JP2014197290A - Smoke detection device and smoke detection method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a smoke detection device capable of distinguishing the existence of smoke with high accuracy by excluding causes of erroneous determination.SOLUTION: The smoke detection device includes: an image memory (10) in which a plurality of images captured in time series by a monitoring camera are stored as time-series images; a preprocessing unit (20) which computes difference values between brightness values of each pixel in the latest image and the previous image on the basis of the time-series images and specifies a target pixel of which the sign of the difference value is equal to that of difference values of peripheral pixels, as a pixel with synchronism and generates a binary image where pixels with synchronism having positive difference values are white, as a positive synchronism image and generates a binary image where pixels with synchronism having negative difference values are white, as a negative synchronism image; and a smoke occurrence detection unit (30) which specifies an area where smoke has occurred, on the basis of a distribution state of brightness variation areas with synchronism and the magnitude of an index value computed about the temporal change in the distribution state in accordance with the positive synchronism image and the negative synchronism image.

Description

  The present invention relates to a smoke detection device that detects the generation of smoke by performing image processing on an image captured by a surveillance camera, and in particular, by a light source such as a change in sunlight or when a mercury lamp or halogen light is turned on. The present invention relates to a smoke detection device and a smoke detection method suitable for detecting smoke in a situation where a relatively slow change in brightness exists.

  From the viewpoint of initial fire extinguishment in the event of a fire or prevention of escape delay in a fire accident, early detection of fire or smoke is very important. Thus, in the field of smoke detection devices, research has been conducted on early detection of smoke by performing image processing on an image captured by a surveillance camera.

  As an example, there is a conventional smoke detection device that detects smoke by installing a camera in a tunnel or the like and performing image processing on an image captured by the camera. In image processing for detecting smoke, generally, a reference image (reference image) is stored in advance, a difference image between the latest captured image and the reference image is calculated, and a region where a change has occurred is calculated. Extraction detects smoke (for example, see Patent Document 1).

  In addition, the reference image is regularly updated in order to cope with the temporal change of the reference image due to the influence of sunlight.

Thus, by performing image processing on the image captured by the camera and performing smoke detection, the following two merits can be obtained.
1) By visually confirming the image of the surveillance camera, it is possible to grasp the smoke detection status in a remote place.
2) It is possible to divert already installed surveillance cameras and construct efficient equipment.

Japanese Patent No. 3909665

However, the prior art has the following problems.
In the prior art, in order to detect smoke, a pixel region where a luminance difference from a frame difference image or a background image exceeds a predetermined threshold is extracted. However, for example, when there is a change in sunshine or a slow change in brightness due to a light source such as a mercury lamp or a halogen light, there is a problem that a portion where no smoke is generated is erroneously detected as smoke. .

  The present invention has been made to solve the above-described problems, and excludes factors that may be erroneously determined as smoke typified by changes in lighting, etc., and identifies the presence of smoke with high accuracy. An object of the present invention is to obtain a smoke detection device and a smoke detection method capable of performing the above.

  A smoke detection device according to the present invention is a smoke detection device that detects the generation of smoke by performing image processing on an image captured by a monitoring camera, and a plurality of images captured in time series by the monitoring camera Based on the image memory that stores the images as time-series images and the time-series images stored in the image memory, the difference value of the luminance value of each pixel in the latest image and the previous image in time series is calculated. Then, a pixel whose difference value of its own pixel corresponding to the target pixel is the same as that of a difference value of one or more pixels at a predetermined surrounding position is identified as a synchronous pixel, and the difference A binary image in which a value is positive and a synchronous pixel is white is generated in time series as a positive synchronization image, and a difference value is negative and a synchronous pixel is white Before generating images in time series as negative sync images Based on the positive synchronization image and the negative synchronization image generated in time series by the processing unit and the preprocessing unit, a quantitative index value regarding the distribution state of the synchronized luminance change region and the temporal change of the distribution state is obtained. And a smoke generation detection unit that identifies a region where smoke is generated based on the calculated index value.

