the updated background image g (t) is calculated by the following formula.

G (t) = α × f (t) + (1−α) × g (t−1) where α is a constant 0 <α <1, and is common to all pixels. It is a value.

In the method described in the above-mentioned JP-A-62-114064, the simple addition method as described above is used to generate the background. There was a problem that it was difficult to update a good background because it was added.

As a method for solving the above problems,
There is a method described in JP-A No. 11-15982. 14
3 is a flowchart showing an outline of the method described in Japanese Patent Laid-Open No. 11-15982. In this method, first, a region determination / combination process is performed on the input image.
That is, the inter-frame difference processing unit performs inter-frame difference processing to determine the moving area. That is, the input image at time t is compared with the image at time t−1 to determine the moving area. After that, the area division processing unit and the image variation area detection processing unit perform the area division processing and the image variation area detection processing, and determine a predetermined area adjacent to the movement area determined by the inter-frame difference processing as a movement area. Then, the composition processing unit masks the moving area in the input image and updates the background image at t-1 o'clock to generate the background image at t o (502). When the background image is updated, the difference from the input image is calculated (503) to detect the moving object (50
4).

That is, the method described in Japanese Patent Laid-Open No. 11-15982 uses both the interframe difference method and the background difference method. First, the background update area is determined by using the interframe difference method. According to the determination result, the background image and the current image up to that point are combined to generate a new background image.

[0011]

However, the above-mentioned Japanese Unexamined Patent Application Publication No.
According to the method described in No. 1-15982, it is possible to update the background image promptly, but in the case of a moving body such as a person that may stop moving, the background image is displayed at the moment when the moving stops. Furthermore, when the movement is started again, it is detected as an animal body, and there is a problem that the animal body cannot be detected continuously.

Therefore, the present invention can detect a moving object more accurately, and in particular, even an moving object such as a person that may stop moving is not immediately recognized as the background and is continuously recognized. It is an object of the present invention to provide a moving body detection device and method that can be recognized as a moving body.

[0013]

The present invention has been made to solve the above problems, and firstly, the moving object is detected by comparing at least the background image and the input image. A moving object detection device, background image data storage means for storing background image data which is data of a background image, input image data which is data of a certain input image in input images sequentially input, and background image data storage An object area detecting means for detecting an object area which is an area of an object that does not exist in the background image data by comparing the background image data stored in the means, and the input used for the comparison in the object area detecting means. By comparing the image data with the previous input image data, which is the input image data input before the input image data, Moving object detection means for detecting a moving object area which is an area of the moving object from the output object area, and a count value for counting and storing a predetermined count value for each pixel in the input image data For each pixel in the storage unit, for each pixel in the input image data, a cumulative value storage unit for counting the cumulative value of the brightness value, and for the pixels other than the pixels detected as the moving object region by the moving object detection unit, A changing unit that increases the count value in the count value storage unit and increases the cumulative value in the cumulative value storage unit, and a pixel in which the count value is increased by the changing unit and the count value becomes a predetermined value Is
Updating means for updating the background image data stored in the background image data storage means.

In the moving object detecting apparatus of the first structure, the object area detecting means is stored in the input image data which is the data of a certain input image in the input images sequentially input, and the background image data storing means. By comparing the background image data with the background image data, an object region that is a region of the object that does not exist in the background image data is detected. It is assumed that the background image data storage unit stores the brightness value data for each pixel. Further, the moving object region detection means compares the input image data used for comparison in the object region detection means with the previous input image data which is the input image data input before the input image data. As a result, the moving object area, which is the moving object area, is detected from the object areas detected by the object area detecting means. Then, the changing means, for the pixels other than the pixels detected as the moving body region by the moving body detecting means,
The count value in the count value storage means is increased, and the cumulative value in the cumulative value storage means is increased, and the updating means increases the count value by the changing means to bring the count value to a predetermined value. For pixels, the background image data stored in the background image data storage means is updated.

Therefore, the value of the count value storage means is increased for the pixels in the area which is not recognized as the moving object area, and the luminance value is updated for the pixels having the predetermined count value to update the background image. Therefore, even a moving object such as a person that may stop moving can be recognized continuously as a moving object without immediately recognizing the background. In addition, by adjusting the count value when updating the background image data, the background can be updated according to the situation. Furthermore, when updating the background image data,
Since it is determined whether or not it is a moving object area, a moving object does not become a background, and similar to the case of the above simple addition method, an object that continues to stop and an object that has moved are the same. Is never added to the background.

The detection of the object area by the object area detecting means, the detection of the moving object area by the moving object detecting means, the increase of the count value and the cumulative value by the changing means, and the updating by the updating means are input. Every time an image is input,
It will be repeated.

Secondly, in the first configuration, the moving object detecting means compares the input image data with the previous input image data to detect pixels having different brightness values, In the object area detected by the object area detecting means, it is determined whether or not the proportion of pixels having different luminance values is equal to or more than a predetermined value, and if the proportion is equal to or more than a predetermined value, the object The feature is that the region is detected as a moving object region.

Thirdly, an object detecting device for detecting an object by comparing at least a background image and an input image, and an input means for sequentially inputting the input image by photographing. A background image data storage unit that stores background image data that is background image data, input image data at time t in the input image data that is the input image data input by the input unit, and a background image data storage unit. Compare with the stored background image data,
First comparing means for detecting pixels having different brightness values; second comparing means for comparing the input image data at time t with the input image data at time t-1 to detect pixels having different brightness values;
Based on the detection result by the first comparing means, an object area that is an area composed of pixels having different brightness values is detected, and in the object area, pixels having different brightness values detected by the second comparing means are detected. A moving object detection unit that determines whether the occupied ratio is greater than or equal to a predetermined value, and detects the object region as a moving object region when the ratio is greater than or equal to a predetermined value, and the input image data. Count value storage means for counting and storing a predetermined count value for each pixel; cumulative value storage means for counting the cumulative value of the brightness value for each pixel in the input image data; For the pixels other than the pixels detected as the moving object region by the moving object detecting unit, the count value in the count value storage unit is increased and the cumulative value in the cumulative value storage unit is increased. And a updating unit that updates the background image data stored in the background image data storage unit for the pixels whose count value has increased to a predetermined value by the changing unit. It is characterized by

In the moving object detecting apparatus of the third structure, when the input image is input by the input means, the first comparing means inputs the input image which is data of the input image input by the input means. The input image data at time t in the data is compared with the background image data stored in the background image data storage unit to detect pixels having different brightness values.
The moving object detection means detects an object area, which is an area composed of pixels having different brightness values, based on the detection result by the first comparison means, and detects the object area by the second comparison means. It is determined whether or not the ratio occupied by the pixels having different luminance values is equal to or more than a predetermined value, and when the ratio is equal to or more than the predetermined value, the object area is detected as a moving object area. Then, the changing means increases the count value in the count value storage means and increases the cumulative value in the cumulative value storage means for the pixels other than the pixels detected as the moving body region by the moving body detection means. Then, the updating unit updates the background image data stored in the background image data storage unit for the pixel whose count value has been increased by the changing unit and the count value becomes a predetermined value.

Therefore, the value of the count value storage means is increased for the pixels in the area which is not recognized as the moving object area, and the luminance value is updated for the pixels for which the count value reaches the predetermined value to update the background image. Therefore, even a moving object such as a person that may stop moving can be continuously recognized as a moving object without immediately recognizing the background. In addition, by adjusting the count value when updating the background image data, the background can be updated according to the situation. Furthermore, when updating the background image data,
Since it is determined whether or not it is a moving object area, a moving object does not become a background, and similar to the case of the above simple addition method, an object that continues to stop and an object that has moved are the same. Is never added to the background.

It should be noted that the object area detection means detects the object area, the input image comparison means detects pixels having different brightness values, the moving object detection means detects the moving object area, and the changing means counts the count value. The increase of the cumulative value and the update by the update unit are repeatedly performed each time the input image is input.

Fourth, a moving object detecting device for detecting an moving object by comparing at least a background image and an input image, and an input means for sequentially inputting the input image by taking an image, A background image data storage unit that stores background image data that is background image data, input image data at time t in the input image data that is the input image data input by the input unit, and a background image data storage unit. Compare with the stored background image data,
An object area detecting unit that detects an object area that is an area composed of pixels having different brightness values by detecting pixels having different brightness values; input image data at the time t;
Input image comparison means for comparing the input image data at −1 o'clock and detecting pixels having different brightness values, and the brightness value detected by the input image comparison means in the object area detected by the object area detection means A moving object detecting means for determining whether or not the ratio occupied by different pixels is equal to or more than a predetermined value, and detecting the object region as a moving object region when the ratio is equal to or more than a predetermined value, Count value storage means for counting and storing a predetermined count value for each pixel in the input image data, and cumulative value storage for counting the cumulative value of the brightness value for each pixel in the input image data Means and the pixels other than the pixels detected as the moving body region by the moving body detection means, the count value in the count value storage means is increased, and at the same time, in the cumulative value storage means. The changing means for increasing the product value, and the updating means for updating the background image data stored in the background image data storage means for the pixel whose count value is increased by the changing means to reach the predetermined value And are included.

