JP7562999B2 - MOTION DETECTION DEVICE AND MOTION DETECTION METHOD - Google Patents

Description

The present invention relates to a technology for detecting moving objects (objects that move).

As a technology for surveillance and the like, a motion detection technology has been proposed that detects moving objects (moving bodies) based on the differences between images captured by a camera. With this motion detection technology, areas where the above differences are large are detected as areas of moving bodies.

In the above-mentioned moving object detection technology, when the contrast of a captured image (captured image) is low, the above-mentioned difference is calculated to be small even in the area of a moving object, so that the moving object cannot be detected with high accuracy.

Patent document 1 proposes a technique for enhancing the contrast of captured images before detecting moving objects.

JP 2014-059798 A

If the contrast is uniform across the entire captured image, moving objects can be detected with high accuracy using images after conventional contrast enhancement (which uniformly enhances (highers) the contrast across the entire captured image).

However, contrast is not necessarily uniform throughout the entire captured image. For example, a camera (IR camera) that detects infrared light (IR (Infrared Ray) light) detects reflected IR light emitted from the IR camera and diffusely reflected by an object. For this reason, the longer the distance from the IR camera to the object, the greater the attenuation of the IR light, and the lower the intensity (detection intensity) of the detected reflected IR light. As a result, in the captured image (IR image), the detection intensity varies according to the distance from the IR camera to each position in the imaging range, and contrast varies within the IR image. Even with a camera that detects natural light, contrast varies within the captured image in scenes with many shadows.

When contrast varies within a captured image, conventional contrast enhancement, which uniformly enhances the contrast of the entire captured image, does not sufficiently enhance the contrast of some areas of the captured image, making it impossible to detect moving objects in those areas with high accuracy.

The present invention has been developed in consideration of the above-mentioned circumstances, and aims to provide a technology that can detect moving objects with high accuracy even when the contrast varies within a captured image.

To achieve the above objective, the present invention adopts the following configuration.

A first aspect of the present invention provides a moving object detection device comprising: a contrast enhancement means for individually enhancing contrast for a plurality of divided regions constituting an area of an image captured within a predetermined range; and a moving object detection means for detecting a moving object within the predetermined range based on a difference between images after contrast enhancement by the contrast enhancement means. The predetermined range is a predetermined range indoors, outdoors, etc. The plurality of images for obtaining the difference are two or more images. The contrast enhancement means may determine a parameter for the contrast enhancement for each of the plurality of divided regions based on pixel values of the divided region.

According to the above-mentioned configuration, contrast enhancement is performed individually for the multiple divided regions that make up the area of the captured image, so that the contrast of each divided region can be sufficiently enhanced, and thus moving objects can be detected with high accuracy in each divided region.

The contrast enhancement may be gamma correction that converts each pixel value of the divided region using a gamma curve based on each pixel value of the divided region. In this case, a gamma curve or the like can be used as the parameter. The contrast enhancement may be processing that changes each pixel value of the divided region so that the difference between the pixel value representing the divided region and each pixel value of the divided region increases. In this case, a representative pixel value (a pixel value representing the divided region) or a coefficient (magnification factor) for increasing the difference from the representative pixel value can be used as the parameter. The pixel value representing the divided region may be the average pixel value of the divided region. In this way, pixel values can be changed (distributed) without bias.

The moving object detection means may detect the moving object by an inter-frame difference method using the difference between an image obtained by performing the contrast enhancement on an image of a current frame and an image obtained by performing the contrast enhancement on an image of a past frame. The past frame is a frame that is a predetermined number of frames before the current frame, and the predetermined number is 1 or more. The moving object detection means may detect the moving object by a background difference method using the difference between an image obtained by performing the contrast enhancement on an image of a current frame and an image obtained by performing the contrast enhancement on a background image prepared in advance. The background image is, for example, an image captured of a predetermined range in the absence of a moving object to be detected (a moving object to be detected).

