JP6679152B1 - Accident analysis device, accident analysis method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique capable of promptly making a match between an accident situation and a past accident case. SOLUTION: Based on the acquired vehicle data and the video data, an acquisition unit that acquires vehicle data measured by a sensor included in the accident vehicle and video data captured by a camera included in the accident vehicle is used to determine the accident vehicle. Compare the analysis unit that analyzes the situation of the accident that occurred, the situation of the accident caused by the accident vehicle analyzed by the analysis unit, and the accident case data that associates the accident situation with information about past accident cases Thus, an accident analysis device having a search unit that searches for information on past accident cases corresponding to an accident situation caused by an accident vehicle is provided. [Selection diagram] Figure 2

Description

  The present invention relates to an accident analysis device, an accident analysis method and a program.

  At present, there is provided an in-vehicle camera called a drive recorder capable of constantly recording an image of a traveling direction and the like while a vehicle is traveling. By attaching a drive recorder to the car that the driver drives, the driver can use it for explaining the situation at the time of the accident and estimating the cause of the accident.

Japanese Patent Laid-Open No. 2018-65680

  When a traffic accident occurs, the insurance company accurately grasps the accident situation by organizing the accident situation, and calculates the negligence rate by comparing the accident situation with past accident cases. Conventionally, these operations have been performed manually, which is a time-consuming task.

  Therefore, it is an object of the present invention to provide a technique that enables a quick match between an accident situation and a past accident case.

  An accident analysis device according to one aspect of the present invention includes an acquisition unit that acquires vehicle data measured by a sensor included in an accident vehicle and image data captured by a camera included in the accident vehicle, and the acquired vehicle data and image data. The analysis unit that analyzes the situation of the accident caused by the accident vehicle, and the situation of the accident caused by the accident vehicle, which is analyzed by the analysis unit, and the information about the accident situation and past accident cases based on And a search unit that searches for information on past accident cases corresponding to the situation of the accident caused by the accident vehicle by comparing the accident case data with the accident case data.

  According to the present invention, it is possible to provide a technique capable of promptly making a match between an accident situation and a past accident case.

It is a figure which shows an example of the accident analysis system which concerns on this embodiment. It is a flow chart which shows an outline of a processing procedure when an accident analysis device analyzes the situation of an accident. It is a figure which shows an example of the information which shows the situation of the accident which an accident analysis device outputs. It is a figure showing an example of hardware constitutions of an accident analysis device. It is a figure showing an example of functional block composition of an accident analysis device. It is a figure for demonstrating the process sequence at the time of an accident analysis device performing peripheral object detection and coordinate specification. It is a figure for explaining a processing procedure when an accident analysis device presumes a relative physical relationship of a vehicle and a peripheral. It is a figure for explaining a processing procedure when an accident analysis device presumes an absolute position of a vehicle and a peripheral. It is a figure for demonstrating the process which estimates the absolute position of a vehicle. It is a figure which shows an example of the image which an accident analysis device produces. It is a figure for demonstrating the collision direction. It is a figure for demonstrating the process which determines a contact target. It is a figure for demonstrating a contact part. It is a figure for demonstrating the process which specifies the positional relationship of a pedestrian, a pedestrian crossing, and a safety zone. It is a figure for demonstrating the advancing direction of a vehicle and another vehicle. It is a figure which shows the specific content regarding the situation of an accident. It is a figure which shows an example of an accident case DB.

  Embodiments of the present invention will be described with reference to the accompanying drawings. In addition, in each of the drawings, those denoted by the same reference numerals have the same or similar configurations.

<System configuration>
FIG. 1 is a diagram showing an example of an accident analysis system according to this embodiment. The accident analysis system 1 shown in FIG. 1 includes an accident analysis device 10 and a terminal 20. The accident analysis device 10 and the terminal 20 are connected via a wireless or wired communication network N and can communicate with each other.