In the moving object detecting apparatus of the fourth configuration, when the input image is input by the input means, the object area detecting means inputs the input image which is data of the input image input by the input means. By comparing the input image data at time t in the data with the background image data stored in the background image data storage means and detecting pixels having different brightness values, an area composed of pixels having different brightness values is detected. Detect an object area. It is assumed that the background image data storage unit stores the brightness value data for each pixel. Further, the input image comparison means is
The input image data at time t-1 is compared with the input image data at time t-1 to detect pixels having different brightness values. Then, the moving object detection means determines whether or not the proportion of pixels having different brightness values detected by the input image comparison means in the object area detected by the object area detection means is equal to or more than a predetermined value. Then, when the ratio is equal to or more than a predetermined value, the object area is detected as a moving object area. Then, the changing means increases the count value in the count value storage means for pixels other than the pixels detected as the moving body region by the moving body detection means,
The updating means increments the cumulative value in the cumulative value storage means, and the updating means stores, in the background image data storage means, pixels whose count value has increased to a predetermined value by the changing means. Update the background image data.

Therefore, the value of the count value storage means is increased for the pixels in the area which is not recognized as the moving object area, and the brightness value is updated for the pixels having the predetermined count value to update the background image. Therefore, even a moving object such as a person that may stop moving can be recognized continuously as a moving object without immediately recognizing the background. In addition, by adjusting the count value when updating the background image data, the background can be updated according to the situation. Furthermore, when updating the background image data,
Since it is determined whether or not it is a moving object area, a moving object does not become a background, and similar to the case of the above simple addition method, an object that continues to stop and an object that has moved are the same. Is never added to the background.

It should be noted that the object area detection means detects the object area, the input image comparison means detects the pixels having different luminance values, the moving object detection means detects the moving object area, and the changing means counts the count value. The increase of the cumulative value and the update by the update unit are repeatedly performed each time the input image is input.

Fifth, in any one of the first to fourth configurations, the changing means stores the count value for the pixel detected by the moving body detecting means as the moving body region. The count value in the means is decreased and the cumulative value in the cumulative value storage means is decreased. As a result, the count value is reduced for the moving object region, so that the moving object region can be made less likely to be updated, and thus the moving object having less movement can be made less likely to become the background. Further, since the cumulative value is reduced, the updated brightness value can be set to an appropriate value.

Sixthly, in the fifth structure, when the changing means decreases the count value in the count value storage means, the count value is decreased by a predetermined number and the cumulative value is decreased. When reducing the cumulative value in the storage means, it is necessary to reduce the cumulative value for the pixel stored in the cumulative value storage means by the value obtained by dividing the cumulative value in the count value storage means by the count value before the reduction. Characterize. Therefore, in particular, when the cumulative value is reduced, the so-called average value of the luminance values of the pixel is reduced, so that the updated luminance value can be set to an appropriate value.

Seventhly, in any one of the first to sixth configurations, when the changing means increases the count value in the count value storing means,
When the count value is increased by a predetermined number and the cumulative value in the cumulative value storage means is increased, it is increased by the luminance value in the input image data.
Therefore, in particular, when the cumulative value is increased, the brightness value is increased by the brightness value, so that the updated brightness value can be set to an appropriate value.

It is preferable that the predetermined values for increasing and decreasing the count value are equal. Basically, when increasing the count value, the count value is increased by 1 and the count value is decreased. When it is made, it is decreased by 1.

Eighthly, in any one of the first to seventh configurations, when the updating means updates the background image data for a predetermined pixel, the accumulated value storage means stores the background image data. The luminance value data is replaced with a value obtained by dividing the cumulative value stored for the pixel by the count value stored for the pixel in the count value storage means. Therefore, when the brightness value is updated for a certain pixel, it is replaced with the average value of the brightness values for the pixel, so that the updated brightness value can be set to an appropriate value.

Ninth, there is provided a moving object detecting method for detecting an moving object by comparing at least a background image with an input image, which is data of a certain input image in sequentially input images. By comparing the image data with the background image data stored in the background image data storage unit that stores the background image data that is the data of the background image, the background image data is present in the background image data. Before the object area detection step of detecting an object area which is an area of an object, input image data used for comparison in the object area detection step, and input image data input before the input image data By comparing with the input image data, the moving object area which is the area of the moving object is detected from the object areas detected in the object area detecting step. The moving object detecting step and the pixels other than the pixels detected as the moving object region in the moving object detecting step are count value storing means, and a predetermined count value is set for each pixel in the input image data. The count value in the count value storage means for counting and storing is increased, and the cumulative value storage means is a cumulative value storage for counting the cumulative value of the brightness value for each pixel in the input image data. For the changing step of increasing the cumulative value in the means, and for the pixel having the count value increased in the changing step to reach the predetermined value, the background image data stored in the background image data storage means is updated. And a renewal step of performing.

In the moving object detecting method of the ninth structure, first, in the object area detecting step, the input image data, which is the data of a certain input image in the input images sequentially input, and the background image data storage means are used. Then, by comparing with the background image data stored in the background image data storage unit that stores the background image data that is the background image data, the object region that is the region of the object that does not exist in the background image data is detected. To do. Then, in the moving object detection step, by comparing the input image data used for comparison in the object area detection step with the previous input image data which is the input image data input before the input image data, A moving object region, which is a moving object region, is detected from the object regions detected in the object region detecting step. After that, in the changing step, for the pixels other than the pixels detected as the moving object region in the moving object detecting step, a count value storage means is provided, and a predetermined count value is set for each pixel in the input image data. The count value in the count value storage means for counting and storing is increased, and the cumulative value storage means is a cumulative value storage for counting the cumulative value of the brightness value for each pixel in the input image data. For the pixels whose cumulative value in the means is increased and the count value is increased in the changing step to reach the predetermined value, the background image data stored in the background image data storage means in the updating step. To update.

Therefore, the value of the count value storage means is increased for the pixels in the area which is not recognized as the moving object area, and the brightness value is updated for the pixels for which the count value reaches the predetermined value to update the background image. Therefore, even a moving object such as a person that may stop moving can be continuously recognized as a moving object without immediately recognizing the background. In addition, by adjusting the count value when updating the background image data, the background can be updated according to the situation. Furthermore, when updating the background image data,
Since it is determined whether or not it is a moving object area, a moving object does not become a background, and similar to the case of the above simple addition method, an object that continues to stop and an object that has moved are the same. Is never added to the background.

The object area detecting step, the moving object detecting step, the changing step and the updating step are repeated every time an input image is input.

Tenth, in the ninth configuration, in the moving object detecting step, the input image data and the previous input image data are compared to detect pixels having different brightness values. In the object area detected in the object area detecting step, it is determined whether or not the proportion of pixels having different brightness values is equal to or more than a predetermined value, and when the proportion is equal to or more than a predetermined value, the The object area is detected as a moving object area.

Eleventh, there is provided a moving object detecting method for detecting a moving object by comparing at least a background image and an input image, and an input step of successively inputting the input images by photographing. T in the input image data which is the data of the input image input in the input step
The input image data at the time is compared with the background image data stored in the background image data storage unit, which is the background image data storage unit that stores the background image data that is the background image data, and the pixels having different brightness values are compared. Comparing the input image data at the time t with the input image data at the time t−1, and detecting a pixel having a different brightness value; and the first comparing step. Based on the detection result in
An object region, which is a region composed of pixels having different brightness values, is detected, and whether or not the proportion of the pixels having different brightness values detected in the second comparison step in the object region is not less than a predetermined value. If it is determined that the ratio is equal to or more than a predetermined value, an object detection step of detecting the object area as an object area, and a pixel other than a pixel detected as an object area in the object detection step. Regarding the pixel, it is a count value storage means, and increases the count value in the count value storage means for counting and storing a predetermined count value for each pixel in the input image data, and at the same time, stores the cumulative value storage means. A changing step of increasing the cumulative value in a cumulative value storage means for counting the cumulative value of the brightness value for each pixel in the input image data; The pixels to which the count value count is increased has reached a predetermined value in the extent, and having a updating step of updating the background image data stored in the background image data storage means.

In the eleventh configuration of the moving object detection method, in the input step, the input images are sequentially input by photographing. In the first comparing step, the input image data at time t in the input image data that is the input image data that is input in the input step, and the background image data storage unit that is the background image data. The background image data stored in the background image data storage unit that stores the image data is compared to detect pixels having different brightness values. Further, in the second comparison step, the input image data at the time t is compared with the input image data at the time t−1, and pixels having different brightness values are detected. Then, in the moving object detecting step, based on the detection result in the first comparing step, an object area that is an area formed of pixels having different brightness values is detected, and in the object area, in the second comparing step. It is determined whether or not the ratio occupied by the detected pixels having different brightness values is equal to or higher than a predetermined value, and if the ratio is equal to or higher than the predetermined value, the object region is detected as a moving object region. Then, in the changing step, with respect to the pixels other than the pixels detected as the moving object region in the moving object detecting step, a count value storage means is provided, and a predetermined count value is set for each pixel in the input image data. The count value in the count value storage means for counting and storing is increased, and the cumulative value storage means is a cumulative value storage for counting the cumulative value of the brightness value for each pixel in the input image data. The pixel whose count value has increased to a predetermined value in the changing step by increasing the cumulative value in the means, the background image data stored in the background image data storage means in the updating step. To update.