The frame difference method and the background difference method each have different advantages and disadvantages. For example, in the frame difference method, if a moving object to be detected stops, the moving object will not be detected. Also, when a person is walking toward the camera, the same color of the person's clothes, etc., may be present in both the current frame and the past frame in a certain area, and the difference between the frames may be small, and the person may not be detected. In the background difference method, the accuracy of moving object detection decreases due to disturbance factors such as changes in sunlight conditions, that is, fluctuations in the surrounding environment. Also, when a moving object (moving object that is not the detection target) is included in the background, the moving object included in the background differs between the image of the current frame and the background image, and unnecessary moving objects may be erroneously detected or background areas may be erroneously detected as moving object areas. Such erroneous detection is particularly noticeable when the background contains many moving objects (moving objects that are not the detection target).

For this reason, the moving object detection means may detect the moving object by a combination of an inter-frame difference method that uses the difference between an image obtained by performing the contrast enhancement on an image of the current frame and an image obtained by performing the contrast enhancement on an image of a past frame, and a background difference method that uses the difference between an image obtained by performing the contrast enhancement on an image of the current frame and an image obtained by performing the contrast enhancement on a previously prepared background image. This makes it possible to eliminate the disadvantages of the inter-frame difference method and the background difference method and utilize their advantages, thereby enabling more accurate detection of moving objects.

The background may change due to changes in sunlight or other light over time, changes in the layout of a room, etc. For this reason, the apparatus may further include an update means for updating the background image of an area determined to be background by the frame difference method based on the image of the current frame. This can improve the accuracy of moving object detection by the background difference method.

The accuracy of moving object detection decreases due to differences in contrast between images after contrast enhancement. For this reason, the contrast enhancement means may generate multiple images for obtaining the difference by contrast enhancement using the same parameters. In this way, it is possible to match the contrast between the images after contrast enhancement, thereby improving the accuracy of moving object detection.

The system may further include a type determination means for determining the type of moving object detected by the moving object detection means. In this way, further effects such as improved security in surveillance can be obtained. Types include vehicles such as cars, motorbikes, and bicycles, people, animals, objects, etc.

The higher the resolution of the captured image, the greater the processing load of contrast enhancement of the captured image. For this reason, the system may further include a reduction means for reducing the image captured within the specified range, and the contrast enhancement means may perform the contrast enhancement on the image reduced by the reduction means. In this way, the processing load of contrast enhancement can be reduced.

The image captured of the specified range may be an IR image. The IR image is an image captured by a camera (IR camera) that detects infrared light (IR (Infrared Ray) light).

The second aspect of the present invention provides a moving object detection method characterized by having a contrast enhancement step for individually enhancing the contrast of a plurality of divided regions constituting an area of an image captured within a predetermined range, and a moving object detection step for detecting a moving object within the predetermined range based on the difference between the images after contrast enhancement by the contrast enhancement step.

The present invention can be understood as a moving object detection system having at least a part of the above configuration or functions. The present invention can also be understood as a moving object detection method or a control method for a moving object detection system that includes at least a part of the above processing, a program for causing a computer to execute these methods, or a computer-readable recording medium on which such a program is non-temporarily recorded. The above configurations and processing can be combined with each other to constitute the present invention as long as no technical contradictions arise.

According to the present invention, moving objects can be detected with high accuracy even when contrast varies within a captured image.

FIG. 1 is a block diagram showing an example of the configuration of a moving object detection device to which the present invention is applied. FIG. 2A is a schematic diagram showing an example of the general configuration of a moving object detection system according to an embodiment of the present invention, and FIG. 2B is a block diagram showing an example of the configuration of a PC (moving object detection device) according to the embodiment. FIG. 3A is a diagram showing an example of conventional contrast enhancement, and FIG. 3B is a diagram showing an example of contrast enhancement according to an embodiment of the present invention. 4A to 4C are diagrams showing an example of contrast enhancement (contrast enhancement for one divided region) according to an embodiment of the present invention. FIG. 5 is a diagram showing an example of contrast enhancement (contrast enhancement for one divided region) according to an embodiment of the present invention. FIG. 6 is a flowchart showing an example of a process flow of the PC according to the embodiment of the present invention.