  The smoke detection method according to the present invention is a smoke detection method for detecting the generation of smoke by performing image processing on an image captured by a monitoring camera, and is captured in time series by the monitoring camera. A storage step for storing a plurality of images in the image memory as a time-series image, and a luminance value of each pixel in the latest image and the previous image in time series based on the time-series image stored in the image memory The pixel having the same sign as that of the difference value of one or more pixels in the surrounding predetermined position is synchronized with the pixel. And a binary image in which the difference value is positive and the synchronous pixel is white is generated as a positive synchronization image in time series, and the differential value is negative and the synchronous pixel is A binary image that is white Time-series preprocessing steps, and the synchronized brightness change region distribution state and distribution state change over time based on the time-sequentially generated positive synchronization image and negative synchronization image A smoke generation detecting step of calculating a quantitative index value related to and identifying a region where smoke is generated based on the calculated index value.

  According to the present invention, attention is paid to the presence or absence of synchronism in which the temporal change of the luminance value in the time-series image data increases or decreases, and on the basis of this synchronism, the smoke determination area is excluded By introducing a plurality of index values for identifying the smoke, the smoke detection device can eliminate the factor that may be erroneously determined as smoke represented by changes in lighting, etc., and can accurately identify the presence of smoke And a smoke detection method can be obtained.

It is a block diagram of the smoke detection apparatus in Embodiment 1 of this invention. It is explanatory drawing regarding the series of processes by the synchronism determination process part in Embodiment 1 of this invention. It is explanatory drawing regarding the synchronous image produced | generated by the synchronous image production | generation part in Embodiment 1 of this invention.

  Hereinafter, preferred embodiments of a smoke detection device and a smoke detection method of the present invention will be described with reference to the drawings.

Embodiment 1 FIG.
FIG. 1 is a configuration diagram of a smoke detection device according to Embodiment 1 of the present invention. The smoke detection apparatus according to the first embodiment includes an image memory 10, a preprocessing unit 20, and a smoke generation detection unit 30. The image memory 10 is configured as an image memory for a plurality of frames so that images captured by the camera 1 can be stored as time-series data for a certain period in the past.

  The preprocessing unit 20 includes a difference image generation unit 21, a synchronization determination processing unit 22, and a synchronization image generation unit 23. Then, the pre-processing unit 20 sequentially generates, in time series, a synchronization image that is used when smoke detection is performed based on an image for a predetermined period of time that is captured by the camera 1 and stored in the image memory 10. It has a function to do.

  The smoke generation detection unit 30 includes a determination value calculation unit 31 and a smoke determination unit 32. Then, the smoke generation detection unit 30 calculates a determination value for detecting the generation of smoke based on the synchronous images sequentially generated in time series by the preprocessing unit 20, and smoke is generated based on the calculation result. It has a function of judging whether or not it has occurred.

Here, the detection principle of the present invention will be described.
Changes in sunshine and slow changes in brightness caused by light sources such as mercury lamps and halogen lights are caused by changes in brightness that are synchronized within a certain range in the camera image (brightness with directionality from dark to bright or from bright to dark). Change).

  On the other hand, when smoke is generated, similarly, a synchronous luminance change occurs. However, when smoke is generated, a behavior different from that in the case of a change in brightness due to sunshine, a light source, or the like is observed in the distribution of the synchronized luminance change region and the temporal change in the distribution state. Therefore, in the present invention, paying attention to such a difference, there is a risk of misjudging it as smoke typified by a change in illumination, etc., based on an index value calculated by quantitatively analyzing the distribution of synchronization. The factor is excluded, and the presence of smoke can be identified with high accuracy.

  Next, a series of image processing contents executed by the preprocessing unit 20 in the smoke detection device according to the first embodiment will be described with reference to FIGS. FIG. 2 is an explanatory diagram relating to a series of processes performed by the synchronization determination processing unit 22 according to Embodiment 1 of the present invention. In FIG. 2, the results of the series of processes (a) to (d) by the synchronization determination processing unit 22 are collectively shown by using patterns 1 to 3 as an example. Although details will be described later, it is assumed that the pattern 1 is “no synchronization”, the pattern 2 is “positive synchronization”, and the pattern 3 is “negative synchronization”. The three types of synchronism determined by the above are illustrated.

  Moreover, FIG. 3 is explanatory drawing regarding the synchronous image produced | generated by the synchronous image generation part 23 in Embodiment 1 of this invention. FIG. 3 collectively shows a positive synchronization image and a negative synchronization image generated by the synchronization image generation unit 23 with respect to the captured image. Therefore, the steps will be described in order with reference to FIGS.

[Step 1] Generation of a difference image (see FIG. 2A)
The difference image generation unit 21 calculates a difference image between the latest image and the previous image in time series based on the time series images for a predetermined period of time captured by the camera 1 and stored in the image memory 10. Generate sequentially. Here, each pixel is represented as a luminance value for 8 bits from 0 to 255, for example.