Therefore, the value of the count value storage means is increased for the pixels in the area which is not recognized as the moving object area, and the brightness value is updated for the pixels for which the count value reaches the predetermined value to update the background image. Therefore, even a moving object such as a person that may stop moving can be recognized continuously as a moving object without immediately recognizing the background. In addition, by adjusting the count value when updating the background image data, the background can be updated according to the situation. Furthermore, when updating the background image data,
Since it is determined whether or not it is a moving object area, a moving object does not become a background, and similar to the case of the above simple addition method, an object that continues to stop and an object that has moved are the same. Is never added to the background.

The object area detecting step, the input image comparing step, the moving object detecting step, the changing step, and the updating step are performed every time an input image is input.
It will be repeated.

Twelfthly, there is provided a moving object detecting method for detecting a moving object by comparing at least a background image with an input image, and an input step of successively inputting input images by photographing. T in the input image data which is the data of the input image input in the input step
The input image data at the time is compared with the background image data stored in the background image data storage unit, which is the background image data storage unit that stores the background image data that is the background image data, and the pixels having different brightness values are compared. By detecting
An object region detection step of detecting an object region which is a region composed of pixels having different brightness values is compared with the input image data at the time t and the input image data at the time t−1, and the brightness values are different. Whether the ratio of pixels having different brightness values detected in the input image comparison step in the input image comparison step of detecting pixels and the object area detected in the object area detection step is not less than a predetermined value. Is judged,
When the ratio is equal to or more than a predetermined value, a count value is applied to a moving object detection step of detecting the object area as a moving object area and pixels other than the pixels detected as the moving object area in the moving object detection step. Storage means for increasing the count value in the count value storage means for counting and storing a predetermined count value for each pixel in the input image data, and cumulative value storage means,
For each pixel in the input image data, the changing step of increasing the cumulative value in the cumulative value storage means for counting the cumulative value of the luminance value, and the count value is increased by the changing step in the changing step. The pixel having the value has a step of updating the background image data stored in the background image data storage means.

In the moving object detecting method of the twelfth structure, in the input step, the input images are sequentially input by photographing. Then, in the object area detecting step, the input image data at time t in the input image data which is the data of the input image input in the input step, and the background image which is the background image data storage means and is the background image data. By comparing the background image data stored in the background image data storage unit that stores the data and detecting pixels having different brightness values, an object area that is an area composed of pixels having different brightness values is detected. . In the input image comparison step, the input image data at time t is compared with the input image data at time t−1 to detect pixels having different brightness values. Then, in the moving object detection step, in the object area detected in the object area detection step, it is determined whether or not the proportion of pixels with different brightness values detected in the input image comparison step is equal to or more than a predetermined value. If it is determined that the ratio is equal to or more than a predetermined value,
The object area is detected as a moving object area. Then, in the changing step, with respect to the pixels other than the pixels detected as the moving object region in the moving object detecting step, a count value storage means is provided, and a predetermined count value is set for each pixel in the input image data. The count value in the count value storage means for counting and storing is increased, and the cumulative value storage means is a cumulative value storage for counting the cumulative value of the brightness value for each pixel in the input image data. The pixel whose count value has increased to a predetermined value in the changing step by increasing the cumulative value in the means, the background image data stored in the background image data storage means in the updating step. To update.

Therefore, the value of the count value storage means is increased for the pixels in the area which is not recognized as the moving object area, and the luminance value is updated for the pixels for which the count value reaches the predetermined value to update the background image. Therefore, even a moving object such as a person that may stop moving can be recognized continuously as a moving object without immediately recognizing the background. In addition, by adjusting the count value when updating the background image data, the background can be updated according to the situation. Furthermore, when updating the background image data,
Since it is determined whether or not it is a moving object area, a moving object does not become a background, and similar to the case of the above simple addition method, an object that continues to stop and an object that has moved are the same. Is never added to the background.

The object area detecting step, the input image comparing step, the moving object detecting step, the changing step, and the updating step are performed every time an input image is input.
It will be repeated.

Thirteenth, in any one of the ninth to twelfth configurations, in the changing step,
For the pixel detected as the moving object region in the moving object detecting step, the count value in the count value storage means is decreased and the cumulative value in the cumulative value storage means is decreased. Therefore, it is possible to make it difficult for the moving object to become a background. Further, since the cumulative value is reduced, the updated brightness value can be set to an appropriate value.

In the fourteenth aspect, in the thirteenth configuration, when the count value in the count value storage means is decreased in the changing step, the count value is decreased by a predetermined number and the accumulated value is decreased. When reducing the cumulative value in the storage means, it is necessary to reduce the cumulative value for the pixel stored in the cumulative value storage means by the value obtained by dividing the cumulative value in the count value storage means by the count value before the reduction. Characterize. Therefore, when the cumulative value is reduced, the so-called average value of the luminance values of the pixel is reduced, so that the updated luminance value can be set to an appropriate value.

Fifteenthly, in any one of the ninth to fourteenth constructions, in the changing step,
When increasing the count value in the count value storage means, the count value is increased by a predetermined number, and when increasing the cumulative value in the cumulative value storage means, it is increased by the luminance value in the input image data. It is characterized by Therefore, especially when increasing the cumulative value,
Since the brightness value is increased by the brightness value, the updated brightness value can be set to an appropriate value.

It is preferable that the predetermined values for increasing and decreasing the count value are equal. Basically, when increasing the count value, the count value is increased by 1 and the count value is decreased. When it is made, it is decreased by 1.

Sixteenth, in any one of the ninth to fifteenth configurations, when the background image data is updated for a predetermined pixel in the updating step,
The brightness value data is replaced with a value obtained by dividing the cumulative value stored for the pixel in the cumulative value storage means by the count value stored for the pixel in the count value storage means. Therefore, when the brightness value is updated for a certain pixel, it is replaced with the average value of the brightness values for the pixel, so that the updated brightness value can be set to an appropriate value.

The seventeenth aspect is a computer-readable program for detecting a moving object by comparing at least a background image and an input image, which is a certain input image in the input images sequentially input. By comparing the input image data that is the data of the background image data with the background image data stored in the background image data storage unit that stores the background image data that is the background image data. Object area detection processing for detecting an object area which is an area of an object that does not exist in the image data, input image data used for comparison in the object area detection processing, and input input before the input image data By comparing with the previous input image data, which is image data, from among the object regions detected in the object region detection processing, A moving object detection process for detecting a moving object region that is a region of an object, and pixels other than the pixels detected as the moving object region in the moving object detection process are count value storage means and are included in the input image data. For each pixel, the count value in the count value storage means for counting and storing a predetermined count value is increased, and the cumulative value storage means is provided for each pixel in the input image data. The changing process for increasing the cumulative value in the cumulative value storage means for counting the cumulative value and the pixel for which the count value is increased in the changing process to reach the predetermined value, the background image data storage And a moving object detection program for causing a computer to execute an updating process for updating the background image data stored in the means.

In the moving object detecting program of the seventeenth structure, first, by the object area detecting process, the input image data which is the data of a certain input image in the input images sequentially input, and the background image data storage means are used. Then, by comparing with the background image data stored in the background image data storage unit that stores the background image data that is the data of the background image, the object region that is the region of the object that does not exist in the background image data is detected. To do. Then, by the moving object detection process, by comparing the input image data used for comparison in the object region detection step and the previous input image data that is the input image data input before the input image data, The object area, which is the area of the object, is detected from the object areas detected by the object area detection processing. Thereafter, by the above-mentioned change processing, with respect to the pixels other than the pixels detected as the moving object region by the moving object detection processing, the count value storage means is provided, and a predetermined count value is set for each pixel in the input image data. The count value in the count value storage means for counting and storing is increased, and the cumulative value storage means is a cumulative value storage for counting the cumulative value of the brightness value for each pixel in the input image data. For the pixels for which the cumulative value in the means is increased and the count value is increased by the change processing to reach the predetermined value, the background image data stored in the background image data storage means by the updating step. To update.

Therefore, the value of the count value storage means is increased for the pixels in the area which is not recognized as the moving object area, and the brightness value is updated for the pixels for which the count value reaches the predetermined value to update the background image. Therefore, even a moving object such as a person that may stop moving can be recognized continuously as a moving object without immediately recognizing the background. In addition, by adjusting the count value when updating the background image data, the background can be updated according to the situation. Furthermore, when updating the background image data,
Since it is determined whether or not it is a moving object area, a moving object does not become a background, and similar to the case of the above simple addition method, an object that continues to stop and an object that has moved are the same. Is never added to the background.

The object area detecting process, the moving object detecting process, the changing process, and the updating process are repeated each time an input image is input.

Eighteenth, in the seventeenth configuration, in the moving object detection processing, the input image data and the previous input image data are compared to detect pixels having different brightness values. In the object area detected in the object area detection process, it is determined whether or not the proportion of pixels having different brightness values is equal to or more than a predetermined value, and when the proportion is equal to or more than a predetermined value, the The object area is detected as a moving object area.

The nineteenth aspect is a computer-readable program for detecting a moving object by comparing at least a background image with an input image, wherein the input images are sequentially input by photographing. Input processing, input image data that is input image data input in the input processing at time t, and background image data storage means that stores background image data that is background image data. The background image data stored in the background image data storage means to detect the pixels having different luminance values, the input image data at the time t, and the input image data at the time t−1. Compare
A second comparison process for detecting pixels having different brightness values;
Based on the detection result of the comparison process, an object region that is a region composed of pixels having different brightness values is detected, and the proportion of the pixels having different brightness values detected by the second comparison process in the object region is determined. An object detection process for determining whether the object area is an object area when the ratio is equal to or more than a predetermined value by determining whether the object is a predetermined value or more, and an object in the object detection process. For the pixels other than the pixels detected as the region, the count value storage means is provided for each pixel in the input image data.
The count value in the count value storage means for counting and storing a predetermined count value is increased, and the cumulative value storage means counts the cumulative value of the brightness value for each pixel in the input image data. Change processing for increasing the cumulative value in the cumulative value storage means for storing, and the pixels whose count value has increased to a predetermined value in the change processing are stored in the background image data storage means. It is a moving object detection program for causing a computer to execute an update process of updating background image data.