<Application Examples>
An application example of the present invention will be described. In the conventional technology for detecting a moving object, when the contrast of a captured image (captured image) is low, the difference between the captured images is calculated to be small even in the area of the moving object, so that the moving object cannot be detected with high accuracy. If the contrast is uniform throughout the captured image, the moving object can be detected with high accuracy by performing conventional contrast enhancement that uniformly emphasizes (increases) the contrast of the entire captured image. However, when the contrast varies within the captured image, the above conventional contrast enhancement does not sufficiently emphasize the contrast of a part of the captured image, and the moving object cannot be detected with high accuracy in that area.

Figure 1 is a block diagram showing an example of the configuration of a moving object detection device 100 to which the present invention is applied. The moving object detection device 100 has a contrast enhancement unit 101 and a moving object detection unit 102. The contrast enhancement unit 101 performs contrast enhancement individually on a plurality of divided regions that make up the region of an image captured within a specified range. The moving object detection unit 102 detects a moving object within the specified range based on the difference between the images after contrast enhancement by the contrast enhancement unit 101. The contrast enhancement unit 101 is an example of the contrast enhancement means of the present invention, and the moving object detection unit 102 is an example of the moving object detection means of the present invention.

The above-described configuration of the moving object detection device 100 performs contrast enhancement individually for the multiple divided regions that make up the area of the captured image, so that the contrast of each divided region can be sufficiently enhanced, and thus moving objects can be detected with high accuracy in each divided region.

<Embodiment>
An embodiment of the present invention will be described.

2A is a schematic diagram showing an example of the general configuration of a moving object detection system according to this embodiment. The moving object detection system according to this embodiment includes a camera 10 and a PC 200 (personal computer; moving object detection device). The camera 10 and the PC 200 are connected to each other by wire or wirelessly. The camera 10 captures an image of a predetermined range indoors or outdoors, and outputs the captured video to the PC 200. The camera 10 may be a camera that detects natural light, but in this embodiment, it is a camera that detects infrared light (IR (Infrared Ray) light). In other words, the image captured by the camera 10 is an IR image. The PC 200 detects a moving object within a predetermined range based on the video captured by the camera 10. The PC 200 displays the detection result of the moving object (presence or absence of a moving object, the area where the moving object is detected, etc.) on the display unit, records the detection result of the moving object in a storage medium, and outputs the detection result of the moving object to another terminal (such as a smartphone of an administrator in a remote location).

In this embodiment, the PC 10 is a separate device from the camera 10, but the PC 200 may be built into the camera 10. The display unit and storage medium described above may or may not be part of the PC 200. The location where the PC 200 is installed is not particularly limited. For example, the PC 200 may or may not be installed in the same room as the camera 10. The PC 200 may or may not be a computer on the cloud. The PC 200 may be a terminal such as a smartphone carried by the administrator.

Fig. 2(B) is a block diagram showing an example configuration of the PC 200. The PC 200 has an input unit 210, a control unit 220, a storage unit 230, and an output unit 240. The control unit 220 has a contrast enhancement unit 221 and a moving object detection unit 222.

The input unit 210 sequentially performs a process of acquiring images (video frames) of a predetermined range from the camera 10 and outputting the images to the control unit 220 (contrast enhancement unit 221). The camera 10 may sequentially capture still images, in which case the input unit 210 sequentially performs a process of acquiring still images of a predetermined range from the camera 10 and outputting the still images to the control unit 220 (contrast enhancement unit 221).

The control unit 220 includes a CPU (Central Processing Unit), RAM (Random Access Memory), ROM (Read Only Memory), etc., and controls each component and performs various information processing.