The generation of the difference image will be described in detail with reference to FIG.
FIG. 2A shows difference values for a total of five pixels, that is, a central pixel as a target pixel and four neighboring pixels adjacent thereto (specifically, four pixels adjacent vertically and horizontally). Are shown for pattern 1 to pattern 3. For example, in the upper pattern 1, the difference value of the center pixel is −5, and the brightness value of the center pixel in the latest image is the brightness value of the center pixel in the previous image in time series. This means that it is reduced by “5”.

  Similarly, in the pattern 1, the difference values of the four neighboring pixels are the difference value “−2” of the upper pixel, the difference value “10” of the right pixel, the difference value “3” of the lower pixel, and the difference of the left pixel, respectively. As an example, the value “7” is obtained.

  Note that if any of the difference values regarding the total of five pixels including the central pixel as the target pixel and the four neighboring pixels adjacent thereto is 0, the target pixel is excluded from the synchronization determination target. That is, the synchronism of the target pixel is determined only when the difference value is generated for all five pixels including the target pixel.

[Step 2] Calculation of individual multiplication values of four neighboring pixels (see FIG. 2B)
Next, the synchronization determination processing unit 22 calculates, for each of the pixels constituting the image, the multiplication value of the difference values between the four neighboring pixels and the central pixel for the difference image obtained in the previous step 1. To do.

The calculation of the multiplication value will be described in detail with reference to FIG.
The result of multiplying the difference value of the central pixel by the difference value of each of the four neighboring pixels based on the difference value shown in FIG. 2A is shown in FIG. That is, if the sign of the difference value of each of the four neighboring pixels matches the sign of the difference value of the center pixel, the sign of the multiplication value is +, and if the sign does not match, multiplication is performed. The sign of the value is-.

[Step 3] Individual synchronicity determination of 4 neighboring pixels (see FIG. 2C)
Next, if the sign of the multiplication value calculated in step 2 is positive for each of the four neighboring pixels, the synchronization determination processing unit 22 determines that there is synchronization with the center pixel of interest, and step When the sign of the multiplication value calculated in 2 is negative, it is determined that there is no synchronization with the center pixel of interest. The individual synchronicity determination results of the four neighboring pixels are shown in FIG.

As described above, when the luminance value changes in the direction of the same sign as the central pixel for each of the four neighboring pixels, the synchronization determination processing unit 22 changes the luminance in synchronization with the central pixel. (That is, there is synchronization). In addition, for each of the four neighboring pixels, the synchronism determination processing unit 22 changes the luminance without synchronizing with the central pixel when the luminance value changes in the direction of the sign different from that of the central pixel. (That is, there is no synchronization).
[Step 4] Judgment of synchronism of the center pixel of interest (see FIG. 2D)
Next, the synchronization determination processing unit 22 determines the synchronization of the pixel of interest at the center based on the presence or absence of individual synchronization of the four neighboring pixels. Specifically, the synchronism determination processing unit 22 performs the following determination for the target pixel.
If it is determined that at least one of the four neighboring pixels is not synchronized, it is determined that there is no synchronization.
If all of the four neighboring pixels are determined to be synchronous and the difference value of the pixel is positive, it is determined that there is positive synchronization.
If all of the four neighboring pixels are determined to be synchronous and the difference value of the pixel is negative, it is determined that there is negative synchronization.
FIG. 2 (d) shows the result of determining the synchronism of the pixel of interest at the center.

[Step 5] Synchronous image generation (see FIG. 3)
Next, the synchronization image generation unit 23 generates a synchronization image based on the synchronization determination result regarding each of the target pixels determined in step 4. Specifically, the synchronism image generation unit 23 generates an image in which the target pixel determined to be synchronous is white, and the target pixel determined to be non-synchronous is black. Are generated as a positive synchronization image and a negative synchronization image.

  FIG. 3 shows a specific example of a positive synchronization image and a negative synchronization image generated by the synchronization image generation unit 23. FIG. 3A shows a time-series image, and shows a state in which a circular area having bright luminance is gradually moving from left to right as time passes. By executing the above-described series of steps 1 to 5 on the time series image by the preprocessing unit 20, the normal synchronization image shown in FIG. ) Are respectively generated.