In the moving object detecting program of the nineteenth structure, the input image is sequentially photographed by the above-described input processing. Then, by the first comparison process, the input image data at time t in the input image data, which is the data of the input image input by the input process,
The background image data storage means compares the background image data stored in the background image data storage means that stores the background image data that is the background image data, and detects pixels having different brightness values. Further, by the second comparison processing, the input image data at the time t is compared with the input image data at the time t−1, and pixels having different brightness values are detected. Then, the moving object detection process detects an object region, which is a region composed of pixels having different brightness values, based on the detection result of the first comparison process, and in the object region, the second comparison process. It is determined whether or not the ratio occupied by the detected pixels having different brightness values is equal to or higher than a predetermined value, and if the ratio is equal to or higher than the predetermined value, the object region is detected as a moving object region. Then, by the change processing, for the pixels other than the pixels detected as the moving body region by the moving body detection processing, the count value storage means is provided, and a predetermined count value is set for each pixel in the input image data. In addition to increasing the count value in the count value storage means for counting and storing, the cumulative value storage means is a cumulative value storage for counting the cumulative value of the brightness value for each pixel in the input image data. For the pixels whose count value has been increased by the changing process by increasing the cumulative value in the means, the background image data stored in the background image data storage means by the updating process. To update.

Therefore, the value of the count value storage means is increased for the pixels in the area which is not recognized as the moving object area, and the brightness value is updated for the pixels for which the count value reaches a predetermined value to update the background image. Therefore, even a moving object such as a person that may stop moving can be recognized continuously as a moving object without immediately recognizing the background. In addition, by adjusting the count value when updating the background image data, the background can be updated according to the situation. Furthermore, when updating the background image data,
Since it is determined whether or not it is a moving object area, a moving object does not become a background, and similar to the case of the above simple addition method, an object that continues to stop and an object that has moved are the same. Is never added to the background.

The object area detecting process, the input image comparing process, the moving object detecting process, the changing process, and the updating process are performed every time an input image is input.
It will be repeated.

The twentieth is a computer-readable program for detecting a moving object by comparing at least a background image with an input image, and the input image is sequentially input by taking a picture. Input processing, input image data that is input image data input in the input processing at time t, and background image data storage means that stores background image data that is background image data. By detecting the pixels having different luminance values by comparing the background image data stored in the background image data storage means with each other, the object area detection for detecting the object area which is an area composed of the pixels having different luminance values Input image comparison in which processing and input image data at time t are compared with input image data at time t−1, and pixels having different brightness values are detected In the object area detected in the object area detection processing, it is determined whether or not the ratio of pixels having different brightness values detected in the input image comparison processing is equal to or more than a predetermined value, and the ratio is determined. Is a predetermined value or more, a count value storage means for a moving object detection process for detecting the object region as a moving object region and pixels other than the pixels detected as the moving object region in the moving object detection process. The count value in the count value storage means for counting and storing a predetermined count value for each pixel in the input image data is increased, Change processing for increasing the cumulative value in the cumulative value storage means for counting the cumulative value of the luminance value for each pixel in, and the count value in the changing processing. For a pixel whose count value has become a predetermined value in addition, a moving object detection program for causing a computer to execute an update process of updating the background image data stored in the background image data storage means. .

In the moving object detection program of the twentieth configuration, the input image is sequentially input by performing the photographing by the input processing. Then, by the object area detection process, the input image data at time t in the input image data which is the data of the input image input by the input process, and the background image which is the background image data storage means and is the background image data. By comparing the background image data stored in the background image data storing means for storing data with each other and detecting pixels having different luminance values, an object area which is an area composed of pixels having different luminance values is detected. . Further, by the input image comparison processing, the input image data at the time t is compared with the input image data at the time t−1, and pixels having different brightness values are detected. Then, in the object area detected by the object area detection processing by the moving object detection processing, it is determined whether or not the proportion of pixels having different brightness values detected by the input image comparison processing is greater than or equal to a predetermined value. When it is determined that the ratio is equal to or more than a predetermined value, the object area is detected as the moving object area. Then, by the change processing, for the pixels other than the pixels detected as the moving body region by the moving body detection processing, the count value storage means is provided, and a predetermined count value is set for each pixel in the input image data. In addition to increasing the count value in the count value storage means for counting and storing, the cumulative value storage means is a cumulative value storage for counting the cumulative value of the brightness value for each pixel in the input image data. For the pixels whose count value has been increased by the changing process by increasing the cumulative value in the means, the background image data stored in the background image data storage means by the updating process. To update.

Therefore, the value of the count value storage means is increased for the pixels in the area which is not recognized as the moving object area, and the brightness value is updated for the pixels for which the count value reaches a predetermined value to update the background image. Therefore, even a moving object such as a person that may stop moving can be recognized continuously as a moving object without immediately recognizing the background. In addition, by adjusting the count value when updating the background image data, the background can be updated according to the situation. Furthermore, when updating the background image data,
Since it is determined whether or not it is a moving object area, a moving object does not become a background, and similar to the case of the above simple addition method, an object that continues to stop and an object that has moved are the same. Is never added to the background.

The object area detecting process, the input image comparing process, the moving object detecting process, the changing process, and the updating process are performed every time an input image is input.
It will be repeated.

The 21st structure may be as follows. That is, in any one of the seventeenth to twentieth configurations, for the pixel detected as the moving object region by the moving object detecting process in the changing process, the count value in the count value storage means is decreased and The cumulative value in the cumulative value storage means is decreased.

The 22nd structure may be as follows. That is, in the twenty-first structure, when the count value in the count value storage means is decreased in the change processing, the count value is decreased by a predetermined number, and the cumulative value in the cumulative value storage means is decreased. In this case, the cumulative value for the pixel stored in the cumulative value storage means is reduced by a value obtained by dividing the cumulative value for the pixel by the count value before reduction in the count value storage means.

The 23rd structure may be as follows. That is, in any one of the seventeenth to twenty-second configurations, when the count value in the count value storage means is increased in the change processing, the count value is increased by a predetermined number, and the accumulated value storage is performed. When the cumulative value in the means is increased, it is characterized in that it is increased by the luminance value in the input image data.

It is preferable that the predetermined values for increasing and decreasing the count value are the same, and basically, for increasing the count value, the count value is increased by 1 and the count value is decreased. When it is made, it is decreased by 1.

The 24th structure may be as follows. That is, in any one of the seventeenth to twenty-third configurations, when the background image data is updated for a predetermined pixel in the update processing, the cumulative value stored for the pixel in the cumulative value storage means It is characterized in that the brightness value data is replaced with a value divided by the count value stored for the pixel in the count value storage means.

[0067]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described with reference to the drawings. As shown in FIG. 1, the moving object detection apparatus A according to the present invention includes a camera 10, a display unit 20, an input unit 30, a storage unit 40, and a CPU 90.
And have.

Here, the camera 10 is a photographing device for photographing a moving image, and has a function of photographing an image at predetermined time intervals (for example, several tens of frames per second). Also,
The display unit 20 is a display device for displaying various information. The input unit 30 is an input device for performing various operations on the moving object detection device A. The camera 10 functions as the input means.

The storage unit 40 is a storage device for storing various kinds of information, and has a data storage unit 50 and a program storage unit 70.

Here, the data storage unit 50 is for storing various data other than the program data, and the data storage unit 50 has a background image data storage unit 52 as shown in FIG. And the input image data storage unit 54
A front input image data storage unit 56, a background difference data storage unit 58, an input image difference data storage unit 60, a moving object region data storage unit 62, a counter 64, and a cumulative value data storage unit 66. Have

Here, the background image data storage unit 52 is a storage unit for storing the background image data, and has a storage area for storing the data of the brightness value for each pixel of the background image. This background image data storage unit 5
2 stores luminance value data for each pixel of the background image (see FIG. 8A). The background image data storage unit 52 functions as the background image data storage means.

The input image data storage unit 54 is a storage unit for storing the input image data, and the camera 10
The input image data storage unit 54 has a storage area for storing the data of the brightness value for each pixel for the input image (for example, frame image) captured by For the image, the data of the luminance value for each pixel is stored (see FIG. 8B).

The previous input image data storage unit 56 is a storage unit for storing the previous input image data, and is one of the previous input images, that is, one of the input images stored in the input image data storage unit 54. The previous input image has a storage area for storing the data of the brightness value of each pixel, and the previous input image data storage unit 56 stores the brightness value of each pixel of the previous input image. Data is stored (see FIG. 8C). That is, assuming that the input image data stored in the input image data storage unit 54 is the input image data at time t, the previous input image data storage unit 5
It can be said that the previous input image data stored in 6 is the input image data at t-1.