  The accident analysis device 10 is imaged by a camera D included in the vehicle C and vehicle data measured by a sensor included in a vehicle C (which may be referred to as “own vehicle” for convenience in the following description) that is an accident vehicle. It has a function of acquiring the video data and analyzing the situation of the accident caused by the vehicle C based on the acquired vehicle data and the video data. Further, the accident analysis device 10 associates the accident situation obtained by the analysis, the accident situation that occurred in the past, and information on the past accident case (hereinafter, referred to as “accident case”) with each other. It has a function of searching an accident case corresponding to the situation of the accident caused by the vehicle C by comparing with the database. Further, the accident analysis device 10 maps the position of the vehicle C and the position of another vehicle obtained by analyzing the condition of the accident on the map data, and thus an image showing the condition of the accident caused by the vehicle C. Has the function of generating. The image showing the situation of the accident caused by the vehicle C may be any image, but may be, for example, a bird's-eye view or a moving image.

  The accident analysis device 10 may be configured by one or a plurality of physical information processing devices or the like, or may be configured by using a virtual information processing device that operates on a hypervisor. However, it may be configured using a cloud server.

  The sensor included in the vehicle C may be, for example, a GPS receiver, a vehicle speed sensor, an acceleration sensor, a geomagnetic sensor, a throttle sensor, and / or a winker detection sensor. The vehicle data measured by the sensor provided in the vehicle C includes, for example, latitude and longitude data indicating the current position of the vehicle C, the vehicle speed of the vehicle C, the acceleration of the vehicle C, and the direction in which the vehicle C is facing (for example, north. May be the angle when 0 is 0 degree, the throttle opening (accelerator opening), the operating condition of the winker, and the like. The present invention is not limited to this, and the vehicle C may further include another sensor. In addition, the sensors listed above may be sensors included in the vehicle C or may be included in the camera D. For example, the GPS receiver, the acceleration sensor, and the geomagnetic sensor may be sensors included in the camera D instead of the vehicle C.

  The camera D is attached to the front part of the vehicle C or the windshield so that an image of at least the traveling direction of the vehicle C can be taken. The camera D may be a drive recorder. Further, the camera D may be capable of photographing the side direction and the rear of the vehicle C.

  The terminal 20 is a terminal operated by, for example, an operator of an insurance company, and displays the situation of an accident analyzed by the accident analysis apparatus 10, an accident case corresponding to the accident state, and an image generated by the accident analysis apparatus 10. As the terminal 20, any information processing apparatus such as a personal computer (PC), a notebook PC, a tablet terminal, and a smartphone can be used as long as the information processing apparatus has a display.

  FIG. 2 is a flowchart showing an outline of a processing procedure when the accident analysis device 10 analyzes the situation of an accident. First, the accident analysis device 10 acquires the vehicle data of the vehicle C in which the accident occurred and the image data at the time of the accident (S10, S11). The vehicle data and the video data are recorded in a non-transitory storage medium such as an SD card or a USB memory, and may be taken into the accident analysis device 10 by connecting the recording medium to the accident analysis device 10. Good. Alternatively, the communication function provided in the vehicle C or the camera D may be used to transmit the radio signal to the accident analysis apparatus 10.

  Next, the accident analysis device 10 analyzes the image data of the accident at each frame to analyze one or more peripheral objects (other vehicles, people, etc.) existing around the vehicle C shown in the image for each frame. , Signs, road structures, etc.), and further the coordinates indicating the position where the specified peripheral object appears in the image (S12). Then, the accident analysis device 10 estimates the relative positional relationship between the vehicle C and the peripheral objects existing around the vehicle C based on the coordinates of the specified peripheral object (S13).

  Subsequently, the accident analysis device 10 estimates the absolute position of the vehicle C using the position information of the vehicle C and / or the image data at the time of the accident acquired by the GPS sensor included in the vehicle C. Further, the accident analysis device 10 is based on the estimated absolute position of the vehicle C and the relative positional relationship between the vehicle C and surrounding objects existing around the vehicle C estimated in the processing procedure of step S13. Then, the absolute position of the surrounding objects existing around the vehicle C is estimated (S14).

  Next, the accident analysis device 10 maps the calculated absolute positions of the vehicle C and surrounding objects existing around the vehicle C on the map data, so that the image showing the situation of the accident (including the bird's-eye view and the moving image). Is generated (S15).

  Subsequently, the accident analysis device 10 determines that the vehicle C is another vehicle, an obstacle, or the like based on the longitudinal and lateral accelerations of the vehicle C at the moment of the collision, which are acquired by the acceleration sensor of the vehicle C. The direction in which the vehicle has collided with the vehicle (for example, the vehicle has collided from the front) is estimated (S16).