  That is, the positive synchronization image shown in FIG. 3B corresponds to an image obtained by binarizing pixels that change in synchronization with the direction in which the brightness is increased as white. Further, the negative synchronization image shown in FIG. 3C corresponds to an image obtained by binarizing pixels that change in synchronization with the direction in which the luminance is darkened as white.

[Step 6] Determination Value Calculation Processing Based on Positive Synchronous Image and Negative Synchronous Image Next, the determination value calculation unit 31 in the smoke generation detection unit 30 performs the positive synchronization image and negative synchronization generated in step 5. Based on the image, a determination value for determining whether or not smoke is generated is calculated. Specifically, the determination value calculation unit 31 in the first embodiment first divides each of the positive synchronization image and the negative synchronization image into a plurality of predetermined areas, and the latest fixed period. Regarding the time-series data of both images at (predetermined number of times), five index values based on the following five determination values are calculated for each divided area.

[First determination value: an index for extracting an area where smoke is likely to be generated]
The determination value calculation unit 31 determines the number of white pixel continuations for a time-series image including only a positive synchronization image, a time-series image including only a negative synchronization image, and a time-series image including both a positive synchronization image and a negative synchronization image. The number of pixels that is equal to or greater than a first predetermined number of times determined in advance is counted, and the density in the area is calculated as a first determination value.

  This first determination value is a first index value for extracting a smoke determination region. That is, since smoke due to fire is generated continuously, it is considered that there is a continuous region with synchronization. Therefore, in at least one of the positive synchronization image and the negative synchronization image, when the first determination value is equal to or greater than the first threshold value determined in advance, the region has a continuous pixel in the synchronization. Can be extracted as a region where smoke may be generated.

[Second determination value: index value for removing the influence of noise]
The determination value calculation unit 31 counts the number of pixels in which the number of continuations of white pixels is equal to or less than a predetermined second predetermined number of times for both the positive-synchronous image and the negative-synchronous image. By dividing by the total number of pixels, the in-area density is calculated and used as the second determination value.

  This second determination value is a second index value for preventing erroneous determination due to the influence of noise. That is, when the in-area density, which is the second determination value calculated for each area, is equal to or higher than a predetermined second threshold value, the area is an area with very few synchronous pixels. It can be judged that there is. Accordingly, in both the positive synchronization image and the negative synchronization image, by excluding the area where the in-area density that is the second determination value is equal to or more than the second threshold from the smoke determination target area, The influence of noise, which is one of the above, can be removed.

[Third judgment value: Indicator for distinguishing smoke from sunshine change, lighting change, etc.]
The determination value calculation unit 31 counts the number of pixels in which the number of continuations of white pixels is equal to or greater than a predetermined third predetermined number of times for both the time-series images of the positive synchronization image and the negative synchronization image. The density is calculated and set as the third determination value.

  This third determination value is a third index value for preventing erroneous determination due to the influence of changes in sunlight or changes in brightness due to light source movement. That is, when a change in brightness occurs due to a change in sunlight or a light source movement, it is considered that a unidirectional change from “dark to light” or “bright to dark” occurs. For this reason, when the appearance state of white pixels within a predetermined period in the synchronous image is viewed, white pixels appear only in either the positive synchronous image or the negative synchronous image.

  On the other hand, when smoke is generated, since there is lightness and darkness caused by the flow of smoke, white pixels appear in both the positive synchronization image and the negative synchronization image. Therefore, when the in-area density, which is the third determination value calculated for each area, is equal to or greater than a predetermined third threshold value, the area is changed in sunshine or brightness change due to light source movement. It can be determined that this is the region where Therefore, in either one of the positive synchronization image and the negative synchronization image, an area where the density within the area, which is the third determination value, is equal to or more than the third threshold is excluded from the smoke determination target area. It is possible to remove the influence of brightness change, which is one of the determination factors. Note that the third predetermined number of times used to use the third determination value is set as a value larger than the first predetermined number of times used to use the first determination value.

[Fourth judgment value: indicator for distinguishing smoke from lighting]
The determination value calculation unit 31 calculates the movement distance of the centroid position of the white image for each of the positive synchronization image and the negative synchronization image and sets it as a fourth determination value.

  The fourth determination value is a fourth index value for preventing erroneous determination due to the influence of brightness fluctuation when the light source is turned on. That is, when the light source is turned on slowly, the change in brightness tends to spread around a certain point. For this reason, the gravity center position of the white pixel appearing in the synchronous image hardly changes.