The background difference data storage unit 58 is a storage unit for storing the background difference data, that is, the data obtained by comparing the background image data with the input image data. The background difference data storage unit 58 has a storage area for storing the background difference data. The background difference data will be described in more detail. Specifically, when the luminance value is different between the background image data and the input image data for each corresponding pixel, a value of 1 is input. On the other hand, when the brightness values are equal, a value of 0 is input (see FIG. 8D). That is, this background difference data is composed of a flag of 1 and a flag of 0, and is stored as, so to speak, binary data. The pixel portion having the value of 1 is a portion having a difference from the background, so to speak, an object is detected. An area in which a plurality of pixels each having a value of 1 are continuously adjacent to each other is referred to as an object detection area.

Further, the input image difference data storage unit 60
Is a storage unit for storing the input image difference data (eg, “interframe difference data”), that is, the data obtained by comparing the input image data with the previous input image data, The input image difference data storage unit 60 has a storage area for storing the input image difference data, and the input image difference data is stored in the input image difference data storage unit 60. It should be noted that the inter-input-image difference data will be described in more detail. Specifically, when the brightness value is different between the corresponding input pixels of the input image data and the previous input image data, the value of 1 is used. Is input, on the other hand, when the brightness values are equal, a value of 0 is input (see FIG. 8E).
That is, the difference data between the input images has a flag of 1 and 0.
The flag is stored as binary data. Since the area of the pixel having the value of 1 is a portion having a difference from the previous input image, so to speak, a moving portion is detected. An area in which a plurality of pixels each having a value of 1 are continuously adjacent to each other is referred to as a motion detection area.

The moving object region data storage unit 62 is a storage unit for storing moving object region data, that is, data obtained by comparing the background difference data and the input image difference data. The moving object region data storage unit 62 has a storage area for storing the moving object region data, and the moving object region data storage unit 62 stores the moving object region data. To describe this moving object region data in more detail, in the pixel group in the object detection region in the background difference data, when the ratio of the input image difference data being 1 is a predetermined ratio or more, the object is detected. This is data obtained by setting pixels in the detection region to a value of 1 and other pixels to a value of 0 (see FIG. 8 (f)). More details will be described later.

The counter 64 is a counter that counts each pixel, and is configured to count an integer between 0 and n (n is an integer). In the body region data, the value of the counter is decremented by 1 for pixels having a value of 1
With respect to the pixels that are, the counter value is incremented by 1. The count value of each pixel in the counter 64 does not become less than 0 and does not become larger than n. The counter 64 functions as the count value storage means.

The cumulative value data storage unit 66 is a storage unit for storing a cumulative value obtained by increasing or decreasing the brightness value for each pixel. Specifically, the cumulative value data storage unit 66 has a value of 1 in the moving object region data. The cumulative value is decreased for the pixel having the value of 0, and the cumulative value is increased for the pixel having the value of 0. Note that when decreasing the cumulative value, the cumulative value before the decrease is decreased by the value obtained by dividing the count value of the corresponding pixel, while when increasing the cumulative value, the corresponding pixel in the input image data at that time is decreased. The brightness value of is increased. Details will be described later. The cumulative value data storage unit 66 functions as the cumulative value storage means.

The count value of each pixel in the counter 64 and the cumulative value of each pixel in the cumulative value data storage unit 66 are all 0 in the initial state.

Next, the program storage unit 70 stores various programs for operating the moving object detection apparatus A. In particular, as shown in FIG. 3, the background difference processing program 72 and the input image difference A processing program 74, a moving object region determination processing program 76, and a counter increase / decrease processing and cumulative value data increase / decrease processing program (hereinafter simply referred to as “increase / decrease program”) 78 are stored.

The background difference processing program 72 is a program for performing background difference processing described later, and the inter-input-image difference processing program 74 is a program for performing inter-input-image difference processing described later. In addition, the moving object region determination processing program 76 is
The increase / decrease program 78 is a program for performing a moving object region determination process, which will be described later, and the counter increase / decrease process and cumulative value data increase / decrease process, which will be described later.

The operation of the moving object detection apparatus A having the above structure will be described with reference to FIGS. First, the input image is stored (S10). That is, the camera 10 inputs the first input image (for example, a frame) and stores the input image data in the input image data storage unit 54. Then, the input image data stored in the input image data storage unit 54 is moved to the previous input image data storage unit 56, and the next input image is input and stored in the input image data storage unit 54. Regarding the background image data, initially, the input image input first is stored in the background image data storage unit 52 as the background image data. In the initial state in which the moving object detection apparatus A is operated, the counter 64 and the cumulative value data storage unit 66 are in the initial state, and all the values for each pixel are 0.
It has become. The step S10 and step S14 described later function as the input step and the input process.

Then, a series of processing is repeated (S11). In this series of processing, as shown in FIG. 5, the processing is performed as follows.

First, background difference processing is performed (S20).
This background difference processing is performed by the background difference processing program. The background difference processing program and the CPU 90 that operates based on the program function as the object area detecting unit and the first comparing unit. The background subtraction process corresponds to the object region detection process, the first comparison process, the object region detection process, and the first comparison process. In this background difference processing, the background image data stored in the background image data storage unit 52 is compared with the input image data stored in the input image data storage unit 54, and the comparison result is the background difference data. Is stored in the background difference data storage unit 58 (see FIG. 8D). Specifically, between the background image data and the input image data,
The brightness values of the respective pixels are compared. If the brightness values are the same, the value is 0, and if the brightness values are different, the value is 1. For example, referring to FIG. 8 as an example, (p, q)
For the pixel at the position of, the luminance value of the background image data is 2
Since the brightness value of the input image data is 4, a value of 1 is given to the background difference data. In addition, (p + 1,
For the pixel at the position q), the brightness value of the background image data is 2, and the brightness value of the input image data is also 2. Therefore, a value of 0 is given to the background difference data. In this background subtraction processing, the object area, which is the area of the object that does not exist in the background image data, is detected by comparing the input image data with the background image data.

Next, difference processing between input images is performed (S2).
1). The input image difference processing is performed by the input image difference processing program. This input image difference processing program and the CPU 9 that operates based on the program
0 functions as the second comparing unit and the input image comparing unit. The input image difference process corresponds to the second comparison process, the input image comparison process, the second comparison process, and the input image comparison process. In this input image difference processing, the input image data stored in the input image data storage unit 54 and the previous input image data storage unit 5 are input.
The previous input image data stored in 6 is compared, and the comparison result is stored in the input image difference data storage unit 60 as the input image difference data. Specifically, the brightness value of each pixel is compared between the input image data (and the previous input image data, and when the brightness values are equal, the value is set to 0, and when the brightness values are different, 1 is set. Where, where
If the input image data is the input image data at time t, it can be said that the previous input image data is the input image data at time t-1. For example, using FIG. 8 as an example,
For the pixel at the position of (p, q), the brightness value of the input image data is 4 and the brightness value of the previous input image data is 4,
A value of 0 is given to the input image difference data. Further, for the pixel at the position of (p + 1, q + 3), the luminance value of the input image data is 6 and the luminance value of the previous input image data is 4, so that a value of 1 is given to the input image difference data.

Next, a moving object region determination process is performed (S2).
2). This moving object region determination processing is performed by the moving object region determination program. The moving object region determination program and the CPU 90 that operates based on the program,
It functions as the moving body detection means. In addition, this moving object region determination processing corresponds to the moving object detection step and moving object detection processing described above. In the moving object region determination process, when the ratio of the input image difference data being 1 in the pixel group in the object detecting region in the background difference data is equal to or more than a predetermined ratio, the pixel in the object detecting region is set to 1 This is a process of setting a value and setting the other pixels to a value of 0.

The moving object region determination processing will be described in more detail. First, the object detection region is extracted from the background difference data (S30). For example, in the background difference data, a region in which a plurality of pixels each having a value of 1 are continuously adjacent to each other is extracted as an object detection region. For example, in the example of FIG. 8, in the background difference data in FIG. 8, as shown in FIG. 9, the hatched regions R1 and R2 are the object detection regions.

Then, the object detection area to be judged is specified (S31). That is, the object detection area for which the determination in step S32 is performed is specified. When a plurality of object detection areas are extracted in step S31, one of them is specified. Note that the object detection area is an area in which a plurality of pixels having a value of 1 are continuously adjacent to each other in the background difference data, but the present invention is not limited to this. When there is only one pixel with a value of 0 in the position between (including vertical, horizontal, and diagonal directions), it may be considered that they are adjacent to each other (in that case, The region R1 and the region R2 are regarded as one object detection region), and in a region composed of a plurality of pixel groups, when the ratio of pixels having a value of 1 is a predetermined ratio or more, the region May be the object detection area, and other methods may be used to detect the area of the object.

Then, a predetermined judgment is made on the object detection area (S32). In this determination, the value of the inter-input-image data is 1 in the pixel group in the object detection area.
The ratio of the pixels that have become is calculated, and it is determined whether or not the ratio is equal to or higher than a predetermined ratio. In FIG.
In the region R1 as the object detection region, the flag in the input inter-image difference data is 1 for two of the six pixels, and therefore 2/6 = 0.33. It is determined whether or not this value is equal to or higher than a predetermined ratio. It should be noted that in the object detection region R2, since the flags in the input image difference data are all 0 for the corresponding pixels, the value of the ratio is 0.

When the ratio is equal to or higher than the predetermined ratio, the object detection area is recognized as the moving object area, and the value of each pixel in the object detecting area is set to 1 in the moving object area data. Yes (S33).