  Subsequently, the accident analysis device 10 detects the vehicle data of the vehicle C, the image data at the time of the accident, the absolute positions of the vehicle C and the surrounding objects existing around the vehicle C estimated in the processing procedure of step S14, and step S16. Information indicating the situation of the accident caused by the vehicle C is analyzed using the collision direction of the vehicle C estimated by the processing procedure and the map data (S17). The information includes a plurality of items (may be referred to as tags), and the situation of the accident is specified by the combination of each item. Subsequently, the accident analysis device 10 acquires the accident case corresponding to the situation of the accident caused by the vehicle C by searching the accident case database using each item as a key (S18).

  The order of the processing procedures described above can be arbitrarily changed as long as there is no contradiction in the processing. For example, the processing procedure of step S15 may be after any of the processing procedures of step S16, step S17, or step S18.

  FIG. 3 is a diagram showing an example of items included in the information indicating the situation of the accident. The information indicating the situation of the accident includes a plurality of items A to N as shown in FIG. It should be noted that “A: contact target (automobile, obstacle, etc.)” shown in FIG. 3 is information indicating a target contacted by the vehicle C. “B: Contact part” indicates a part (for example, the front surface, the right side surface, the front left bumper, etc.) where the vehicle C contacts the object. “C: Road type” indicates the type of road on which the vehicle C was traveling at the time of the accident (straight line, curve, intersection, T-shaped road, highway, etc.). “D: Signal color of own vehicle and other vehicle” indicates the signal color of the vehicle C side and the signal color of the other party in the accident at the intersection. "E: Positional relationship between pedestrians and pedestrian crossings / safety zones" refers to whether the pedestrian had an accident on the pedestrian crossing or the safety zone, or when the pedestrian was not on the pedestrian crossing or safety zone. Shows you.

  "F: Own vehicle / other vehicle traveling direction" means whether vehicle C and the other vehicle were traveling straight ahead, were changing lanes, were turning left, or were turning right at the time of the accident. Indicates whether or not you were turning right. The “G: lane position on the highway” indicates the lane in which the vehicle C was traveling (main line, overtaking lane) when an accident occurred on the highway. “H: Priority at intersection” indicates which of the vehicle C and the other vehicle should have given priority to the intersection and currency when an accident occurs at the intersection. “I: Temporary stop, presence / absence of signal violation” indicates whether the vehicle C and the other vehicle are in violation of stop or ignore the signal. “J: Whether speed limit is kept” indicates whether the vehicle C and the other vehicle were keeping the speed limit immediately before the accident. "K: speed before collision of own vehicle / other vehicle" indicates vehicle speed of vehicle C and other vehicle immediately before the accident. “L: Presence or absence of obstacle on road” indicates whether or not there is an obstacle on the road on which the vehicle C was traveling at the time of the collision. “M: Open / close door of collision target” indicates whether or not the door of the opponent vehicle is opened at the time of the collision. “N: Presence or absence of winker before collision” indicates whether or not the vehicle C operated the winker before the collision.

<Hardware configuration>
FIG. 4 is a diagram illustrating a hardware configuration example of the accident analysis device 10. The accident analysis device 10 includes a processor 11 such as a CPU (Central Processing Unit) and a GPU (Graphical processing unit), a storage device 12 such as a memory, a HDD (Hard Disk Drive) and / or an SSD (Solid State Drive), a wired or wireless device. It has a communication IF (Interface) 13 for performing communication, an input device 14 for receiving an input operation, and an output device 15 for outputting information. The input device 14 is, for example, a keyboard, a touch panel, a mouse, and / or a microphone. The output device 15 is, for example, a display and / or a speaker.

<Function block configuration>
FIG. 5 is a diagram showing a functional block configuration example of the accident analysis device 10. The accident analysis device 10 includes a storage unit 100, an acquisition unit 101, an analysis unit 102, a generation unit 103, a search unit 104, and an output unit 105. The storage unit 100 can be realized by using the storage device 12 included in the accident analysis device 10. Further, the acquisition unit 101, the analysis unit 102, the generation unit 103, the search unit 104, and the output unit 105 are provided by the processor 11 of the accident analysis device 10 executing the program stored in the storage device 12. Can be realized. Further, the program can be stored in a storage medium. The storage medium storing the program may be a non-transitory computer-readable storage medium (Non-transitory computer readable medium). The non-transitory storage medium is not particularly limited, but may be a storage medium such as a USB memory or a CD-ROM.