The contrast enhancement unit 221 performs contrast enhancement individually on a plurality of divided regions that constitute the region of the captured image obtained by the camera 10. For example, the contrast enhancement unit 221 performs contrast enhancement on the image output by the input unit 210. Then, the contrast enhancement unit 221 outputs the image after contrast enhancement to the moving object detection unit 222. The contrast enhancement unit 221 is an example of a contrast enhancement means of the present invention.

In this embodiment, contrast enhancement is performed individually for the multiple divided regions that make up the area of the captured image, so the contrast of each divided region can be sufficiently enhanced. Specifically, the contrast enhancement unit 221 dynamically (adaptively) determines the contrast enhancement parameters for each of the multiple divided regions based on the pixel values of the divided region. Then, the contrast enhancement unit 221 performs contrast enhancement for each divided region using the determined parameters. This allows the contrast of each divided region to be sufficiently enhanced.

This will be specifically explained using Figures 3(A) and 3(B). Figure 3(A) shows an example of conventional contrast enhancement, and Figure 3(B) shows an example of contrast enhancement according to this embodiment. In Figures 3(A) and 3(B), people H1 and H2 are captured as moving objects in the captured image. Before contrast enhancement, the contrast around person H1 is much lower than the contrast around person H2.

In the contrast enhancement of FIG. 3(A), the contrast of the entire captured image is enhanced uniformly, so even if the contrast around person H2 can be sufficiently enhanced, the contrast around person H1 (which has extremely low contrast) cannot be sufficiently enhanced. For this reason, person H1 cannot be detected in moving object detection. Furthermore, if the contrast around person H1 is sufficiently enhanced, whiteout occurs around person H2, reducing the contrast.

In this embodiment, contrast enhancement is performed for each divided region, so that, as shown in FIG. 3(B), the contrast around person H1 and the contrast around person H2 can both be sufficiently enhanced. Specifically, the contrast of person H2 can be sufficiently enhanced without affecting the contrast around person H1, and the contrast of person H1 can be sufficiently enhanced without affecting the contrast around person H2. In other words, the contrast of a divided region can be sufficiently enhanced without affecting the contrast of other divided regions.

Note that in FIG. 3(B), 12 rectangular regions of the same size arranged in a matrix of 4 horizontal by 3 vertical are shown as the multiple divided regions, but the number, arrangement, size, shape, etc. of the divided regions are not particularly limited. For example, the number of divided regions may be more or less than 12, the multiple divided regions may be arranged in a houndstooth pattern, the multiple divided regions may not be the same size or shape, and the divided regions may be trapezoidal, triangular, etc.

A specific method of contrast enhancement will be described. For example, the contrast enhancement unit 221 performs gamma correction to enhance contrast by converting each pixel value of a divided region using a gamma curve based on each pixel value of the divided region. This process will be described in detail with reference to Figures 4(A) to 4(C). For simplicity, the process for one divided region will be described.

Figures 4(A) to 4(C) show examples of a histogram before contrast enhancement, a gamma curve, and a histogram after contrast enhancement. The histogram before contrast enhancement shows the number of pixels (frequency) in a divided region for each pixel value before contrast enhancement. The gamma curve shows the correspondence between pixel values before contrast enhancement (before conversion) and pixel values after contrast enhancement (after conversion). The histogram after contrast enhancement shows the number of pixels (frequency) in a divided region for each pixel value after contrast enhancement. Figures 4(A) and 4(B) show a case where the histogram before contrast enhancement has one peak (frequency peak), and Figure 4(C) shows a case where the histogram before contrast enhancement has two peaks (frequency peaks). The spread of the peaks (range of pixel values) is different between Figures 4(A) and 4(B).

