Light-receiving device, impact detection method, camera, and monitoring system

Abstract

The present invention relates to a light-receiving device that operates in a camera mounted on a moving body, and that detects an impact on the moving body on the basis of the result of comparing a generated image and a past image which was generated in the past from the image.

Description

Light Receiving Device, Impact Detection Method, Camera, and Monitoring System

[0001] The present technology relates to a light receiving device, an impact detection method, a camera, and a monitoring system.

[0002] Conventionally, there is a technology related to parking monitoring for automobiles that detects the movement of a person approaching an automobile, a collision with the automobile, or the like, and records video in accordance with the detection result.

[0003] Accordingly, a drive recorder has been proposed, which is a system that controls video recording from a camera based on a signal from a Doppler sensor capable of detecting human movement, turns on the power of the camera based on detecting movement of a person, and starts recording video from the camera when movement of a person is detected again after a predetermined time has elapsed (Patent Document 1)

[0004] Japanese Patent Application Laid-Open No. 2023-121886

[0005] Since such a monitoring system for detecting a person often needs to operate for a long time, operation with lower power consumption is required.

[0006] The present technology has been made in view of such problems, and an object thereof is to provide a light receiving device, an impact detection method, a camera, and a monitoring system capable of operating for monitoring with low power consumption.

[0007] In order to solve the above-described problem, a first technology is a light receiving device that operates in a camera mounted on a moving body and detects an impact applied to the moving body based on a result of comparison between a generated image and a past image generated before the image.

[0008] A second technology is an impact detection method, which detects an impact applied to a moving body based on a result of comparison between a generated image and a past image generated before the image in a light receiving device operating in a camera mounted on the moving body.

[0009] A third technology is a camera including a light receiving device that operates in a camera mounted on a moving body and detects an impact applied to the moving body based on a result of comparison between a generated image and a past image generated before the image.

[0010] The fourth technology is a surveillance system consisting of an ECU that controls a moving object and a camera equipped with a light-receiving device that detects impacts on the moving object based on a comparison between the generated image and past images generated in the past.

[0011] This is a diagram showing a camera 200 mounted on an automobile 300. This is a block diagram showing the configuration of the monitoring system 10. This is a block diagram showing the configuration of the camera 200. This is a block diagram showing the configuration of the light receiving device 100. This is a flowchart showing the processing in the monitoring system 10 and the light receiving device 100. This is an explanatory diagram of a low-resolution image. This is an explanatory diagram of motion detection. This is an explanatory diagram of impact detection. This is an explanatory diagram of an image with high resolution in a part of the area. This is an explanatory diagram of an image with high resolution in a part of the area. This is a diagram showing an example of detection based only on a specific area.

[0012] The embodiments of this technology will be described below with reference to the drawings. The description will be in the following order: <Embodiments> [Configuration of the monitoring system 10] [Configuration of the light receiving device 100 and camera 200] [Processing in the monitoring system 10 and light receiving device 100] <Modifications>

[0013] <Embodiment> [Configuration of the monitoring system 10] The configuration of the monitoring system 10 will be described with reference to Figures 1 and 2. The monitoring system 10 consists of a camera 200 equipped with a light receiving device 100 according to this technology and an ECU 310 (Electronic Control Unit) that controls the automobile 300.

[0014] As shown in Figure 1, multiple cameras 200 are mounted on the vehicle 300 to photograph the area around the vehicle 300 as a moving object for parking surveillance. In this embodiment, a total of four cameras 200A, 200B, 200C, and 200D are mounted on the front, rear, left, and right sides of the vehicle 300, forming a surround camera system. This allows for the detection of impacts to the vehicle 300 and motion and human activity around the vehicle 300 for parking surveillance. In the following description, cameras 200A to 200D will be referred to simply as camera 200 unless it is necessary to distinguish between them. The position of the cameras 200 on the vehicle 300 is not limited to the example shown in Figure 1, and the cameras 200 can be mounted in various positions on the vehicle. Furthermore, there is no limit to the number of cameras 200 mounted on the vehicle 300; it may be one to three, or four or more. In the following description, impact detection, motion detection, and human detection for parking surveillance may be collectively referred to as "detection for parking surveillance."

