JP5149918B2 - Vehicle periphery monitoring device - Google Patents

Description

  The present invention relates to an apparatus for monitoring the periphery of a vehicle using a captured image obtained by an infrared camera, and more specifically, a vehicle periphery monitoring apparatus that extracts an object by binarization processing of the captured image. About.

  2. Description of the Related Art Conventionally, there has been proposed an apparatus that mounts an infrared camera on a vehicle, binarizes a captured image captured around the vehicle, and extracts a high-temperature object such as a pedestrian or animal. In Patent Document 1 below, a method is proposed in which a luminance value histogram of a captured image obtained using an infrared camera is created, and a threshold value to be divided into a background image and an object image is determined based on the luminance value histogram. Has been. By the binarization process using such a threshold value, the high temperature object is extracted separately from the background.

JP 2003-216949 A

  In addition to living bodies such as pedestrians and animals, various artificial structures such as utility poles and walls can exist around the vehicle. In order to extract a living body such as a pedestrian or an animal from a background as a high-temperature object, it is desirable that such an artificial structure is classified as a background in the binarization process. However, depending on the type of man-made structure and its arrangement, and the surrounding environment of the vehicle such as the temperature around it, the man-made structure may be hotter than other background parts. Even if such a conventional method is employed, the artificial structure may be classified as a high-temperature object. Therefore, there is a demand for a technique that can extract a desired object from the background with better accuracy in the binarization process without depending on the surrounding environment of the vehicle.

  According to one aspect of the present invention, a vehicle periphery monitoring device that monitors the periphery of a vehicle using a captured image obtained by an infrared camera mounted on the vehicle includes an outside air temperature detecting unit that detects the outside air temperature of the vehicle, Temperature difference calculating means for obtaining a temperature difference between the surface temperature of the object estimated based on the temperature and the outside air temperature, and the imaging based on the luminance value of the background in the captured image and the luminance difference corresponding to the temperature difference Object luminance value calculating means for calculating the luminance value of the object in the image, and an object for extracting the object by binarizing the captured image obtained by the infrared camera using the luminance value of the object as a threshold value Extraction means.

  The relationship between the surface temperature of the object and the outside air temperature is determined in advance, and the surface temperature can be estimated from the outside air temperature. This invention is made based on this knowledge, and calculates the temperature difference between the surface temperature of the object estimated based on the detected outside air temperature and the outside air temperature. Since the luminance value of the background is considered to correspond to the outside air temperature, the luminance value corresponding to the object can be calculated based on the luminance value of the background and the luminance difference corresponding to the temperature difference. By binarizing the calculated luminance value as a threshold value, it is possible to extract the object separately from the background portion other than the object. Therefore, for example, when a pedestrian is used as an object, by predetermining the relationship between the surface temperature of the pedestrian and the outside air temperature, it is possible to prevent an artificial structure or the like from being erroneously extracted as the object. can do.

  Other features and advantages of the present invention will be apparent from the detailed description that follows.

The block diagram which shows the structure of the periphery monitoring apparatus according to one Example of this invention. The figure for demonstrating the attachment position of the camera according to one Example of this invention. 3 is a flowchart illustrating a process in an image processing unit according to an embodiment of the present invention. The figure which shows the map which prescribes | regulates the relationship between the outside temperature and the surface temperature of a target object according to one Example of this invention. The figure for demonstrating the threshold value setting of a binarization process according to one Example of this invention. The figure showing the comparison of the conventional binarized image and the binarized image according to one Example of this invention. The figure for demonstrating the estimation of the magnitude | size of the target object in a captured image according to one Example of this invention. The figure for demonstrating the setting of an object area | region and object determination according to one Example of this invention. The figure for demonstrating the other method of the setting of an object area | region according to one Example of this invention. The figure for demonstrating the method of setting the threshold value of a binarization process using road surface temperature according to one Example of this invention.