As shown in Figures 4(A) to 4(C), the contrast enhancement unit 221 determines a gamma curve that stretches the histogram (increases the spread of the distribution). Specifically, the contrast enhancement unit 221 determines the pixel value (peak pixel value) Pa with the highest frequency and the spread Pb of the peak for each peak (peak of frequency) in the histogram before contrast enhancement. Then, the contrast enhancement unit 221 determines the position at which the gamma curve is inclined according to the peak pixel value Pa, and determines the range at which the gamma curve is inclined (range of pixel values before contrast enhancement) according to the spread Pb. Specifically, the contrast enhancement unit 221 determines a gamma curve that, centered on the peak pixel value Pa, widens the peak as the spread Pb becomes narrower. As a result, as shown in Figures 4(A) and 4(B), gamma correction can be performed to sufficiently widen the peak regardless of the spread Pb. Furthermore, as shown in Figure 4(C), when there are multiple peaks in the histogram before contrast enhancement, gamma correction can be performed to sufficiently widen each peak. A sufficiently broadened peak corresponds to a sufficient enhancement of contrast.

This method is also called histogram equalization or histogram equalization. In other words, it can be said that the contrast enhancement unit 221 achieves contrast enhancement by histogram equalization.

In addition, instead of the peak pixel value Pa (mode), other pixel values (such as the average value or the median value) that represent the peak in the histogram before contrast enhancement may be used. In addition, in the case of the above method, a gamma curve, peak pixel value Pa, spread Pb, etc. may be used as the above parameters. In addition, although the gamma curve is illustrated as a connection of straight lines in which the pixel value after conversion changes linearly with respect to the change in the pixel value before conversion, the gamma curve may be a curve in which the pixel value after conversion changes nonlinearly with respect to the change in the pixel value before conversion.

As a contrast enhancement, the contrast enhancement unit 221 may perform a process of changing each pixel value of the divided region so as to increase the difference between the pixel value representing the divided region and each pixel value of the divided region. This process will be specifically described with reference to FIG. 5. For simplicity, the process for one divided region will be described.

5 shows an example of a histogram before contrast enhancement and a histogram after contrast enhancement. For example, the contrast enhancement unit 221 changes pixel values using the following formula 1. In formula 1, "Pave" is the average pixel value of the divided region before contrast enhancement, and "P
" is the pixel value before contrast enhancement, "C" is a coefficient (magnification) for increasing the difference from the average pixel value Pave, and "P'" is the pixel value after contrast enhancement. The coefficient C is a value greater than 0 and less than or equal to 1. This makes it possible to realize the change in pixel value as described above. In other words, it is possible to increase the deviation of each pixel value from the average pixel value Pave, and to sufficiently enhance the contrast.
P'=Pave-C×(Pave-P)...(Formula 1)

In addition, instead of the average pixel value Pave (average value), other representative pixel values (pixel values representing the divided regions) before contrast enhancement, such as the median value, may be used. By using the average value or median value as the representative pixel value, the pixel values can be changed (distributed) without bias. In addition, in the case of the above method, the average pixel value Pave, the coefficient C, etc. can be used as the above parameter.

It is also preferable that the coefficient C is determined for each divided region. For example, it is preferable that the contrast enhancement unit 221 determines the coefficient C so that the difference in pixel value between a pixel having a pixel value smaller than the average pixel value Pave and a pixel having a pixel value larger than the average pixel value Pave is greater than a predetermined threshold value. The contrast enhancement unit 221 may determine the coefficient C based on the variance of all pixel values in the divided region. For example, the contrast enhancement unit 221 may increase the coefficient C as the variance decreases.

Returning to the explanation of FIG. 2B, the moving object detection unit 222 detects a moving object in a predetermined range based on the difference between the multiple images (multiple images after contrast enhancement) output by the contrast enhancement unit 101. The moving object detection unit 222 then outputs the detection result of the moving object to the output unit 240. Any algorithm may be used for the moving object detection, and the moving object may be detected by a simple method of detecting an area where the difference is greater than a predetermined threshold as the moving object area, or it may not be. The moving object may be detected using a trained model generated by existing machine learning, and specifically, the moving object may be detected using a trained model generated by deep learning (e.g., R-CNN, Fast R-CNN, YOLO, SSD, etc.). In addition, in this embodiment, the difference between two images is obtained, but the number of images to obtain the difference may be more than two. The moving object detection unit 222 is an example of a moving object detection means of the present invention.