[0015] The camera 200 is equipped with a light-receiving device 100 and captures RGB (Red, Green, Blue) or monochrome images or videos. The light-receiving device is also called an image sensor, and the light-receiving device 100 can be, for example, a CMOS (Complementary Metal Oxide Semiconductor) image sensor, a CCD (Charge Coupled Device) image sensor, or an intelligent vision sensor. An intelligent vision sensor is a sensor that uses a stacked structure in which a pixel chip and a logic chip are stacked, and the logic chip is equipped with image analysis functions such as AI.

[0016] It is desirable that the camera 200 is pre-calibrated and that its position, orientation, and camera parameters are known.

[0017] In this embodiment, all cameras 200 mounted on the automobile 300 are equipped with the light-receiving device 100 according to this technology. However, it is not necessary for all cameras 200 to be equipped with the light-receiving device 100 according to this technology; rather, any one of the multiple cameras 200 mounted on the automobile 300 may be equipped with the light-receiving device 100 according to this technology.

[0018] The automobile 300 is comprised of an ECU 310, an IMU 320 (Inertial Measurement Unit), and a storage device 330.

[0019] The ECU 310 performs electronic control of the entire vehicle 300 and its individual parts. The vehicle 300 may be configured to have one ECU 310, or it may be configured to have a central ECU that controls the entire vehicle 300 and processes collected data, and multiple zone ECUs that control the individual parts of the vehicle 300.

[0020] The ECU 310 may be equipped with motion detection and human detection functions. The motion detection and human detection functions provided by the ECU 310 may be the same functions as those provided by the light receiving device 100, or they may be functions that detect in different ways.

[0021] The ECU310 has two operating modes related to parking surveillance: a power-saving mode and a monitoring mode. The power-saving mode reduces power consumption by limiting some functions and processes, and is an operating mode that waits for data from the camera 200 and various sensors. The monitoring mode is an operating mode that processes the data supplied from the camera 200 and various sensors, and performs functions such as recording determination, recording, notification, and alarm.

[0022] The IMU 320 is a sensor for detecting impacts to the vehicle 300 by detecting the movement of the vehicle 300. The IMU 320 outputs the detection result to the ECU 310. The vehicle 300 may be equipped with, in place of or in addition to, the IMU 320, an acceleration sensor, angular velocity sensor, gyroscope, etc., for two or three axes.

[0023] The storage 330 is a storage medium that stores various data (images, timestamps, location information, etc.) in accordance with the control of the ECU 310 when the light receiving device 100 detects an impact on the automobile 300 or a moving object around the automobile 300. The storage 330 is composed of, for example, a hard disk, an SSD (Solid State Drive), or flash memory.

[0024] Furthermore, the vehicle 300 may be equipped with a notification device and an alarm device for parking surveillance. The notification device, when the light receiving device 100 detects an impact on the vehicle 300 or a moving object around the vehicle 300, will notify the owner of the vehicle 300 or the management company, etc., in accordance with the control of the ECU 310. The notification method may be, for example, sending information to an electronic device such as a smartphone or smart key owned by the owner of the vehicle 300 or to the management company's system. The alarm device, when the light receiving device 100 detects an impact on the vehicle 300 or a moving object, will output an alarm in accordance with the control of the ECU 310.

[0025] Furthermore, the automobile 300 may be equipped with a positioning device, a door locking device, sonar, a distance sensor, a millimeter-wave radar, and the like for parking surveillance.

[0026] [Configuration of the light receiving device 100 and camera 200] Next, the configuration of the camera 200 and the light receiving device 100 will be described with reference to Figures 3 and 4.

[0027] The camera 200 is comprised of a light receiving device 100, a control unit 210, a storage unit 220, and a communication unit 230.

[0028] The control unit 210 consists of a CPU (Central Processing Unit), ROM (Read Only Memory), and RAM (Random Access Memory). The CPU functions as an arithmetic processing unit that performs various processing tasks and controls the camera 200 as a whole and its individual components. The CPU executes various processes according to the program stored in the ROM or the program loaded into the RAM from the storage unit 220. The RAM stores data and other information necessary for the CPU to perform various processes as appropriate.

[0029] The storage unit 220 is a storage medium such as a hard disk, SSD, or flash memory. Various applications, image data, and information are stored in the storage unit 220. However, if all images captured and output data from the camera 200 are saved to the storage 330 provided by the automobile 300, the camera 200 does not need to have a storage unit 220.