  On the other hand, when smoke is generated, the brightness gradation moves slowly along the flow of the hot air flow, so that the barycentric position of the white pixel appearing in the synchronous image changes. Therefore, when the distance of the center of gravity position, which is the fourth determination value calculated for each area, is equal to or smaller than a predetermined fourth threshold value, the brightness of the area changes when the light source is turned on. It can be determined that this is the area. Therefore, in either the positive synchronization image or the negative synchronization image, by excluding the region where the distance of the center of gravity position, which is the fourth determination value, is equal to or less than the fourth threshold from the target region for smoke determination, It is possible to remove the influence of the brightness fluctuation when the light source is turned on, which is one of the erroneous determination factors.

[Fifth judgment value: Index value for removing the influence of brightness fluctuations caused by factors other than smoke]
The determination value calculation unit 31 calculates the sum of the difference values of all the pixels for each area of the time-series image for a predetermined number of times, and sets it as the fifth determination value.

  This fifth determination value is a fifth index value for removing the influence of brightness fluctuations caused by factors other than smoke. That is, the difference in brightness change due to smoke is relatively small, and it is considered that the fifth determination value does not increase. Therefore, when the total value of the time-series difference values, which are the fifth determination values calculated for each area, is equal to or greater than a predetermined fifth threshold value, the area has a luminance change due to the influence of smoke. It can be determined that the region is not generated. Therefore, by excluding the area where the fifth determination value is equal to or greater than the fifth threshold from the smoke determination target area, the influence of brightness fluctuation due to factors other than smoke, which is one of the erroneous determination factors, is removed. be able to.

  As described above, the determination value calculation unit 31 calculates the five determination values from the first to the fifth based on the time-series positive synchronization image and negative synchronization image generated by the preprocessing unit 20. Can be calculated. Therefore, the smoke determination unit 32 specifies a target region for smoke determination based on the first index value based on these five determination values, and makes an erroneous determination based on the second index value to the fifth index value. Identify the factor area.

  And the smoke determination part 32 excluded the erroneous determination factor by excluding the area specified based on the second index value to the fifth index value from the area specified based on the first index value. A highly accurate smoke detection process can be realized. Summarizing specific detection processing by the smoke determination unit 32, the following five conditions are determined.

[Condition 1] An area where the first determination value calculated for the normal synchronization image is equal to or greater than the first threshold is extracted as a smoke detection candidate area based on the normal synchronization. Similarly, a region where the first determination value calculated for the negative synchronization image is equal to or greater than the first threshold is extracted as a smoke detection candidate region based on negative synchronization (extraction of a smoke detection candidate region based on synchronization) ).
[Condition 2] Region in which the second determination value calculated for the positive synchronization image is greater than or equal to the second threshold and the second determination value calculated for the negative synchronization image is greater than or equal to the second threshold Is specified as the first exclusion region (removal of noise influence).
[Condition 3] An area where the third determination value calculated for the positive synchronization image is equal to or greater than the third threshold is specified as a second exclusion area based on the normal synchronization. Similarly, a region in which the third determination value calculated for the negative synchronization image is equal to or greater than the third threshold is specified as a second exclusion region based on negative synchronization (removal of effects such as sunshine change and illumination change). .
[Condition 4] An area where the fourth determination value calculated for the normal synchronization image is equal to or smaller than the fourth threshold is specified as a third exclusion area based on the normal synchronization. Similarly, an area in which the fourth determination value calculated for the negative synchronization image is equal to or less than the fourth threshold is specified as a third exclusion area based on the negative synchronization (removal of influence at the time of lighting).
[Condition 5] An area where the fifth determination value based on the difference value is greater than or equal to the fifth threshold is specified as a fourth exclusion area (removal of the influence of brightness fluctuations caused by factors other than smoke).

  Then, the smoke determination unit 32 selects, from the smoke detection candidate area based on the normal synchronization, the first exclusion area, the second exclusion area based on the normal synchronization, the third exclusion area based on the normal synchronization, and the fourth The area excluding the exclusion area consisting of the four exclusion areas is specified as the smoke detection area based on the normal synchronization.

  Similarly, the smoke determination unit 32 determines, from the smoke detection candidate area based on negative synchronization, the first exclusion area, the second exclusion area based on negative synchronization, the third exclusion area based on negative synchronization, and the first An area excluding the four exclusion areas among the four exclusion areas is specified as a smoke detection area based on negative synchronization.