The storage unit 230 stores programs executed by the control unit 220, various data used by the control unit 220, and the like. For example, the storage unit 230 is an auxiliary storage device such as a hard disk drive or a solid state drive.

The output unit 240 displays the detection result (moving object detection result) output by the moving object detection unit 222 on a display unit, records it on a storage medium, or outputs it to another terminal (such as a smartphone of an administrator in a remote location). Note that the processing of the moving object detection unit 222 and the output unit 240 may be controlled so that these processes are performed.

Figure 6 is a flowchart showing an example of a processing flow of the PC 200. The PC 200 repeatedly executes the processing flow of Figure 6. The repetition period of the processing flow of Figure 6 is not particularly limited, but in this embodiment, the processing flow of Figure 6 is repeated at the frame rate of the image capture by the camera 10 (e.g., 30 fps).

First, the input unit 210 acquires an image of a specified range (image of the current frame; current image) from the camera 10 (step S601).

Next, the contrast enhancement unit 221 determines the parameters for contrast enhancement (step S602), and enhances the contrast of the current image acquired in step S601 using the determined parameters (step S603). Details of the processes of steps S602 and S603 are as described above. Then, the contrast enhancement unit 221 records the parameters used for contrast enhancement in step S603 (the parameters determined in step S602) in the storage unit 230 (step S604). The processes of steps S602 to S604 are performed individually for each divided region. Furthermore, the contrast enhancement unit 221 records the current image acquired in step S601 in the storage unit 230 for the process of step S606 described later.

Next, the contrast enhancement unit 221 acquires a background image prepared in advance in the storage unit 230, and enhances the contrast of the acquired background image by the above-mentioned method (step S605). The background image is, for example, an image captured of a predetermined range in the absence of a moving object to be detected (a moving object to be detected). Then, the contrast enhancement unit 221 acquires a previous image (an image of a previous frame) recorded in the storage unit 230, and enhances the contrast of the acquired previous image by the above-mentioned method (step S606). The previous image is an image of a frame a predetermined number of frames before the current image, and the predetermined number is 1 or more. The processing of steps S605 and S606 is performed individually for each divided area.

The accuracy of moving object detection decreases due to differences in contrast between images after contrast enhancement. For this reason, in this embodiment, the contrast enhancement unit 221 uses the same parameters (the parameters determined in step S602; the parameters recorded in step S604) in the contrast enhancement of steps S605 and S606 as in the contrast enhancement of step S603. In this way, it is possible to match the contrast between the images after contrast enhancement, and thus improve the accuracy of moving object detection.

However, this is not limited to the above, and the contrast enhancement unit 221 may determine the parameters to be used in the contrast enhancement in step S605 based on the background image, and determine the parameters to be used in the contrast enhancement in step S606 based on the previous image. The contrast enhancement unit 221 may use the parameters determined based on the background image or the previous image in the contrast enhancement in steps S603, S605, and S606.

Next, the moving object detection unit 222 detects a moving object by a background subtraction method using the difference between the current image after contrast enhancement in step S603 and the background image after contrast enhancement in step S605 (step S607). Then, the moving object detection unit 222 detects a moving object by an inter-frame subtraction method using the difference between the current image after contrast enhancement in step S603 and the previous image after contrast enhancement in step S606 (step S608). Next, the moving object detection unit 222 generates detection information indicating the moving object detection result based on the detection results of steps S607 and S608 (step S609).

Note that only one of the frame difference method and the background difference method may be used. When only the background difference method is used, the processes of steps S606 and S608 are omitted. Recording of the current image in the storage unit 230 may also be omitted. When only the frame difference method is used, the processes of steps S605 and S607 are omitted.