[0030] The communication unit 230 connects the camera 200 and the ECU 310 to enable communication. Communication methods include SerDes-based data transfer methods such as GMSL (Gigabit Multimedia Serial Link) and FPD-link (registered trademark), as well as methods such as MIPI (Mobile Industry Processor Interface) and I2C (Inter-Integrated Circuit). Alternatively, communication methods such as HDMI (High-Definition Multimedia Interface), USB (Universal Serial Bus), Wi-Fi, Bluetooth (registered trademark), Wireless LAN (Local Area Network), NFC (Near Field Communication), and Ethernet (registered trademark) may be used.

[0031] As will be explained in more detail later, when the light receiving device 100 detects an impact, or when it detects a moving object and a person, the camera 200 sends an activation trigger to the ECU 310 via processing by the control unit 210 to operate the ECU 310 in monitoring mode.

[0032] Furthermore, after the camera 200 sends a startup trigger to the ECU 310, that is, after the ECU 310 starts up in monitoring mode, it continuously captures images at predetermined time intervals and streams the images to the ECU 310.

[0033] The camera 200 may also include an input unit for user operation, a display unit for displaying images and GUI (Graphical User Interface), and drives and connection ports for connecting external storage media, etc.

[0034] The camera 200 may be equipped with known functions (such as subject detection, face detection, and scene detection) that can detect various types of information from images generated during shooting. These information detection functions can be implemented using machine learning or deep learning methods, template matching methods, matching methods based on the brightness distribution information of the subject, artificial intelligence methods, instance segmentation, and the like.

[0035] As shown in Figure 4, the light receiving device 100 operates in a camera 200 mounted on a vehicle 300 as a mobile device, and is composed of a pixel chip 110, a readout circuit 120, and a logic chip 130.

[0036] The pixel chip 110 has multiple pixels arranged in a two-dimensional array, each performing photoelectric conversion and outputting a pixel signal.

[0037] The readout circuit 120 reads out pixel signals from each pixel of the pixel chip 110 and supplies them to the logic chip 130.

[0038] The logic chip 130 is equipped with signal processing circuits, a DSP (Digital Signal Processor) memory, and other components, enabling it to perform image generation processing, ISP (Image Signal Processor) processing, AI (Artificial Intelligence) processing, and more.

[0039] In this technology, the logic chip 130 functions as a signal processing unit 140, an impact detection unit 150, a motion detection unit 160, and a human detection unit 170.

[0040] The signal processing unit 140 generates an image by applying predetermined signal processing, such as CDS (Correlated Double Sampling), AGC (Auto Gain Control), and A / D (Analog / Digital) conversion, to the pixel signals supplied from the readout circuit 120.

[0041] The impact detection unit 150 detects an impact on the automobile 300 based on the image generated by the light receiving device 100.

[0042] The motion detection unit 160 detects moving objects as subjects in the image generated by the light receiving device 100.

[0043] The person detection unit 170 detects a person as the subject in the image generated by the light receiving device 100.

[0044] In this technology, impact detection and motion detection are performed by the light receiving device 100 in order to achieve low power consumption and cost reduction, which are challenges when realizing high-precision parking surveillance.

[0045] The light receiving device 100 can generate a high-resolution image using pixel signals read out from all pixels included in the pixel chip 110. Furthermore, the light receiving device 100 can generate medium-resolution images and low-resolution images having lower resolution than the high-resolution image through reduction processing that combines pixel addition or thinning in the pixel chip 110 with processing in the signal processing unit 140.

[0046] [Processing in Monitoring System 10 and Light Receiving Device 100] With reference to FIG. 5, processing in the monitoring system 10 and the light receiving device 100, and the impact detection method executed by the light receiving device 100 will be described.

[0047] In step S101, the ECU 310 operates in a power saving mode. In the power saving mode, even if the ECU 310 is provided with monitoring functions such as impact detection, moving object detection, and human detection, these processes are not performed.

[0048] The ECU 310 may shift to the power saving mode in response to an input operation from a user, or may shift to the power saving mode triggered by events such as turning off of the engine of the automobile 300 or removal of the key of the automobile 300. Note that the ECU 310 may be in an off state instead of the power saving mode.

[0049] Furthermore, when the ECU 310 is in the power saving mode, the IMU 320 is turned off and does not operate. This can reduce power consumption. The IMU 320 may also be turned off in response to an input operation from a user, or may be turned off triggered by events such as turning off of the engine of the automobile 300 or removal of the key of the automobile 300.