  Next, an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of a vehicle periphery monitoring device according to an embodiment of the present invention. The device is mounted on a vehicle and can detect two far-infrared cameras 1R and 1L, a sensor 5 for detecting the temperature of the outside air around the vehicle (outside air temperature), and image data obtained by the cameras 1R and 1L. The image processing unit 2 for detecting an object in front of the vehicle based on the vehicle, the speaker 3 for generating an alarm by voice based on the detection result, and the image obtained by the camera 1R or 1L are displayed, and the driver The head up display (henceforth HUD) 4 which performs the display for making it recognize the target object ahead of a vehicle is provided.

  As shown in FIG. 2, the cameras 1 </ b> R and 1 </ b> L are disposed at the front part of the vehicle 10 at positions symmetrical with respect to the central axis passing through the center of the vehicle width. The two cameras 1R and 1L are fixed to the vehicle so that their optical axes are parallel to each other and their height from the road surface is equal. The infrared cameras 1R and 1L have a characteristic that the higher the temperature of the object than the background temperature, the higher the level of the output signal (that is, the greater the luminance value in the captured image).

  The image processing unit 2 includes an A / D conversion circuit that converts an input analog signal into a digital signal, an image memory that stores a digitized image signal, a central processing unit (CPU) that performs various arithmetic processing, and a data RAM (Random Access Memory) used to store data, ROM (Read Only Memory) that stores programs executed by the CPU and data used (including tables and maps), driving signals for the speaker 3, display signals for the HUD 4, etc. Is provided. The output signals of the cameras 1R and 1L and the output signal of the sensor 5 are converted into digital signals and input to the CPU. As shown in FIG. 2, the HUD 4 is provided so that a screen 4 a is displayed at a front position of the driver on the front window of the vehicle 10. Thus, the driver can visually recognize the screen displayed on the HUD 4.

  FIG. 3 is a flowchart showing a process executed by the image processing unit 2. The process can be performed at predetermined time intervals.

  In steps S11 to S13, the output signals of the cameras 1R and 1L (that is, captured image data) are received as input, A / D converted, and stored in the image memory. The stored image data is a gray scale image including luminance information.

  The following steps S14 to S19 are processes for extracting a desired object separately from the background in the binarization process. In this embodiment, a pedestrian is described as the desired object.

  In step S14, the outside air temperature i (° C.) detected by the outside air temperature sensor 5 is acquired. In step S15, a background luminance value Tb is determined.

  The luminance value of the background can be determined by an arbitrary method. In this embodiment, a luminance value histogram is created based on the grayscale image, and the highest luminance value is used as the background luminance value Tb. This is because the captured image usually has the largest area occupied by the background.

  In step S16, a map as shown in FIG. 4 is referred to based on the detected outside air temperature i. Here, the map will be described. The head of a pedestrian is a part where the skin is exposed to the outside air mainly on the face and there are few things that obstruct the heat source. Therefore, focus on the head surface exposed to the outside air. As a result of examining the relationship between the temperature of the head surface (hereinafter referred to as the surface temperature) fa (° C.) and the outside air temperature i (° C.) through experiments and simulations, the relationship between the two is shown in FIG. It turned out that there was such a relationship. In this figure, the horizontal axis indicates the outside air temperature i (° C.), and the vertical axis indicates the surface temperature fa (° C.). As shown in this figure, the surface temperature fa can be estimated from the outside air temperature i.

  The surface temperature fa changes as indicated by the curve 101 with respect to the outside air temperature i, and the surface temperature fa increases as the outside air temperature i increases. For a given outside air temperature i, the difference in surface temperature fa (i) relative to the outside air temperature i is represented by the difference between the curve 101 and the line 103 (this is a straight line representing fa = i). Is called the surface temperature difference and is represented by F (i). That is, F (i) = surface temperature fa (i) −outside air temperature i. As shown in the figure, the surface temperature difference F (i) tends to decrease as the outside air temperature i increases.

  In this embodiment, a predetermined margin range (margin) T (° C.) is set with F (i) as the center in order to further improve the object extraction accuracy. The upper limit of the margin range is indicated by a dotted line 101U, and the difference between the upper limit and the outside air temperature i is represented by F (i) max. The lower limit of the margin range is indicated by a dotted line 101L, and the difference between the lower limit and the outside air temperature i is represented by F (i) min.