However, the frame difference method and the background difference method each have different advantages and disadvantages. For example, in the frame difference method, when a moving object to be detected stops, the moving object cannot be detected. In addition, when a person walks toward the camera, in a certain area, the same color of the person's clothes, etc., is present in both the current frame and the past frame, and the difference between the frames becomes small, and the person may not be detected. In the background difference method, the accuracy of moving object detection decreases due to disturbance factors such as changes in sunlight conditions, that is, fluctuations in the surrounding environment. In addition, when a moving object (a moving object that is not a detection target) is included in the background, the moving image included in the background is different between the image of the current frame and the background image, and an unnecessary moving object is erroneously detected, or a background area is erroneously detected as a moving object area. In particular, when the background includes many moving objects (moving objects that are not a detection target), such erroneous detection becomes noticeable.

For this reason, in this embodiment, moving objects are detected by combining the frame-to-frame difference method and the background difference method. This makes it possible to eliminate the disadvantages of the frame-to-frame difference method and the background difference method and utilize their advantages, thereby enabling moving objects to be detected with higher accuracy. For example, in step S609, the moving object detection unit 222 can narrow down the moving object area to an area in which moving objects have been detected by both the frame-to-frame difference method and the background difference method, and generate detection information. The moving object detection unit 222 can also generate detection information in such a way that the detection results obtained by the frame-to-frame difference method and the detection results obtained by the background difference method can be distinguished.

Finally, the output unit 240 outputs the detection information generated in step S609 to a display unit, a storage medium, a smartphone, etc. (step S610).

As described above, according to this embodiment, contrast enhancement is performed individually for each of the divided regions that make up the area of the captured image, so that the contrast of each divided region can be sufficiently enhanced, and thus moving objects can be detected with high accuracy in each divided region.

Note that the background may change due to changes in sunlight or other light over time, changes in the layout of the room, etc. For this reason, the control unit 220 may further update the background image based on the current image for areas determined to be background by the frame-to-frame difference method. This can improve the accuracy of moving object detection using the background difference method. The background image may be updated (updated for a portion of the area) by replacing it with the current image, or by replacing it with an image obtained by weighting and combining the current image and the background image. The background image may be updated only for areas determined to be background by both the frame-to-frame difference and the background difference (areas where no moving object was detected). The background image may be updated at a predetermined time interval, such as every hour, rather than for each frame.

The control unit 220 may further perform type determination (moving object recognition; moving object identification) to determine the type of the moving object detected by the moving object detection unit 222. The type may be a vehicle such as a car, a motorcycle, or a bicycle, or a person, an animal, or an object. In this case, the detection information may also include information on the type of the moving object. By doing so, further effects such as improved security in surveillance can be obtained. Any algorithm may be used for the type determination. For example, the type of the moving object may be determined by an existing type determination, and specifically, the type of the moving object may be determined using a classifier (recognizer; discriminator) that combines image features such as HoG and Haar-like with boosting. The type of the moving object may be determined using a trained model generated by existing machine learning, and specifically, the type of the moving object may be determined using a trained model generated by deep learning (for example, R-CNN, Fast R-CNN, YOLO, SSD, etc.).

In addition, the higher the resolution of the captured image, the greater the processing load of contrast enhancement of the captured image. For this reason, the control unit 220 may further perform a reduction process to reduce the image captured within a specified range, and then perform contrast enhancement on the reduced image. In this way, the processing load of contrast enhancement can be reduced.

＜Other＞
The above-described embodiment merely describes an exemplary configuration of the present invention. The present invention is not limited to the above-described specific embodiment, and various modifications are possible within the scope of the technical concept thereof.

<Appendix 1>
A contrast enhancement means (101, 221) for individually enhancing contrast for a plurality of divided regions constituting an image region obtained by capturing a predetermined range;
and a moving object detection means (102, 222) for detecting a moving object within the predetermined range based on a difference between images after contrast enhancement by the contrast enhancement means.

<Appendix 2>
A contrast enhancement step (S602, S604, S605) of individually enhancing contrast for a plurality of divided regions constituting an image region obtained by capturing a predetermined range;
a moving object detection step (S606 to S608) of detecting a moving object within the predetermined range based on a difference between images after contrast enhancement in the contrast enhancement step.