[0050] Next, in step S102, the camera 200 starts operating in a monitoring mode that executes the processing according to the present technology. The camera 200 may transition to the monitoring mode in response to an input operation from a user, or may transition to the monitoring mode upon receiving a notification that the ECU 310 has shifted to the power saving mode, or the ECU 310 may control the camera 200 to cause it to transition to the monitoring mode.

[0051] In surveillance mode, the camera 200 does not stream images from the camera 200 to the ECU 310, and the light receiving device 100 performs surveillance detection processing, including impact detection, motion detection, and human detection, independently of the ECU 310.

[0052] When the camera starts operating in surveillance mode, in step S103, the light receiving device 100 generates a low-resolution image and supplies the low-resolution image to the impact detection unit 150 and the motion detection unit 160.

[0053] Performing impact detection and motion detection based on high-resolution images increases the processing load and power consumption. Therefore, as shown in Figure 6, the light receiving device 100 performs a reduction process such as pixel addition, and generates a low-resolution image by treating multiple pixels virtually as one pixel. In the example in Figure 6, the low-resolution image shown in Figure 6B is generated by treating multiple pixels, the number of pixels included in the frame indicated by the thick line in Figure 6A, as one pixel. When a low-resolution image is generated by area averaging, the effect of noise can be reduced, so detection accuracy can be improved even at night.

[0054] Next, in steps S104 and S105, based on the low-resolution image, the impact detection unit 150 detects impacts on the automobile 300, and the motion detection unit 160 detects motion objects around the automobile 300. By performing impact detection and motion detection based on the low-resolution image, the light receiving device 100 can reduce the processing load, speed up processing, and further reduce power consumption.

[0055] Impact detection and motion detection are performed by comparing the latest image generated by the light-receiving device 100 with identical pixels in past images generated before the latest image, and based on the number of pixels with differences. Past images are, for example, images taken immediately before the latest image among a group of images generated consecutively at predetermined time intervals. Note that there may be more than one past image.

[0056] For example, in the case of the past image shown in Figure 7A and the latest image shown in Figure 7B, since person H is moving to the right, when comparing identical pixels in the past and latest images, multiple pixels corresponding to person H, indicated by the shaded area in Figure 7C, are detected as pixels with differences.

[0057] On the other hand, if an object (such as another car) collides with the car 300, the camera 200 mounted on the car 300 will shake due to the impact of the collision, so it is thought that the entire image (all subjects present in the image) will move. Therefore, for example, if an object collides with the car 300 at the time the latest image shown in Figure 8B is captured, comparing the same pixels in the past image shown in Figure 8A and the latest image shown in Figure 8B, not only will multiple pixels corresponding to the moving object, person H, be detected as differences, but as shown in Figure 8C, pixels corresponding to other subjects (buildings in Figure 8) will also be detected as differences. Therefore, the number of pixels with differences will be greater when there is a collision with the car 300 than when there is no collision with the car 300 but there are moving objects around the car 300.

[0058] Therefore, the number of pixels with differences is compared using two thresholds. The thresholds include a first threshold for detecting impacts and a second threshold that is smaller than the first threshold and is used for detecting moving objects. For moving objects, the entire image does not move, so the second threshold for motion detection is smaller than the first threshold for impact detection. For impacts, the entire image moves due to the shaking of the camera 200, resulting in a larger number of pixels with differences, so the first threshold for impact detection is larger than the second threshold for motion detection.

[0059] If the number of pixels with differences is greater than or equal to the first threshold, the impact detection unit 150 detects an impact on the automobile 300. This is because the camera 200 is thought to be shaking due to the impact on the automobile 300, resulting in an increase in the number of pixels with differences.

[0060] If the number of pixels with differences is equal to or greater than the second threshold and less than the first threshold, the motion detection unit 160 detects a moving object as being present around the automobile 300. This is because, although there are pixels with differences due to the presence of a moving object, it is thought that there are not as many pixels with differences as when the automobile 300 is subjected to an impact.

[0061] If the number of pixels with differences is less than the second threshold, the motion detection unit 160 will not detect a moving object, and the impact detection unit 150 will not detect an impact. This is because, since the number of pixels with differences is small, it is assumed that there is no moving object and no impact has been applied to the automobile 300.

[0062] In this way, this technology performs impact detection and motion detection based on the number of pixels with differences in an image.