  The storage unit 100 stores an accident case DB (DataBase) that associates an accident situation with an accident case of an accident, and a map data DB. The accident case DB may include information indicating the rate of negligence. The accident case and the error rate corresponding to the situation of the accident that occurred may be searchable using the information indicating the situation of the accident as a key. The map data DB includes various data such as road data, road width, advancing direction, road type, traffic sign (temporary stop, entry prohibition, etc.), speed limit, signal position, number of intersection roads at intersections, and the like. The storage unit 100 may be realized by an external server that can communicate with the accident analysis device 10.

  The acquisition unit 101 has a function of acquiring vehicle data measured by a sensor included in the vehicle C (accident vehicle) and video data captured by a camera included in the vehicle C.

  The analysis unit 102 has a function of analyzing the situation of an accident caused by the vehicle C based on the vehicle data and the image data acquired by the acquisition unit 101. It should be noted that in the situation of the accident analyzed by the analysis unit 102, at least the color of the signal when the vehicle C passes through the intersection, the priority relationship when the vehicle C passes through the intersection, and the vehicle speed of the vehicle C exceeding the speed limit. Whether or not it may be included.

  Further, the vehicle data of the vehicle C includes at least information indicating the absolute position of the vehicle C measured by the GPS device included in the vehicle C, and the analysis unit 102 includes information indicating the absolute position of the vehicle C, The absolute position of the other vehicle is estimated based on the information indicating the relative positional relationship between the vehicle C and the other vehicle, which is obtained by analyzing the image of the other vehicle shown in the image data. Good. The analysis unit 102 may also estimate the absolute positions of the vehicle C and the other vehicle in time series.

  Further, the analysis unit 102 estimates the absolute position of the vehicle C by analyzing the image data, and the absolute position of the accident vehicle estimated by analyzing the image data and the absolute position of the vehicle C measured by the GPS device. The situation of the accident may be analyzed by regarding the absolute position obtained by averaging the position and the predetermined weight as the absolute position of the vehicle C.

  In addition, the analysis unit 102 compares the image size of the portion of the video data where the other vehicle is seen with the data indicating the correspondence relationship between the image size and the distance, and the relative position between the vehicle C and the other vehicle. The situation of the accident may be analyzed by estimating the distance between the vehicle C and the other vehicle, which is one of the information indicating the relationship.

  In addition, the analysis unit 102 compares the difference between the coordinates of the location of the other vehicle in the image data and the center coordinate of the image data with the data indicating the correspondence between the coordinates and the angle, and compares the difference between the vehicle C and the other. The situation of the accident may be analyzed by estimating the angle difference between the traveling direction of the vehicle C and the direction in which another vehicle is present, which is one of the pieces of information indicating the relative positional relationship with the vehicle. .

  The generation unit 103 has a function of mapping an absolute position of the vehicle C and an absolute position of another vehicle to map data to generate an image showing the situation of an accident caused by the vehicle C. The image may include a bird's-eye view or a moving image. For example, the generation unit 103 may generate a moving image in which images showing the situation of an accident are arranged in chronological order.

  The search unit 104 compares an accident situation caused by the vehicle C analyzed by the analysis unit 102 with an accident case DB (accident case data) that associates the accident situation with past accident cases, It has a function of searching for an accident case corresponding to the situation of an accident caused by the vehicle C. Further, the search unit 104 may search the fault rate of the vehicle C in the accident caused by the vehicle C as an accident case corresponding to the situation of the accident caused by the vehicle C.

  Further, the accident case may include one or more correction ratios (correction information) for correcting the error ratio. The search unit 104 may correct the fault rate of the accident vehicle according to the correction rate when the correction rate corresponding to the situation of the accident caused by the accident vehicle exists in the correction rates of 1 or more. The correction ratio will be described later.