  Finally, the smoke determination unit 32 specifies an area belonging to at least one of the smoke detection area based on the positive synchronization and the smoke detection area based on the negative synchronization as the area where the smoke is generated.

  As described above, according to the first embodiment, a synchronization image is generated from a plurality of images captured in time series based on time-series fluctuations of difference values. And it has the structure which calculates the some parameter | index for identifying a smoke generation | occurrence | production area | region and removing a misjudgment factor based on the time series data of a difference value and a synchronous image. As a result, by focusing on the synchronization, it is possible to obtain a smoke detection device and a smoke detection method capable of realizing highly accurate smoke detection processing excluding erroneous determination factors.

  In particular, a plurality of images suitable for identifying various misjudgment factors based on the state of occurrence of synchronization and the transition of synchronization for each predetermined region using a plurality of images within a certain period in the past. The index value is calculated and the smoke generation area is specified. Therefore, by appropriately selecting a threshold value for determining each index value or a threshold value applied to a determination condition using the index value, it is possible to realize appropriate smoke detection according to the environment.

  In Embodiment 1 described above, the case where the smoke detection region is specified using all of the index values 1 to 5 has been described, but the present invention is not limited to this. Depending on the environment in which smoke is detected, it is possible to use a threshold value that uses or does not use some of the five index values.

  In the above-described first embodiment, the case where the inspection area in one screen is divided into a plurality of predetermined areas has been described. However, the divided areas do not necessarily have to be divided into equal sizes, and are to be detected. It can be set appropriately according to the environment and image. It is also possible to set an area in the image where there is no risk of smoke as an area not to be detected.

  Moreover, in Embodiment 1 mentioned above, in order to calculate an index value, the 1st-3rd predetermined number of times and the 1st-5th threshold value are used, but these are separate for every division | segmentation area. It can also be set as a value, and can be set appropriately according to the environment or image to be detected.

  In the first embodiment described above, a case is described in which the presence or absence of synchronization is determined using four neighboring pixels that are adjacent in the vertical and horizontal directions of the target pixel as one or more pixels in a predetermined position around the target pixel. did. However, as one or more pixels at a predetermined position around the target pixel, one or more of four neighboring pixels are used, an image adjacent in the diagonal direction of the target pixel is used, or the target pixel is The setting can be appropriately changed according to the environment and the image to be detected, such as using the surrounding 24 pixels defined by 5 × 5 as the center.

  DESCRIPTION OF SYMBOLS 1 Camera, 10 image memory, 20 Pre-processing part, 21 Difference image generation part, 22 Synchronous determination processing part, 23 Synchronous image generation part, 30 Smoke generation | occurrence | production detection part, 31 Determination value calculation part, 32 Smoke determination part

Claims (5)