[0063] By performing impact detection based on the number of pixels with differences, the impact can be detected due to the shaking of the automobile 300 even if the object that collided with the automobile 300 is not visible in the image generated by the light receiving device 100.

[0064] Furthermore, by performing collision detection based on the number of pixels with differences, it is possible to detect any impact, such as when another vehicle collides with vehicle 300, or when a person strikes vehicle 300 with a blunt object.

[0065] Furthermore, preprocessing such as geometric transformation or brightness scaling may be performed on low-resolution images before impact detection and motion detection.

[0066] Let's return to the explanation of the flowchart in Figure 5. If the impact detection unit 150 detects an impact, the process proceeds to step S106 (Yes in step S104).

[0067] Next, in step S106, the camera 200 sends an activation trigger as a notification to the ECU 310.

[0068] Then, in step S107, the ECU 310, having received the activation trigger, starts up in monitoring mode. After the camera 200 transmits the activation trigger, that is, after the ECU 310 starts up in monitoring mode, the light receiver 100 continuously generates high-resolution images at predetermined time intervals, and the camera 200 starts streaming images to the ECU 310. Switching the operating mode of the light receiver 100 may be controlled by the camera 200, or the light receiver 100 may have a processing unit that controls its operation. These high-resolution images are, for example, HD (High Definition) images, 4K images, 8K images, etc., which are images with the highest resolution that the light receiver 100 can generate. The ECU 310 then continuously performs motion detection and human detection based on the high-resolution images, and records the high-resolution images.

[0069] If the ECU 310 is always operating in monitoring mode, power consumption will increase. Therefore, as shown in step S101, the ECU 310 first operates in power-saving mode, and when it receives a start trigger from the camera 200, it operates in monitoring mode in step S107, thereby reducing power consumption and realizing a low-power monitoring system 10.

[0070] The explanation returns to step S105. If the motion detection unit 160 detects motion in step S105, the process proceeds to step S108 (Yes in step S105).

[0071] Next, in step S108, the light receiving device 100 generates a medium-resolution image and supplies it to the human detection unit 170. This medium-resolution image has a higher resolution than the low-resolution image generated in step S103 and used for impact detection and motion detection, but a lower resolution than the high-resolution image streamed to the ECU 310 after the ECU 310 starts up in monitoring mode. By using a medium-resolution image for human detection, more accurate human detection can be achieved than when using a low-resolution image, and the processing load can be reduced and power consumption reduced compared to when using a high-resolution image.

[0072] Next, in step S109, the human detection unit 170 performs human detection based on the medium-resolution image. Human detection can be performed using methods such as machine learning or deep learning, including CNN (Convolutional Neural Network), template matching, matching based on the brightness distribution information of the subject, AI, or instance segmentation.

[0073] If the human detection unit 170 detects a person, the process proceeds to step S106 (Yes in step S107). Then, in step S106, the camera 200 sends an activation trigger to the ECU 310, and in step S107, the ECU 310 starts up in monitoring mode. Note that preprocessing such as geometric transformation or brightness scaling may be performed on the high-resolution image before human detection.

[0074] On the other hand, if the human detection unit 170 does not detect a person in step S109, the process proceeds to step S110 (No. of step S109). If a certain amount of time has elapsed since the detection process by the human detection unit 170, the process proceeds to step S103 (No. of step S110). If no person is detected for a certain amount of time, the light receiving device 100 generates a low-resolution image again and performs impact detection and motion detection based on the low-resolution image. If a certain amount of time has not elapsed, the process proceeds to step S108 (No. of step S110).

[0075] The processing in the monitoring system 10 and the light receiving device 100 in this technology is carried out as described above.

[0076] In Figure 6, a low-resolution image was generated by performing a reduction process, such as pixel addition, on all pixels of the light-receiving device 100, and treating multiple pixels virtually as one pixel. However, it is not always necessary to perform the reduction process on all pixels of the light-receiving device 100. As shown in Figure 9B, the original resolution of the light-receiving device 100 may be maintained without reduction processing on pixels in some areas. In Figure 9B, as an example, high resolution is maintained in nine specific areas without reduction processing. This allows for high-resolution impact detection and motion detection in certain areas, enabling the detection of minor collisions and fine movements. However, the number of areas that maintain high resolution is not limited to nine. Furthermore, the areas that maintain high resolution do not need to be evenly distributed and can be arranged in any way.