Next, if there is another object detection area, the above steps S31 to S33 are repeated (S34).

As described above, when the processes of steps S31 to S33 are completed for all the object detection areas, the process proceeds to step S35, and the flags of the remaining pixels are set to 0.
In this way, the moving object region data is created. The created moving body region data is stored in the moving body region data storage unit 6
Stored in 2. In this moving object area data, the area where the flag is 1 is the moving object area. For example,
In FIGS. 8 and 9, the region R3 is the moving body region.

In FIG. 8 and FIG. 9, the object detection area R1
Is recognized as a moving object region, the flag is set to 1 (S
33), the object detection area R2 is not recognized as a moving object area and the flag is set to 0 (S35).

In the above description, the object detection area is extracted during the moving object area determination process (S30).
Alternatively, this may be performed during the background difference processing (S20). In this case, the program that operates in this way and the CPU 90 that operates according to the program
It functions as the object area detection means, and the background difference processing in that case corresponds to the object area detection step and the object area detection processing.

Next, the detected moving object area is displayed (S23). That is, the display unit 20 displays the moving object region.

Next, a counter increasing / decreasing process and a cumulative value data increasing / decreasing process (hereinafter referred to as an increasing / decreasing process) are performed (S24).
This increase / decrease processing is performed by the increase / decrease program.
In this increase / decrease process, the value of the counter is increased / decreased, the value of the accumulated value data is increased / decreased, and the background image data is updated for pixels whose counter value has reached a predetermined value.

The increase / decrease process will be described in more detail with reference to FIG. 7. The processes of steps S41 to S52 described below are performed for each pixel. That is, first, as i = 1 (i is a number for specifying the pixel to be processed, and values from 0 to j correspond to each pixel) (S40), the first Let the pixel be the pixel to be processed.

Then, it is determined whether or not the flag in the moving object detection area for the pixel is 1 (S4).
1). If it is 1, the process proceeds to step S42 and 0
If so, the process proceeds to step S47. For example,
In the body region data in FIGS. 8 and 9, (p,
For the pixel at the position of q), the flag is 1, so
On the other hand, the process proceeds to step S42, and since the flag is 0 for the pixel at the position (p + 1, q), step S42 is performed.
It will move to 47.

Then, in step S42, it is determined whether or not the count value is 0 (S42), and if it is not 0, the count value is decremented by 1 (S43). That is, the count value of the pixel in the counter 64 is decreased by 1. For example, the moving body region data is shown in FIG.
The contents of the counter are shown in Fig. 1, and the counter value before the increase / decrease is shown in Fig. 1.
In the case of the state shown in 0 (a), since the value of the moving object region data is 1 for the pixel at the position of (p, q), the count value is as shown in FIG. Decrease from 2 to 1 as shown in b). On the other hand, step S42
If the count value is 0, the value of the cumulative value data for the pixel is also set to 0 (S46), and the step S
And shift to 54.

After step S43, it is determined whether or not the cumulative value is 0 in the cumulative value data (S4
4) If not 0, a process of reducing the accumulated value data is performed (S45). The value to be decreased is a value obtained by dividing the cumulative value of the corresponding pixel in the cumulative value data by the count value of the counter. For example, when the moving object region data has the content shown in FIG. 8 and the cumulative value data before increase / decrease is in the state shown in FIG. 11A, the pixel at the position (p, q) is 30/6 (=
Therefore, the cumulative value data is reduced from 30 to 25, as shown in FIGS. 11 (a) and 11 (b). So to speak, it will reduce the average of the cumulative values. Since the cumulative value of this cumulative value data is not less than 0, it can be reduced to 0.
If it is less than 0, the cumulative value after the increase or decrease is set to 0.

On the other hand, when the flag of the pixel is 0 in step S41, the process proceeds to step S47,
The count value in the pixel is incremented by 1. For example, if the moving object region data has the content shown in FIG. 8 and the value of the counter before increase / decrease is in the state shown in FIG. Indicates that the value of the moving object region data is 0, so the count value is increased from 9 to 10 as shown in FIGS. 10 (a) and 10 (b).

Thereafter, it is judged whether or not the count value for the pixel has reached a predetermined value, that is, the value of n (S48). If the count value has reached the predetermined value, the following step S49 is performed. ~ Perform step S52. That is,
In step S49, a process of increasing the cumulative value data for the pixel is performed. Here, the value to be increased is the brightness value of the pixel in the input image data at that time. For example, when the moving object region data and the input image data have the contents shown in FIG. 8 and the accumulated value data before increase / decrease is in the state shown in FIG. 11A, (p + 1,
For the pixel at the position of q), the brightness value of the corresponding pixel in the input image data (that is, 2) is increased, so the accumulated value data is as shown in FIGS.
As shown in, increase from 27 to 29.

Next, the process proceeds to step S50, and the background image data for the pixel is updated (S50). Here, as an updating method, the cumulative value in the cumulative value data is divided by the number of counters. For example, when the process proceeds to step S50, the cumulative value for the pixel is m
In the case of, the value of m / n (= m ÷ n) is set as the luminance value of the pixel in the background image data. That is,
The brightness value of the pixel in the background image data is replaced.

After that, the count value of the pixel in the counter 64 is set to 0 (S51), and at the same time,
The cumulative value for the pixel in the cumulative value data is set to 0 (S52). That is, the count value and the accumulated value data are cleared for the pixel.

On the other hand, when the value of the counter has not reached the predetermined value in step S48, the process proceeds to step S53 and the process of increasing the accumulated value data is performed. This step S53 is the same process as step S49, and the value to be increased is the brightness value of the pixel in the input image data at that time.

When the processing for a certain pixel is completed as described above, the value of i is incremented by 1 (S54), and the above processing is repeated until the value of i becomes j + 1 (S5).
5). That is, the counter is incremented or decremented for each pixel and the cumulative value data is incremented or decremented, and the background image data is updated for the pixel having a predetermined count value.

When the increase / decrease processing for all the 1st to jth pixels is completed, the increase / decrease processing in step S24 is completed.

In the above, the program for performing the process of increasing / decreasing the count value (S43, S47) and the process of increasing / decreasing the cumulative value (S45, S49, S53) and the CPU 90 operating according to the program are Functioning as a means, these processes are
This corresponds to the above change process. Further, the program for performing the process (S50) for updating the background image data and the CPU 90 operating according to the program function as the updating means, and these processes correspond to the updating process and the updating process.

In the above description, if the count value is 0 in step S42, the value of the accumulated value data for the pixel is also set to 0 (S46), and step S5 is executed.
Although it is supposed to shift to No. 4, it may be performed as follows. That is, after reducing the count value by 1 in step S43, it is determined whether or not the count value after reduction is 0,
When it becomes 0, the value of the cumulative value data is also set to 0, and the process proceeds to step S54. If the count value is not 0, the process directly proceeds to step S44. In this case, if the count value is 0 in step S42, the cumulative value data has already become 0, so the process can proceed directly to step S54. Also in this case, if the count value is 0 in step S42, the process may be moved to step S54 after the process of setting the accumulated value data to 0 is inserted as a precaution, just in case.

Next, in step S25, the updated background image data is displayed (S25). This is to display the background image data on the display unit 20 when the brightness value is updated for at least one of the pixels.

After the series of processes as described above, the previous input image data is stored (S12). That is, the input image data stored in the input image data storage unit 54 is stored in the previous input image data storage unit 56. If the processing is not to be ended (S13), the input image data is stored via the camera 10 (S14), and the next series of processing is performed. In this way, a series of processing is executed each time the next input image is captured.

The above operation will be described below with reference to a specific image example. That is, as shown in FIG. 12, it is assumed that input images are sequentially input from frame 1. At the stage of frame 2, the background image data (see FIG.
The difference between (a2)) and the input image data (FIG. 12 (b2)) is the background difference data shown in FIG. 12 (d2). Further, the input image difference data becomes the input image difference data shown in FIG. 12 (e2), and at this stage, both the person G1 and the chair G2 are recognized as the moving object area (S).
22).

Then, at the stage of frame 3, the person G
If 1 moves the hand, the difference data between the input images is
As shown in FIG. 12 (e3), it becomes a part of the hand movement,
As a result, in the object detection region R10, assuming that a predetermined ratio or more is 1 in the input image difference data, only the object detection region R10 is detected as the moving object region, and the object detection region R11 for the chair is , An example not detected as a moving body region is shown (see S22).

Then, as the processing is repeated thereafter, the count value of the pixel for the chair G2 becomes a predetermined value in the u-1th frame, and the background image data is updated. In Fig. 1, an example in which a chair is captured in the background image data is shown. At the stage of this frame u, the chair is also the background, and therefore the background difference data is only the person as shown in FIG. 12 (du).

As described above, according to the moving object detection apparatus A of the present embodiment, the counter is increased or decreased as described above, and the luminance value is updated for the pixel for which the value of the counter reaches the predetermined value to update the background. Since the image is updated, even a moving object such as a person that may stop moving can be continuously recognized as a moving object without immediately recognizing the background. Further, in the above description, the background image data is updated when the value of the counter reaches n. However, by adjusting the value of this n, the background can be updated according to the situation.
That is, it is preferable to provide setting means for setting the count value (that is, the value of n) when updating the background image data. Specifically, a program for setting the count value is provided so that it can be set from the input unit 30. By doing this, when the value of n is made small, the object that has been stopped for a short time is taken as the background, whereas when the value of n is made large, only the object that has been stopped for a long time becomes the background. Can be.