  The output unit 105 displays information indicating the situation of the accident analyzed by the analysis unit 102, an accident case corresponding to the state of the accident searched by the search unit 104, and an image indicating the situation of the accident generated by the generation unit 103 on the terminal 20. It has a function to output to.

<Processing procedure>
Next, the processing procedure when the accident analysis device 10 analyzes the accident situation caused by the vehicle C will be described in detail along with the processing procedure described in FIG. In the following description, it is assumed that the accident analysis device 10 has already acquired the vehicle data and the video data from the vehicle C that has caused the accident. Further, it is assumed that the vehicle data and the video data each include time information or synchronization information. That is, in the present embodiment, when analyzing the video data at a certain time point, it is possible to perform the analysis using the vehicle data corresponding to the time point. Conversely, when analyzing the vehicle data at a certain time point, It is possible to perform analysis using the video data corresponding to the time point.

(Detection of surrounding objects and identification of coordinates)
FIG. 6 is a diagram for explaining a processing procedure when the accident analysis device 10 detects a peripheral object and specifies coordinates. The processing procedure corresponds to the processing procedure of step S12 of FIG.

  The analysis unit 102 analyzes each image obtained by decomposing the video data into each frame, and thereby analyzes the vehicles, people, bicycles, traffic signs, and traffic signals (signal colors and arrow signals in the image). Include), structures around the road (electric poles, street lights, guardrails, etc.), fallen objects, road surface letters, pedestrian crossings, lanes, and more. In addition, the analysis unit 102 specifies the coordinates of the region in which the peripheral object appears in the image for each of the specified one or more peripheral objects.

  The example of FIG. 6 shows an image example of the X-th frame in each image obtained by decomposing the video data for each frame. The X-axis indicates the number of pixels in the left-right direction with 0 at the left end (coordinates in the X direction), and the Y-axis shows the number of pixels in the vertical direction with 0 at the bottom end (coordinates in the Y direction). In the example of FIG. 6, a truck is shown in the back of the lane in which the vehicle C is traveling (the second lane from the left), and the passenger car is in the lane to the right of the lane in which the vehicle C is traveling. Is reflected. The analysis unit 102 can recognize a recognition target (for example, another vehicle, a person, a bicycle, a road sign, a traffic light, a structure around a road (a utility pole, a street light, a guardrail, or the like), a fallen object, a lane, or the like). A learned model that has learned the ability may be provided, and by inputting an image into the learned model, one or more types of peripheral objects and areas of the peripheral objects shown in the image may be specified. . Such processing can be realized by using existing technologies such as YOLO (Your Only Look Once) and Mask R-CNN (Mask Regional Convolutional Neural Network).

  In the example of A in FIG. 6, the analysis unit 102 displays the track in the area C1 (a rectangular area in which the points of X = 700 and Y = 300 are in the upper left and the points of X = 900 and Y = 300 are in the lower right). And that the passenger car is shown in the area C2 (the rectangular area in which the points of X = 1400 and Y = 250 are in the upper left and the points of X = 1750 and Y = 50 are in the lower right). . Further, similarly, it is specified that the electric poles are imaged in the areas C3 to C6 (coordinates not shown).

  The example of B of FIG. 6 shows an example of the result of the analysis unit 102 identifying the peripheral object for each frame.

  Further, the analysis unit 102 collates the analysis results of the images for each frame, and thereby, the vehicle is blinking the blinker, whether the door of the vehicle is opened and closed, and whether a person is standing or sitting. You may make it specify whether it is or is falling. For example, the analysis unit 102 may specify whether the blinker blinks or not by determining whether or not the color of the blinker portion of the vehicle recognized in each image changes periodically.

  Further, the analysis unit 102 may specify whether or not the vehicle door is opened / closed by determining a change in the vehicle door portion recognized in each image. In addition, the analysis unit 102 identifies whether the person is standing, sitting, or falling based on the ratio of the vertical length and the horizontal length of the area in which the person is shown. May be. For identifying the posture of a person, conventional techniques such as Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields can be used.

(Estimation of relative positional relationship between own vehicle and surrounding objects)
FIG. 7 is a diagram for explaining a processing procedure when the accident analysis device 10 estimates a relative positional relationship between the vehicle C and a peripheral object. The processing corresponds to the processing procedure of step S13 in FIG.