[0077] Since the camera 200 mounted on the automobile 300 is fixed in position, as shown in Figure 10A, when the automobile 300's body is the subject of the image captured by the camera 200, the same part of the automobile body will always be captured. In the example in Figure 10A, an image captured by camera 200A installed on the front of the automobile 300 is shown, and the hood of the automobile 300 is visible in the image. Therefore, as shown in Figure 10B, by increasing the resolution of pixels that include the boundary between the automobile 300's body and the background or ground, impact detection and motion detection can be performed with higher accuracy.

[0078] In the example in Figure 9, high resolution is maintained by not performing a reduction process on a specific region. However, the specific region can be made high-resolution by performing a predetermined process on regions other than the specific region. Specific processes include decimation, FD (Floating Diffusion) addition, filtering, addition, and pixel binning. This method also allows for high-resolution impact detection and motion detection in the specific region, enabling the detection of minor collisions and subtle movements.

[0079] Furthermore, as shown in Figure 11A, if only a portion of the image is high resolution and other areas are low resolution, or as shown in Figure 11B, if the entire image is low resolution, the light receiving device 100 may perform detection for parking surveillance based only on a specific area of ​​the image (shown by a thick line in Figures 11A and 11B). The specific area should be determined based on the parking position of the car, the position of the subject, etc. This also reduces the processing load and power consumption.

[0080] According to this technology, by mounting a camera 200 equipped with a light-receiving device 100 having a detection function for parking surveillance on a vehicle 300, it is not necessary to operate the power-hungry ECU 310 or IMU 320 for parking surveillance, thus realizing a parking surveillance system 10 with low power consumption.

[0081] Furthermore, by operating the ECU 310 in power-saving mode until the light-receiving device 100 detects an impact or moving object, and by not performing impact detection using the IMU 320, a monitoring system 10 with low power consumption can be realized.

[0082] A high-performance monitoring system 10 can be realized by operating the ECU 310 in monitoring mode after impact detection or detection of a moving object and a person by the light receiving device 100, and performing high-precision detection for parking monitoring using the ECU 310.

[0083] Since the only addition to an existing parking surveillance system is a camera 200 equipped with a light-receiving device 100 related to this technology, the cost required for the addition can be kept low, making it possible to realize the surveillance system 10 related to this technology at a low cost.

[0084] The light-receiving device 100 has an impact detection function that enables the construction of a more robust parking surveillance system. By simply providing the light-receiving device 100, a low-power, low-cost, and high-performance parking surveillance system can be provided.

[0085] Conventionally, an IMU 320 has been installed in the automobile 300 to detect impacts. However, when the IMU 320 is installed inside or near the center of the car, there is a problem in that it cannot detect minor impacts. With this technology, impact detection is performed by a light-receiving device 100 equipped in a camera 200 mounted on the body of the automobile 300, so even minor impacts can be detected.

[0086] Since the light receiving device 100 has a detection function for parking surveillance, detection is possible for each camera 200. Therefore, by mounting each of the multiple cameras 200 so as to photograph different directions from the vehicle 300, unlike when the IMU 320 is mounted approximately in the center of the vehicle 300, it is possible to detect the direction from which the impact came and the position on the vehicle 300 that the impact occurred, enabling more detailed impact detection.

[0087] <Modifications> Although embodiments of this technology have been described in detail above, this technology is not limited to the embodiments described above, and various modifications are possible based on the technical concept of this technology.

[0088] In addition to, or instead of, human detection, processes such as object detection or scene recognition may be performed to determine whether or not to record the image.

[0089] In this embodiment, a camera 200 equipped with a light-receiving device 100 is mounted on an automobile 300. However, this technology is not limited to automobiles 300 and can be applied to various mobile devices such as motorcycles, bicycles, drones, airplanes, personal mobility devices, ships, and robots. Furthermore, this technology can be applied to surveillance systems in buildings such as houses, apartments, and office buildings, in addition to mobile devices.