Further, when updating the background image, since it is determined whether or not it is a moving object region, a moving object does not become the background, and it is different from the case of the above simple addition method. , An object that continues to stop and an object that has moved are not added as a background.

In the above description, the count value is increased / decreased in the increase / decrease process (step S24 in FIG. 5, see FIG. 7), but the background difference data (see FIG. 8) is used.
For a pixel that has a value of 0 in (d),
Since the value also becomes 0 in the moving object region data, the count value may be incremented by 1 for those pixels in the background difference processing. In that case, the increase / decrease of the count value in the increase / decrease process is performed for the remaining pixels.

In the above description, when the background difference data is created, if the brightness values are different, the value is set to 1,
Although it has been described that the value is 0 when the brightness values are the same,
1 and 0 may be reversed. Further, when the difference data between the input images is created, if the brightness values are different, the value is set to 1 and
Although it has been described that the value is 0 when the brightness values are the same,
In this case as well, 1 and 0 may be reversed. Also, in the moving object region data, 1 and 0 may be reversed. In the moving object region data, when 1 and 0 are reversed, the count value is decreased when 0 and the count value is 1 when
It will increase the count value.

In the above description, when the cumulative value data is decreased, the average value of the cumulative values, that is, the value obtained by dividing the cumulative value by the count value is subtracted, but the present invention is not limited to this. The brightness value in the input image data may be subtracted. That is, for example, in the example of FIGS. 8 and 11, since 4 is subtracted for the pixel at the position of (p, q), the cumulative value data changes from 12 to 8.

In the above description, when the cumulative value data is increased, the brightness value of the pixel in the input image data at that time is increased. However, the present invention is not limited to this, and the average value of the cumulative values is You may increase. For example, in the example of FIGS. 8 and 11, the pixel at the position of (p + 1, q) is increased by 27/9 (= 27 ÷ 9) = 3, so that the accumulated value data is changed from 27 to 30. It will change.

A storage unit for storing the data stored in the data storage unit 50 and the program storage unit 7
The storage unit for storing the data stored in 0 may be separate.

[0122]

According to the moving object detecting device, the moving object detecting method, and the moving object detecting program according to the present invention, the value of the count value storage means is increased for the pixels in the area not recognized as the moving object area. , As the background image is updated by updating the brightness value for the pixels whose count value has reached a predetermined value, even a moving object that may stop moving, such as a person, can be immediately recognized as the background. Instead, it can be continuously recognized as a moving body. In addition, by adjusting the count value when updating the background image data, the background can be updated according to the situation. Furthermore, when updating the background image data, since it is determined whether or not it is a moving object region, a moving object does not become the background, and it stops as in the case of the above simple addition method. Objects that have continued to move and objects that have moved are not added as a background.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a moving object detection apparatus according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a configuration of a data storage unit.

FIG. 3 is an explanatory diagram showing a configuration of a program storage unit.

FIG. 4 is a flowchart for explaining the operation of the moving object detection apparatus according to the embodiment of the present invention.

FIG. 5 is a flowchart for explaining the operation of the moving object detection apparatus according to the embodiment of the present invention, in particular, a flowchart for explaining step S11 in FIG. 4 in more detail.

FIG. 6 is a flowchart for explaining the operation of the moving object detection apparatus according to the embodiment of the present invention, and in particular, a flowchart for explaining step S22 in FIG. 5 in more detail.

FIG. 7 is a flowchart for explaining the operation of the moving object detection apparatus according to the embodiment of the present invention, in particular, a flowchart for explaining step S24 in FIG. 5 in more detail.

FIG. 8 is an explanatory diagram for explaining the operation of the moving object detection apparatus according to the embodiment of the present invention.

FIG. 9 is an explanatory diagram for explaining the operation of the moving object detection device according to the embodiment of the present invention.

FIG. 10 is an explanatory diagram for explaining the operation of the moving object detection apparatus according to the embodiment of the present invention.

FIG. 11 is an explanatory diagram for explaining the operation of the moving object detection apparatus according to the embodiment of the present invention.

FIG. 12 is an explanatory diagram for explaining the operation of the moving object detection apparatus according to the embodiment of the present invention.

FIG. 13 is a flowchart for explaining a conventional example.

FIG. 14 is a flowchart for explaining a conventional example.

[Explanation of symbols]

A Moving object detection device 10 cameras 20 Display 30 Input section 40 storage 50 data storage 52 background image data storage unit 54 Input image data storage 56 Previous input image data storage 58 background difference data storage 60 Input image difference data storage unit 62 Animal Body Area Data Storage 64 counter 66 Cumulative value data storage

Claims (20)