  The map as shown in FIG. 4 is stored in advance in the memory of the image processing unit 2. The image processing unit 2 refers to the map based on the detected outside air temperature i (° C.) to obtain the surface temperature fa corresponding to the outside air temperature i, and the surface temperature between the surface temperature fa and the outside air temperature i. A difference F (i) is calculated, and an upper limit value F (i) max and a lower limit value F (i) min for the surface temperature difference F (i) are calculated using a predetermined margin range T. Here, the margin range T may be changed according to the outside air temperature i, or may be constant.

  Alternatively, the upper limit value F (i) max and the lower limit value F (i) min for the surface temperature difference F (i) with respect to each outside air temperature i are stored in the memory, and the surface temperature fa is obtained from the outside air temperature i. The upper limit value F (i) max and the lower limit value F (i) min may be obtained directly from the outside air temperature i by skipping.

Returning to FIG. 3, in step S17, a luminance difference corresponding to the upper limit value F (i) max and the lower limit value F (i) min of the surface temperature difference F (i) is calculated. The ratio of the luminance value change with respect to the temperature change is determined by the specifications of the far-infrared camera, and this is represented by the parameter SiTF. Thus, in correspondence with the upper limit value F (i) max and the lower limit value F (i) min of the surface temperature difference F (i), the upper limit value dTmax and the lower limit value dTmin of the luminance difference are expressed by the following equation (1). Are calculated respectively.
dTmax = SiTF × F (i) max
dTmin = SiTF × F (i) min (1)

In step S18, a threshold for binarization processing is calculated. Referring now to FIG. 5, an example of the brightness value histogram of the grayscale image acquired in step S13 is shown. As described above, in step S15, the highest luminance value (peak luminance value) is set as the background luminance value Tb. It can be considered that the background luminance value Tb corresponds to the outside air temperature i. Therefore, as shown in the following formula (2), the luminance value Tcmax of the surface temperature having the upper limit value F (i) max of the surface temperature difference with respect to the outside air temperature i is different from the luminance value Tb of the background. Has an upper limit dTmax. Similarly, the brightness value Tcmin of the surface temperature having the upper limit value F (i) min of the surface temperature difference with respect to the outside air temperature i has the lower limit value dTmin of the brightness difference with respect to the background brightness value Tb.
Tcmax = Tb + dTmax
Tcmin = Tb + dTmin (2)

  The upper limit luminance value Tcmax and the lower limit luminance value Tcmin are set as threshold values for the binarization process. An area 111 defined by these two threshold values is shown in FIG. 5, and the area 111 is a luminance area of an object to be extracted.

  In step S19, using the threshold value set in step S18, the grayscale image acquired in step S13 (in this embodiment, the captured image obtained by the camera 1R is used, but the captured image obtained by the camera 1L is used. However, binarization processing is performed. For each pixel of the captured image, when the luminance value of the pixel is within the luminance region 111, the pixel is determined to be an object to be extracted, and the pixel is set as a white region of value 1, and the luminance value of the pixel is the luminance value. When it is outside the area 111, it is determined as the background, and the pixel is set to a black area having a value of zero.

  Here, referring to FIG. 6, a diagram schematically showing an image is shown. (A) shows a grayscale image (captured image), and in the figure, the difference in gradation is represented by the difference in the type of hatching. (B) shows the image binarized by the conventional method, and (c) shows the image binarized by the method of steps S14 to S19. In the figure, the black area is represented by a hatched area.

  In addition to the pedestrian 121, an artificial structure such as a utility pole 125 or an automobile 127 is captured in the gray scale image. Conventionally, depending on the threshold used for binarization, as shown in (b), these artificial structures 125 and 127 may also be extracted as objects similar to pedestrians, that is, as white areas. It was.