[0090] This technology can also be configured as follows: (1) A light-receiving device that operates in a camera mounted on a moving object and detects an impact on the moving object based on the result of comparing the generated image with a past image generated in the past of the said image. (2) The light-receiving device according to (1) that compares identical pixels in the said image and the past image and detects an impact on the moving object based on the number of pixels with differences. (3) The light-receiving device according to (2) that detects an impact on the moving object when the number of pixels with differences is equal to or greater than a first threshold. (4) The light-receiving device according to any one of (1) to (3) that detects moving objects around the moving object based on the result of comparing the said image and the past image. (5) The light-receiving device according to (4) that compares identical pixels in the said image and the past image and detects a moving object based on the number of pixels with differences. (6) The light-receiving device according to (5) that detects a moving object when the number of pixels with differences is less than a first threshold and the number of pixels with differences is equal to or greater than a second threshold which is smaller than the first threshold. (7) A light receiving device according to any one of (1) to (6) that generates a low-resolution image by treating multiple pixels constituting an image as a single pixel, and detects an impact on the moving object based on the low-resolution image. (8) A light receiving device according to (7) that reduces the resolution of areas other than a predetermined region of the image. (9) A light receiving device according to (4) that detects a person around the moving object based on the image. (10) A light receiving device according to (9) that detects a person when a moving object is detected. (11) A light receiving device according to (9) or (10) that detects a person based on an image with a higher resolution than the image used to detect the moving object. (12) A light receiving device according to any one of (1) to (12) in which the moving object is an automobile. (13) A light receiving device according to (12) that notifies the ECU (Electronic Control Unit) that controls the automobile when an impact is detected. (14) A light receiving device according to (12) that notifies the ECU that controls the automobile when a moving object is detected around the moving object, and further when a person is detected around the moving object.(15) An impact detection method for detecting an impact on a moving body based on the result of comparing a generated image with a past image generated in the past than the generated image, using a light-receiving device operating in a camera mounted on a moving body. (16) A camera equipped with a light-receiving device that operates in a camera mounted on a moving body and detects an impact on the moving body based on the result of comparing a generated image with a past image generated in the past than the generated image. (17) A surveillance system comprising an ECU that controls a moving body, and a camera equipped with a light-receiving device that detects an impact on the moving body based on the result of comparing a generated image with a past image generated in the past than the generated image.

[0091] 10... Surveillance system 100... Light receiving device 132... Impact detection unit 133... Motion detection unit 134... Human detection unit 200... Camera 300... Automobile 310... ECU

Claims

1. A light-receiving device that operates in a camera mounted on a moving object and detects an impact on the moving object based on a comparison between the generated image and past images generated prior to the said image.

2. The light receiving device according to claim 1, which compares identical pixels in the image and the past image and detects an impact on the moving object based on the number of pixels with differences.

3. The light receiving device according to claim 2, which detects an impact on the moving object when the number of pixels having a difference is equal to or greater than a first threshold.

4. The light receiving device according to claim 1, which detects moving objects around the moving object based on the result of comparing the aforementioned image with the aforementioned past image.

5. The light receiving device according to claim 4, which compares identical pixels in the aforementioned image with the aforementioned past image and detects the moving object based on the number of pixels having a difference.

6. The light receiving device according to claim 5, which detects the moving body when the number of pixels having the difference is less than a first threshold and the number of pixels having the difference is greater than or equal to a second threshold that is less than the first threshold.

7. The light receiving device according to claim 1, which generates a low-resolution image by treating multiple pixels constituting an image as a single pixel, and detects an impact on the moving object based on the low-resolution image.

8. The light receiving device according to claim 7, which reduces the resolution of areas other than a predetermined region of the image.

9. The light receiving device according to claim 4, which detects people around the moving object based on the image.

10. The light receiving device according to claim 9, which detects a person when it detects a moving object.

11. The light receiving device according to claim 9, which detects a person based on an image having a higher resolution than the image used to detect the moving object.

12. The light receiving device according to claim 1, wherein the moving body is an automobile.

13. The light receiving device according to claim 12, which, when it detects the impact, notifies the ECU (Electronic Control Unit) that controls the automobile.

14. The light receiving device according to claim 12, which detects a moving object around the moving object and further detects a person around the moving object, and notifies the ECU that controls the automobile.

15. An impact detection method for detecting an impact on a moving object based on the results of comparing a generated image with a past image generated prior to the generated image, in a light-receiving device operating in a camera mounted on a moving object.

16. A camera mounted on a moving object, equipped with a light-receiving device that operates and detects an impact on the moving object based on a comparison between the generated image and past images generated prior to the said image.

17. A surveillance system comprising: an ECU for controlling a moving object; and a camera equipped with a light-receiving device that detects impacts on the moving object based on a comparison between a generated image and past images generated prior to the generated image.

Images