  The analysis unit 102 determines the relative positional relationship between the vehicle C and the peripheral objects existing around the vehicle C for each peripheral object based on the region in which one or more peripheral objects specified in the image are reflected. To estimate. More specifically, the analysis unit 102 estimates the distance from the vehicle C to the surrounding object (d1 and d2 in B of FIG. 7) based on the size of the area where the surrounding object is shown. In addition, the analysis unit 102 sets the traveling direction of the vehicle C (center direction of the image) as a reference (0 degree) based on the difference between the center coordinates of the area in which the peripheral object is captured and the center coordinates of the image in the left-right direction. In this case, the angles in the left-right direction (θ1 and θ2 in B of FIG. 7) at which the surrounding object exists are estimated.

[Example of calculating the distance to surrounding objects]
A specific example will be described with reference to FIG. First, the analysis unit 102 stores data indicating how many pixels occupy in the image when the camera D is used to photograph an object 1 m in length in the vertical direction (or horizontal direction) 1 m ahead. It has already been completed. Since the data varies depending on the angle of view (lens angle) of the camera, it may be stored in association with each model of the camera D. Further, depending on the model of the camera D, a wide-angle lens or a fisheye lens is often used. Therefore, the analysis unit 102 performs the distortion correction on the video data captured by the camera D according to the lens characteristics of the camera D, and then calculates the distance to the peripheral object and the peripheral object as described below. You may make it calculate the angle of a left-right direction. The distortion correction can be realized by using a conventional technique such as Cupic interpolation.

  In the example of FIG. 7A, when an object 1 m in length is placed 1 m ahead in the vertical direction and is imaged, it is assumed that the number of pixels is 100 pixels in the vertical direction. Similarly, when an object with a length of 1 m is placed 10 m ahead in the vertical direction (or the horizontal direction) and photographed, the number of pixels in the image is checked in advance. In the example of FIG. 7, it is assumed that when an object having a length of 1 m is placed 10 m ahead in the vertical direction and the image is captured, the number of pixels is 10 pixels in the vertical direction.

  Further, the size in the vertical direction (or the horizontal direction) is set in advance for each peripheral object. For example, in the case of a truck, the vertical length (height) may be set to 2 m, and in the case of a passenger car, the length may be set to 1.5 m.

  As described above, assuming that the number of pixels in the vertical direction of an object 1 m in length 1 m ahead is 100 pixels, and the number of pixels in the vertical direction of an object 1 m in length 10 m ahead is 10 pixels, the height is It can be calculated that when the 2m track is 1m ahead, the vertical pixel count is 200 pixels, and when the 2m high track is 10m ahead, the vertical pixel count is 20 pixels. More specifically, when the number of pixels is X and the distance is Y, the following formula is established.

Y = 200 ÷ X (Formula 1)
Next, the analysis unit 102 uses Equation 1 to calculate the distance between the vehicle C and the truck. In the example of A of FIG. 7, the vertical length of the region C1 is 50 ((400−300) / 2) = 50 pixels. Therefore, according to the equation 1, it can be calculated that Y = 200/50 = 4 m.

  If the height at which the camera D is mounted on the vehicle C (height from the ground) and the direction in which the camera D should be directed are fixed with high accuracy, the distance to the surrounding object can be determined by using only the Y-axis value. It is possible to specify. However, according to the processing procedure described above, even if the mounting position of the camera D is different depending on the driver, such as when renting a drive recorder, the distance between the vehicle C and surrounding objects can be specified. Will be possible.

[Example of calculating the angle in the left-right direction where peripheral objects exist]
A specific example will be described with reference to FIG. First, it is assumed that the analysis unit 102 stores data indicating the angle of view (lens angle) of the camera D. Since the data varies depending on the model of the camera D, it may be stored in association with each model of the camera D.

  Here, by dividing the angle of view of the camera D by the number of pixels in the horizontal direction of the screen, it becomes clear how many times one pixel corresponds to the angle of view. For example, when the total number of pixels in the horizontal direction of the screen is 2000 pixels and the angle of view of the camera D is 80 degrees, 200 pixels corresponds to 8 degrees. In other words, by calculating the difference in the number of pixels in the horizontal direction between the center position of the area in which the peripheral object appears in the image and the center position of the image, the peripheral object sets the traveling direction of the vehicle C to 0 degrees. In addition, it is possible to calculate the angle (the angle on the horizontal plane) in the left-right direction in which the peripheral object exists.