A smoke detection device that detects the generation of smoke by performing image processing on an image captured by a surveillance camera,
An image memory for storing a plurality of images taken in time series by the monitoring camera as time series images;
Based on the time-series image stored in the image memory, the difference value of the luminance value of each pixel in the latest image and the previous image in time series is calculated, and its own pixel corresponding to the target pixel A pixel in which the sign of the difference value is the same as the sign of the difference value of one or more pixels at a predetermined surrounding position is identified as a synchronous pixel, and the difference value is positive and synchronicity A binary image in which a certain pixel is white is generated as a positive synchronization image in time series, and a binary image in which the difference value is negative and a synchronization pixel is white is defined as a negative synchronization image. A pre-processing unit that generates the sequence;
Based on the positive-synchronous image and the negative-synchronous image generated in time series by the pre-processing unit, a quantitative index value regarding the distribution state of the synchronized luminance change region and the temporal change of the distribution state is obtained. A smoke detection device comprising: a smoke generation detection unit that calculates and identifies a region where smoke is generated based on the calculated magnitude of the index value. The smoke detection device according to claim 1,
The preprocessing unit further calculates the index value based on the difference value calculated in time series,
The smoke generation detection unit divides the positive synchronization image and the negative synchronization image into a plurality of predetermined areas, and as the quantitative index value in each of the plurality of areas,
For each of the positive-synchronized image and the negative-synchronous image generated in chronological order, a pixel in which white pixels exist as the first predetermined number of times or more with respect to the total number of pixels in the area for the latest predetermined number of times Calculate the percentage of the number as the first judgment value,
For each of the positive-synchronous image and the negative-synchronous image generated in chronological order, a pixel in which white pixels exist below the second predetermined number of times with respect to the total number of pixels in the area for the latest predetermined number of times Calculate the percentage of the number as the second judgment value,
For each of the positive-synchronous image and the negative-synchronous image generated in chronological order, the white pixels are larger than the first predetermined number of times with respect to the total number of pixels in the area for the latest predetermined number of times. The ratio of the number of pixels existing as a third predetermined number of times or more is calculated as the third determination value,
For each of the positive-synchronous image and the negative-synchronous image generated in time series, the position of the center of gravity of the white pixel in the previous image in time series and the center of gravity in the area of the white pixel in the latest image The center-of-gravity distance, which is the difference with the position, is calculated as the fourth determination value,
For the difference value calculated in time series, the sum of the difference values in the area for the latest predetermined number of times is calculated as the fifth determination value,
The smoke generation detection unit,
An area where the first determination value calculated for the positive synchronization image is greater than or equal to a first threshold is identified as a smoke detection candidate area based on positive synchronization, and the first calculation calculated for the negative synchronization image Identifying a region where the determination value is equal to or greater than the first threshold as a smoke detection candidate region based on negative synchronization,
An area in which the second determination value calculated for the positive synchronization image is greater than or equal to a second threshold and the second determination value calculated for the negative synchronization image is greater than or equal to a second threshold. Identified as the first exclusion area,
An area in which the third determination value calculated for the positive synchronization image is equal to or greater than a third threshold is specified as a second exclusion area based on positive synchronization, and the second calculation area calculated for the negative synchronization image is specified. A region where the determination value of 3 is equal to or greater than the third threshold is specified as a second exclusion region based on negative synchronization,
An area where the fourth determination value calculated for the positive synchronization image is equal to or smaller than a fourth threshold is specified as a third exclusion area based on positive synchronization, and the first calculation area calculated for the negative synchronization image is determined. A region where the determination value of 4 is equal to or less than the fourth threshold is specified as a third exclusion region based on negative synchronization,
An area where the fifth determination value is equal to or greater than a fifth threshold is specified as a fourth exclusion area,
From the smoke detection candidate region based on the positive synchronization, the first exclusion region, the second exclusion region based on the positive synchronization, the third exclusion region based on the positive synchronization, and the fourth exclusion region A region excluding the four exclusion regions is identified as a smoke detection region based on positive synchronization, and from the smoke detection candidate region based on the negative synchronization, the first exclusion region, the first synchronization region based on the negative synchronization 2 excluded areas, a third excluded area based on the negative synchrony, and an area excluding the excluded area consisting of four of the fourth excluded areas are specified as smoke detecting areas based on negative synchrony, and at least the A smoke detection device that identifies a region belonging to any one of a smoke detection region based on positive synchronization and a smoke detection region based on negative synchronization as a region where the smoke is generated. The smoke detection device according to claim 1 or 2,
The smoke generation detection unit switches the exclusion process by the fourth exclusion region from the first exclusion region to an invalid state according to the setting of the second threshold to the fifth threshold. . The smoke detection device according to any one of claims 1 to 3,
The said pre-processing part judges the presence or absence of a synchronism using four adjacent pixels adjacent to the said pixel of interest up and down and right and left as one or more pixels in the said surrounding predetermined positions. Smoke detection apparatus. A smoke detection method for detecting the generation of smoke by performing image processing on an image captured by a surveillance camera,
A storage step of storing a plurality of images captured in time series by the monitoring camera in an image memory as time series images;
Based on the time-series image stored in the image memory, the difference value of the luminance value of each pixel in the latest image and the previous image in time series is calculated, and its own pixel corresponding to the target pixel A pixel in which the sign of the difference value is the same as the sign of the difference value of one or more pixels at a predetermined surrounding position is identified as a synchronous pixel, and the difference value is positive and synchronicity A binary image in which a certain pixel is white is generated as a positive synchronization image in time series, and a binary image in which the difference value is negative and a synchronization pixel is white is defined as a negative synchronization image. A pre-processing step to generate in the series;
Based on the positive-synchronous image and the negative-synchronous image generated in time series by the preprocessing step, a quantitative index value regarding the distribution state of the synchronized luminance change region and the temporal change of the distribution state is obtained. A smoke detection method comprising: a smoke generation detection step for calculating and identifying a region where smoke is generated based on the calculated magnitude of the index value.

Images