  On the other hand, according to the above technique of the present invention, the surface temperature of the object (in this embodiment, a pedestrian) with respect to the outside air temperature is estimated, and based on the temperature difference between the estimated surface temperature and the outside air temperature, Since the luminance area of the object is set, only the head portion of the pedestrian 121 can be extracted as shown in the white area 131 of (c) (this area is hereinafter referred to as the head area). Even if the artificial structure 125 and the pedestrian 121 are overlapped and imaged as shown in (a), only the pedestrian 121 can be easily extracted as shown in the white area 131 of (c). . As described above, according to the present invention, the object can be extracted with better separation from the background portion other than the object.

  Returning to FIG. 3, in step S20, the size of the pedestrian whole body in the captured image is estimated based on the extracted head region. The estimation can be realized by an arbitrary method, and an example thereof will be specifically described here.

  Referring to FIG. 7, in (a), the head region extracted from the binary image is represented by a black region, and the width thereof is represented by w (represented by the number of pixels). The width w can be calculated, for example, by setting a rectangle circumscribing the head region and obtaining the width. Parts other than the pedestrian's head are shown by dotted lines, but this is the part that has not yet been extracted. In step S20, an object is to estimate the height h (expressed by the number of pixels) in the captured image of the pedestrian.

  For this estimation, as shown in (b), a general size of a pedestrian in real space, that is, a head width Wa and a height Ha are set in advance. For example, a value based on the average value of adults can be set in Wa and Ha (for example, Wa is 20 (cm) and Ha is a value within 160 to 170 (cm)).

  Further, (c) is a diagram showing the arrangement relationship between the camera 1R and the object on the XZ plane, and (d) is an arrangement relationship between the camera 1R and the object on the YZ plane. FIG. Here, X indicates the vehicle width direction of the vehicle 10, Y indicates the vehicle height direction of the vehicle 10, and Z indicates the distance direction from the vehicle 10 to the object. The camera 1R includes an image sensor 11R and a lens 12R. f indicates the focal length of the lens 12R.

In the figure of (c), when the distance to the object is Z (cm), the distance Z is calculated as the following formula (3). Here, pcw indicates a pixel interval in the X direction, that is, a length (cm) per pixel.
Z = Wa * f / (w * pcw) (3)

From the diagram of (d), using the distance Z, the height h (cm) in the captured image of the pedestrian can be calculated as in the following equation (4). Here, pch represents a pixel interval in the Y direction, that is, a length (cm) per pixel.
h = (Ha / pch) × f / Z (4)

  Thus, the size of the pedestrian in the captured image can be estimated to have the width w and the height h. In consideration of the fact that the width of the body is generally wider than the width of the head, alternatively, a value obtained by adding a predetermined margin value to the width w of the head region is set to the width w described above. It may be used instead.

  Returning to FIG. 3, in step S <b> 21, an object region is set on the captured image (may be a grayscale image or a binary image) according to the size of the pedestrian estimated in step S <b> 20. Here, referring to FIG. 8A, the head region 131 extracted as described above is shown. As shown by a thick frame, an object region 141 having a width w of the head region 131 and a height h from the apex of the head region 131 (y coordinate value yu in the figure) is set. Thus, the position of the object in the captured image is specified.

  Returning to FIG. 3, in this embodiment, step S <b> 22 is executed to execute the object determination process for the set object area 141, and the object imaged in the object area 141 is walked. To determine whether or not For example, a pedestrian can be determined by any appropriate object determination method using a known shape matching method (for example, Japanese Patent Application Laid-Open No. 2007-264778). This process is performed using a grayscale image. FIG. 8B shows a pedestrian 151 whose shape has been determined in this way.

  If it is determined in step S22 that the object is a pedestrian, the process proceeds to step S23 to perform an alarm determination process. In this process, it is determined whether or not an alarm is actually output to the driver. If the determination result is affirmative, an alarm is output.