100: Moving object detection device 101: Contrast enhancement unit 102: Moving object detection unit 10: Camera 200: PC (moving object detection device)
210: Input unit 220: Control unit 221: Contrast enhancement unit 222: Moving object detection unit 230: Storage unit 240: Output unit

Claims (13)

A contrast enhancement unit that performs contrast enhancement individually on a plurality of divided regions that constitute an image region obtained by capturing a predetermined range;
a moving object detection unit that detects a moving object within the predetermined range based on a difference between the images after contrast enhancement by the contrast enhancement unit ,
the contrast enhancement means determines a parameter of the contrast enhancement to be applied to each divided region based on a pixel value of each divided region in the image of the current frame, and performs the contrast enhancement on each of the image of the current frame and another image used in the difference processing, using the parameter of each divided region determined based on the image of the current frame;
The moving object detection means detects a moving object within the specified range based on a difference between the image of the current frame after contrast enhancement and the other image after contrast enhancement, in which the contrast enhancement is performed using the parameters of each divided area determined based on the image of the current frame . 2. The moving object detection device according to claim 1 , wherein the contrast enhancement is gamma correction that converts each pixel value of the divided region using a gamma curve based on each pixel value of the divided region. 2. The moving object detection device according to claim 1 , wherein the contrast enhancement is a process of changing each pixel value of the divided region so as to increase a difference between a pixel value representing the divided region and each pixel value of the divided region. The moving object detection device according to claim 3 , wherein the pixel value representing the divided region is an average pixel value of the divided region. The moving object detection device according to any one of claims 1 to 4, characterized in that the moving object detection means detects the moving object by an inter-frame difference method that uses the difference between an image obtained by performing the contrast enhancement on an image of a current frame and an image obtained by performing the contrast enhancement on an image of a past frame . The moving object detection device according to any one of claims 1 to 4, characterized in that the moving object detection means detects the moving object by a background subtraction method that uses the difference between an image obtained by performing the contrast enhancement on an image of a current frame and an image obtained by performing the contrast enhancement on a previously prepared background image. The moving object detection means
A frame difference method using a difference between an image obtained by performing the contrast enhancement on an image of a current frame and an image obtained by performing the contrast enhancement on an image of a past frame; and
The moving object detection device according to any one of claims 1 to 4, characterized in that the moving object is detected by a combination of a background subtraction method that uses the difference between an image obtained by performing the contrast enhancement on an image of the current frame and an image obtained by performing the contrast enhancement on a previously prepared background image. 8. The moving object detection device according to claim 7 , further comprising an update unit that updates the background image based on the image of the current frame for an area determined to be background by the interframe difference method. 9. The moving object detection device according to claim 1, further comprising a type determination unit for determining a type of the moving object detected by the moving object detection unit. The image capturing device further includes a reduction unit for reducing the size of the image captured within the predetermined range.
10. The moving object detection device according to claim 1, wherein the contrast enhancing means performs the contrast enhancement on the image after reduction by the reduction means. 11. The moving object detection device according to claim 1, wherein the image captured within the predetermined range is an IR image. a contrast enhancement step of performing contrast enhancement individually on a plurality of divided regions constituting an image region obtained by capturing a predetermined range;
a moving object detection step of detecting a moving object within the predetermined range based on a difference between images after the contrast enhancement step ;
In the contrast enhancement step, a parameter of the contrast enhancement to be applied to each divided region is determined based on a pixel value of each divided region in the image of the current frame, and the contrast enhancement is performed on each of the image of the current frame and another image used in the difference processing, using the parameter of each divided region determined based on the image of the current frame;
A moving object detection method characterized in that in the moving object detection step, the contrast enhancement is performed using the parameters of each divided area determined based on the image of the current frame, and a moving object within the specified range is detected based on the difference between the image of the current frame after contrast enhancement and the other image after contrast enhancement . A program for causing a computer to execute each step of the moving object detection method according to claim 12 .

Images