[Claims] 1. A moving object detecting device for detecting an moving object by comparing at least a background image with an input image, the background image data storing means for storing background image data which is data of the background image, By comparing the input image data, which is the data of a certain input image in the input image to be input, with the background image data stored in the background image data storage means, an object that is an area of an object that does not exist in the background image data Object region detecting means for detecting a region, input image data used for comparison in the object region detecting means, and previous input image data which is input image data input before the input image data are compared. Thereby, a moving object detecting unit for detecting a moving object region which is a moving object region from the object region detected by the object region detecting unit, A count value storage unit for counting and storing a predetermined count value for each pixel in the input image data, and a cumulative value for counting a cumulative value of luminance values for each pixel in the input image data. With respect to the storage unit and the pixels other than the pixels detected as the moving object region by the moving object detection unit, the changing unit that increases the count value in the count value storage unit and increases the cumulative value in the cumulative value storage unit And update means for updating the background image data stored in the background image data storage means for the pixels whose count value has increased to a predetermined value by the changing means. And a moving object detection device. 2. The moving object detecting means compares the input image data with the previous input image data to detect pixels having different brightness values, and detects the pixels in the object area detected by the object area detecting means. , Determining whether or not the ratio occupied by pixels having different brightness values is equal to or greater than a predetermined value, and detecting the object region as a moving object region when the ratio is equal to or greater than a predetermined value. The moving object detection apparatus according to claim 1. 3. A moving object detecting device for detecting an moving object by comparing at least a background image and an input image, the input means inputting the input images sequentially by photographing, and the background image data. Background image data storage means for storing certain background image data, input image data at time t in the input image data which is data of the input image input by the input means, and background image stored in the background image data storage means First comparing means for comparing data with each other to detect pixels having different brightness values, and comparing the input image data at time t with the input image data at time t-1 to detect pixels with different brightness values. Based on the detection results by the second comparing means and the first comparing means, an object area which is an area composed of pixels having different brightness values is detected, and in the object area, It is determined whether or not the ratio occupied by the pixels having different brightness values detected by the second comparing means is equal to or more than a predetermined value, and when the ratio is equal to or more than the predetermined value, the object area is defined as the moving object area. A moving object detecting means for detecting as a value, a count value storing means for counting and storing a predetermined count value for each pixel in the input image data, and a brightness value for each pixel in the input image data. For a cumulative value storage unit for counting the cumulative value, and for pixels other than the pixels detected as the moving object region by the moving object detection unit, the count value in the count value storage unit is increased and the cumulative value storage is performed. The changing means for increasing the cumulative value in the means, and the pixel for which the count value is increased by the changing means to reach the predetermined value, are stored in the background image data storage means. Animals detection apparatus, characterized in that it comprises updating means for updating 憶 background image data. 4. A moving object detecting apparatus for detecting an moving object by comparing at least a background image with an input image, the input means inputting the input images one after another by photographing, and the data of the background image. Background image data storage means for storing certain background image data, input image data at time t in the input image data which is data of the input image input by the input means, and background image stored in the background image data storage means Object region detecting means for detecting an object region which is a region composed of pixels having different luminance values by comparing the data with pixels having different luminance values; and input image data at the time t, Input image comparison means for comparing the input image data at t-1 o'clock and detecting pixels having different brightness values, and an object area detected by the object area detection means Then, it is determined whether or not the ratio occupied by the pixels having different brightness values detected by the input image comparison means is equal to or more than a predetermined value, and if the ratio is equal to or more than the predetermined value, the object area is Moving body detection means for detecting as a moving body region, count value storage means for counting and storing a predetermined count value for each pixel in the input image data, and for each pixel in the input image data, With respect to the cumulative value storage means for counting the cumulative value of the brightness value and the pixels other than the pixels detected as the moving body region by the moving body detection means, the count value in the count value storage means is increased and The changing means for increasing the cumulative value in the cumulative value storage means, and the pixel for which the count value is increased by the changing means and the count value becomes the predetermined value, the background image data Animals detection apparatus, characterized in that it comprises updating means for updating the background image data stored in the 憶 means. 5. The changing means reduces the count value in the count value storage means and decreases the cumulative value in the cumulative value storage means for the pixels detected as the moving body region by the moving body detection means. The moving object detection device according to claim 1, 2 or 3 or 4. 6. The changing means reduces the count value by a predetermined number when decreasing the count value in the count value storage means, and when decreasing the cumulative value in the cumulative value storage means, 6. The moving object detection according to claim 5, wherein the cumulative value for the pixel stored in the cumulative value storage means is reduced by a value obtained by dividing the cumulative value in the count value storage means by the count value before the reduction. apparatus. 7. The changing means increases the count value by a predetermined number when increasing the count value in the count value storage means, and when increasing the cumulative value in the cumulative value storage means, 7. The moving object detection apparatus according to claim 1, wherein the moving object detection device increases the brightness value of the input image data. 8. When the updating means updates the background image data for a predetermined pixel, the cumulative value stored for the pixel in the cumulative value storage means is stored for the pixel in the count value storage means. 8. The moving object detection device according to claim 1, wherein the brightness value data is replaced with a value divided by the counted value. 9. A moving object detecting method for detecting an moving object by comparing at least a background image and an input image, wherein input image data which is data of a certain input image in sequentially input images, and a background. By comparing with the background image data stored in the background image data storage unit, which is the image data storage unit and stores the background image data that is the data of the background image, in the area of the object that does not exist in the background image data. An object area detecting step of detecting a certain object area, input image data used for comparison in the object area detecting step, and previous input image data which is input image data input before the input image data. By comparing, the moving object detection for detecting the moving object area which is the moving object area from the object areas detected in the object area detecting step Degree and, for the pixels other than the detected pixel with the moving object region in said animal object detection process, a count value memory unit,
For each pixel in the input image data, the count value in the count value storage means for counting and storing a predetermined count value is increased, and the accumulated value storage means is provided for each pixel in the input image data. In the changing step for increasing the cumulative value in the cumulative value storage means for counting the cumulative value of the luminance value, and for the pixel in which the count value is increased in the changing step and the count value becomes a predetermined value, An updating step of updating the background image data stored in the background image data storage means, the moving object detection method. 10. The moving object detecting step compares the input image data with the previous input image data to detect pixels having different brightness values, and detects the object area detected in the object area detecting step. In, it is determined whether the ratio of pixels having different brightness values is equal to or more than a predetermined value, and if the ratio is equal to or more than a predetermined value, the object region is detected as a moving object region. The method for detecting an animal body according to claim 9. 11. A moving object detecting method for detecting an moving object by comparing at least a background image with an input image, the input step of sequentially inputting the input image by photographing, and inputting in the input step. The input image data at time t in the input image data which is the data of the input image, and the background image data storage means, which is the background image data storage means for storing the background image data which is the data of the background image. The first comparison step of comparing the background image data with each other to detect pixels having different luminance values, and comparing the input image data at the time t with the input image data at the time t−1, to determine pixels having different luminance values. Based on the second comparison step for detecting and the detection result in the first comparison step, an object area, which is an area composed of pixels having different brightness values, is detected, and the object area is detected. In step 2, it is determined whether or not the ratio occupied by pixels having different brightness values detected in the second comparison step is equal to or more than a predetermined value, and if the ratio is equal to or more than the predetermined value, the object area is A moving object detecting step of detecting as a moving object area, and a pixel other than a pixel detected as an moving object area in the moving object detecting step is a count value storage means,
For each pixel in the input image data, the count value in the count value storage means for counting and storing a predetermined count value is increased, and the accumulated value storage means is provided for each pixel in the input image data. In the changing step of increasing the cumulative value in the cumulative value storage means for counting the cumulative value of the luminance value, and the pixel whose count value is increased in the changing step and the count value becomes a predetermined value, An updating step of updating the background image data stored in the background image data storage means, the moving object detection method. 12. A moving object detection method for detecting an moving object by comparing at least a background image and an input image, the input step of sequentially inputting the input image by performing photographing, and the input step in the input step. The input image data at time t in the input image data which is the data of the input image, and the background image data storage means, which is the background image data storage means for storing the background image data which is the data of the background image. An object region detection step of detecting an object region which is a region composed of pixels having different brightness values by comparing the image data with background image data and detecting pixels having different brightness values; and input image data at time t And the input image data at t-1 o'clock, and an input image comparison step of detecting pixels having different brightness values, and an object area detection step. In the body region, it is determined whether or not the ratio occupied by the pixels having different brightness values detected in the input image comparison step is equal to or more than a predetermined value, and when the ratio is equal to or more than the predetermined value, the object is A moving object detecting step of detecting the area as a moving object area, and a pixel other than the pixels detected as the moving object area in the moving object detecting step is a count value storage means,
For each pixel in the input image data, the count value in the count value storage means for counting and storing a predetermined count value is increased, and the accumulated value storage means is provided for each pixel in the input image data. In the changing step of increasing the cumulative value in the cumulative value storage means for counting the cumulative value of the luminance value, and the pixel whose count value is increased in the changing step and the count value becomes a predetermined value, An updating step of updating the background image data stored in the background image data storage means, the moving object detection method. 13. The pixel detected as a moving body region in the moving body detecting step in the changing step,
13. The count value in the count value storage means is decreased and the cumulative value in the cumulative value storage means is decreased.
The method for detecting an animal body according to. 14. In the changing step, when the count value in the count value storage means is decreased, the count value is decreased by a predetermined number, and when the cumulative value in the cumulative value storage means is decreased, 14. The moving object detection according to claim 13, wherein the cumulative value for the pixel stored in the cumulative value storage means is decreased by a value obtained by dividing the cumulative value in the count value storage means by the count value before the decrease. Method. 15. In the changing step, when the count value in the count value storage means is increased, the count value is increased by a predetermined number, and when the cumulative value in the cumulative value storage means is increased, 15. The moving object detection method according to claim 9, 10 or 11 or 12 or 13 or 14, wherein the brightness is increased by the brightness value in the input image data. 16. When updating the background image data for a predetermined pixel in the updating step, the cumulative value stored for the pixel in the cumulative value storage means is stored for the pixel in the count value storage means. 16. The method for detecting a moving object according to claim 9, wherein the brightness value data is replaced with a value divided by the count value. 17. A computer-readable program for detecting a moving object by comparing at least a background image with an input image, which is data of a certain input image in input images sequentially input. By comparing the image data with the background image data stored in the background image data storage unit that stores the background image data that is the background image data, the background image data is present in the background image data. Object area detection processing for detecting an object area that is an area of a non-object, input image data used for comparison in the object area detection processing, and input image data input prior to the input image data. By comparing with the input image data, the object area is detected from the object area detected in the object area detection processing. That the moving object detection process of detecting a moving object region, for the pixels other than the detected pixel with the moving object region in said animal object detection process, a count value memory unit,
For each pixel in the input image data, the count value in the count value storage means for counting and storing a predetermined count value is increased, and the accumulated value storage means is provided for each pixel in the input image data. In the change processing for increasing the cumulative value in the cumulative value storage means for counting the cumulative value of the brightness value, and for the pixel whose count value has increased to a predetermined value in the change processing, A moving object detection program for causing a computer to execute an update process for updating the background image data stored in the background image data storage means. 18. In the moving object detection process, the input image data is compared with the previous input image data to detect pixels having different brightness values, and the object region detected in the object region detection process is detected. In, it is determined whether the ratio of pixels having different brightness values is equal to or more than a predetermined value, and if the ratio is equal to or more than a predetermined value, the object region is detected as a moving object region. The moving object detection program according to claim 17. 19. A computer-readable program for detecting a moving object by comparing at least a background image and an input image, the input processing inputting successive input images by photographing. Background image data storage that stores the input image data at time t in the input image data that is the input image data input in the input processing, and the background image data storage unit that stores the background image data that is the background image data. The first comparison process of comparing the background image data stored in the means to detect pixels having different brightness values, the input image data at the time t and the input image data at the time t−1 are compared to obtain the brightness. A second comparison process for detecting pixels having different values, and an object region which is a region composed of pixels having different brightness values based on the detection result of the first comparison process. Is detected, it is determined whether or not the ratio of the pixels having different brightness values detected by the second comparison processing in the object region is equal to or more than a predetermined value. In some cases, a moving object detection process for detecting the object region as a moving object region, and a pixel other than the pixel detected as the moving object region in the moving object detection process is count value storage means,
For each pixel in the input image data, the count value in the count value storage means for counting and storing a predetermined count value is increased, and the accumulated value storage means is provided for each pixel in the input image data. In the change processing for increasing the cumulative value in the cumulative value storage means for counting the cumulative value of the luminance value, and for the pixel whose count value has increased to a predetermined value in the change processing, A moving object detection program for causing a computer to execute an update process for updating the background image data stored in the background image data storage means. 20. A computer-readable program for detecting a moving object by comparing at least a background image and an input image, the input processing inputting successive input images by photographing. Background image data storage that stores the input image data at time t in the input image data that is the input image data input in the input processing, and the background image data storage unit that stores the background image data that is the background image data. An object area detection process for detecting an object area that is an area composed of pixels having different brightness values by comparing the background image data stored in the means and detecting pixels having different brightness values; Image comparison processing for comparing the input image data at time t-1 and the input image data at time t-1 and detecting pixels having different brightness values, In the object area detected in the area detection processing, it is determined whether or not the ratio occupied by pixels having different brightness values detected in the input image comparison processing is a predetermined value or more, and the ratio is a predetermined value or more. In the case where the object area is a moving object area, the moving object detecting process detects the object area as a moving object area, and the pixels other than the pixels detected as the moving object area in the moving object detection process are count value storage means.
For each pixel in the input image data, the count value in the count value storage means for counting and storing a predetermined count value is increased, and the accumulated value storage means is provided for each pixel in the input image data. In the change processing for increasing the cumulative value in the cumulative value storage means for counting the cumulative value of the brightness value, and for the pixel whose count value has increased to a predetermined value in the change processing, A moving object detection program for causing a computer to execute an update process for updating the background image data stored in the background image data storage means.