  For example, it is determined from the output of a brake sensor (not shown) whether or not the driver of the vehicle is performing a brake operation. If the brake operation is not performed, an alarm output can be performed. In the alarm output, an audio alarm is issued through the speaker 3, and an image obtained by, for example, the camera 1R is displayed on the screen 4a by the HUD 4, and a pedestrian is highlighted. The highlighting may be performed by any method, for example, it can be highlighted by surrounding it with a colored frame. In this way, the driver can more reliably recognize the pedestrian existing in front of the vehicle. The alarm output may be performed using either one of the alarm and the image display.

  As another method of step S20 described above, for example, from the height of the head region 131 (the height of a rectangle circumscribing the head region 131 can be used and represented by the number of strokes) and the number of head and body, You may calculate the height h of the pedestrian in a captured image. For example, if the height of the head region 131 is hb and the average number of heads of an adult is 7, the pedestrian height h can be estimated as h = 7 × hb.

  Further, as yet another method of steps S20 and S21, a method of determining the road surface from the luminance value of the lower region of the head region 131 and specifying the object region 141 may be employed. This method will be briefly described with reference to FIG. 9. FIG. 9A is a grayscale image (in the figure, items other than the head region 131 are omitted), and below the extracted head region 131. Then, a mask 161 having a predetermined size is set, and a variance of luminance values in an area covered by the mask (alternatively, a standard deviation which is a square root of the variance may be used) is calculated. Since the road surface is considered to be imaged as an image region having a substantially uniform luminance value, if the variance is higher than a predetermined value, it is determined that the region where the mask is set is not a road surface. In that case, as shown in (b), the mask 161 is moved downward, and the variance is calculated again. This process is repeated while moving the mask 161 downward. When the area covered by the mask 161 is only the road surface, the dispersion shows a low value. As shown in (c), if the position of the mask 161 whose variance is lower than a predetermined value is obtained, the position between the position of the mask 161 and the previous position of the mask 161 (indicated by a dotted line) is obtained. The boundary (y coordinate value is yb) can be determined as the base of the object region 141. Thus, the object region 141 having the width w and the height from the top of the head region (y coordinate is yu) to the boundary is extracted.

  In the above-described embodiment, the highest luminance value is set as the background luminance value Tb in the luminance value histogram, and this is associated with the outside air temperature. Alternatively, the background brightness value Tb may be determined as follows by distinguishing between the outside air temperature and the road surface temperature. That is, in the case of a camera arranged in front of the vehicle as shown in FIG. 2, since the area of the road surface occupied in the captured image is large, usually the highest luminance value can be associated with the road surface temperature. Therefore, the relationship between the road surface temperature and the outside air temperature is previously defined in a map (not shown) and stored in the memory. This relationship can be obtained by experiment, simulation, or the like.

  Based on the detected outside air temperature i, the map is referred to and a corresponding road surface temperature R is obtained. The temperature difference between the road surface temperature R and the outside air temperature i is calculated. Using the parameter SiTF described above, the temperature difference is converted into a luminance difference dTi. Here, referring to FIG. 10, a luminance value histogram similar to FIG. 5 is shown. The luminance value Tr having the highest frequency is associated with the road surface temperature R. Since the road surface temperature R is usually higher than the outside air temperature i, the brightness value Tr corresponding to the outside air temperature i is calculated by subtracting it from the road surface brightness value Tr by the calculated brightness difference dTi. The background luminance value Tb in step S15 in FIG. When the outside air temperature i is higher than the road surface temperature R, the brightness difference Ti may be added to the road surface brightness value Tr to obtain the brightness value Ti corresponding to the outside air temperature. The luminance region 111 of the object is specified by the luminance values Tcmax and Tcmin having the luminance differences dTmax and dTmin with respect to the background luminance value Tb. Thus, by distinguishing between the outside air temperature and the road surface temperature, the luminance value of the background can be obtained more accurately. Therefore, the threshold used for the binarization process can be set more appropriately, and the accuracy of extracting the object can be further increased.

  Preferably, the temperature difference between the road surface temperature and the outside air temperature may fluctuate depending on the values of external environmental parameters such as weather conditions (whether it is sunny, wind speed, rainfall, etc.) and the elapsed time since sunset. Therefore, a map may be created and stored for each external environment parameter value set in advance, and a map corresponding to the external environment parameter value for the day may be used.

  Similarly, the map of FIG. 4 may be set for each value of an external environment parameter such as a weather condition and stored in the memory. For example, a map can be created and stored for each day on which the wind speed is greater than or equal to a predetermined value and another day on which the wind speed is not, and a map corresponding to the wind speed on that day can be used.

  In the above embodiment, as described with reference to FIG. 4, the upper limit value F (i) max and the lower limit value F (i) min that define the margin range T are set for the surface temperature difference F (i). did. By setting such a margin range, it is possible to set a threshold value for binarization processing so that an object can be extracted more reliably and with good accuracy. However, instead of setting such a margin range T, a luminance difference dT corresponding to the surface temperature difference F (i) is calculated, and this is added to the background luminance value Tb to obtain the object. The luminance value Tc may be calculated. Binarization can be performed using the luminance value Tc as a threshold. For example, a pixel having a luminance value that matches the luminance value Tc is determined to be an object and set in a white region, and a pixel having a luminance value that does not match the luminance value Tc is determined not to be an object. Set to black area. Further, a predetermined range centered on the luminance value Tc may be set in the luminance region of the target object.

  Furthermore, the said Example demonstrates the target object which should be extracted in a binarization process to the example of a pedestrian. Alternatively, the object may be another living body such as an animal. For example, for a predetermined animal, a map as shown in FIG. 3 is created in advance by experiments, simulations, or the like, and the threshold for binarization processing as described above is set using the map. In the case of animals, the whole body is often exposed to the outside air. Therefore, the whole body is extracted in the binarization process in step S19 of FIG. Therefore, steps S20 and S21 may be skipped, and in step S22, shape determination or the like may be performed on the region extracted in step S19 to determine whether the object is an animal. If it is determined to be an animal, an alarm determination in step S23 is performed.

  Further, as shown in the map of FIG. 3, the present invention is applicable to an object having a surface temperature that can preliminarily define a relationship with the outside air temperature through experiments, simulations, and the like. However, it is not necessarily limited to living bodies such as humans and animals.

1R, 1L infrared camera (imaging means)
2 Image processing unit 3 Speaker 4 Head-up display

Claims (4)

A vehicle periphery monitoring device that monitors the periphery of a vehicle using a captured image obtained by an infrared camera mounted on the vehicle,
An outside air temperature detecting means for detecting the outside air temperature of the vehicle;
A temperature difference calculating means for obtaining a temperature difference between a surface temperature of a portion of the living body estimated based on the outside air temperature and the outside air temperature, which is exposed to the outside air; and
Obtains the luminance difference corresponding from the temperature difference, the brightness value of the background in the captured image, by adding the luminance difference corresponding to the temperature difference, the biological luminance values to calculate a luminance value of the living body in the captured image A calculation means ;
An object extraction means for binarizing a captured image obtained by the infrared camera using the luminance value of the living body as a threshold, and extracting only a portion of the living body exposed to the outside air ;
A vehicle periphery monitoring device comprising: Road surface temperature estimating means for estimating a road surface temperature based on the outside air temperature;
An outside air temperature corresponding brightness value calculating means for obtaining a brightness value corresponding to the outside air temperature based on a brightness difference corresponding to a temperature difference between the road surface temperature and the outside air temperature,
A luminance value corresponding to the outside air temperature is set to the luminance value of the background;
The vehicle periphery monitoring apparatus according to claim 1. The portion of the estimated living body exposed to the outside air is a head of the living body, and the biological luminance value calculating means obtains the luminance difference from the temperature difference with reference to a map. The vehicle periphery monitoring apparatus according to 2. The outside air temperature corresponding brightness value calculating means corrects the brightness value having the highest frequency in the brightness value histogram of the captured image based on the brightness difference corresponding to the temperature difference between the road surface temperature and the outside air temperature. Obtaining a luminance value corresponding to the outside air temperature;
The vehicle periphery monitoring apparatus according to claim 2.

Images