  For example, the center position in the X direction of the area where the track is shown in A of FIG. 7 is X = 800. Further, it is assumed that the angle of view of the camera D is 80 degrees and the total number of pixels in the horizontal direction of the screen is 2000 pixels. The center position of the image is X = 1000. Therefore, the center position of the area where the track is shown and the center position of the image are separated by 200 pixels (1000-800). As described above, since 200 pixels correspond to 8 degrees, the angle in the left-right direction between the vehicle C and the truck (θ1 in B of FIG. 7) can be calculated to be 8 degrees.

  It should be noted that with the calculation method described above, it is impossible to calculate the distance and the angle when a peripheral object extremely approaches the vehicle C and a part of the peripheral object goes out of the image. However, when a part of the surrounding object is outside the image, it can be estimated that the distance between the surrounding object and the vehicle C is within a certain distance. It is also possible to estimate the position of the peripheral object that has disappeared from the image based on the change in the position of the peripheral object in the images of the preceding and succeeding frames.

  For example, in the image of a certain frame X, a part of the peripheral object is out of the image, but in the image of the frame X-1 that is one frame before, the entire peripheral object is shown, and the vehicle C and Suppose the distance to the surroundings was 10 m. Further, in the image of the frame X-2 that is one frame before, it is assumed that the distance between the vehicle C and the surrounding object is 11 m. In this case, the distance between the vehicle C and the surrounding object in the image of the frame X can be estimated to be 9 m. When the surrounding object is a vehicle, the distance may not be estimated by recognizing the image of the entire vehicle, but the image may be recognized only in a part of the vehicle (for example, a license plate). Thereby, even when a part of the vehicle goes out of the image, it is possible to estimate the distance and the angle with the vehicle C as long as the license plate is shown in the image.

(Estimate the own vehicle position and the absolute position of surrounding objects)
FIG. 8 is a diagram for explaining a processing procedure when the accident analysis device 10 estimates the absolute positions of the vehicle C and surrounding objects. The process corresponds to the process procedure of step S14 of FIG.

  First, the analysis unit 102 estimates the absolute position of the vehicle C. The analysis unit 102 may estimate the absolute position of the vehicle C using the position information of the vehicle C included in the vehicle data of the vehicle C and acquired by the GPS sensor included in the vehicle C. Alternatively, the analysis unit 102 analyzes the image data at the time of the accident by using a conventional technique called Structure From Motion or Simultaneous Localization and Mapping (hereinafter, referred to as “SFM” for convenience), of the vehicle C. The absolute position of the vehicle C may be estimated more accurately by combining the absolute position and the absolute position of the vehicle C estimated by the GPS sensor.

[Estimation of absolute position of vehicle C based on SFM]
By using SFM, the feature points included in the image of each frame in the video data are extracted, the corresponding points (corresponding feature points) in the image of each frame are identified from the extracted feature points, and the identified corresponding points The movement of the camera can be reproduced by following the movement. The feature points can be extracted by using a conventional technique called SIFT feature amount, for example. Further, the identification of the corresponding points can be realized by using a conventional technique called FLANN (Fast Library for Approximate Nearest Neighbors), for example.

  Here, since the camera that captured the video data is the camera D mounted on the vehicle C, the reproduced camera movement can be regarded as the movement of the vehicle C. Since the vehicle data of the vehicle C includes the vehicle speed data, the distance traveled by the vehicle C between the frames can be estimated by matching the vehicle speed data and the frame rate in the video data. it can. For example, when the vehicle C is traveling at 60 km / h and the frame rate of video data is 15 frames per second, the moving distance of the vehicle C per frame is about 1 m.

  In addition, in order to obtain the movement of the vehicle C with respect to the surrounding environment, not the other vehicles running in parallel, the analysis unit 102 excludes the feature points of the moving peripheral object when extracting the feature points. , Feature points are extracted for peripheral objects that are fixedly installed. For example, the analysis unit 102 may extract feature points for traffic signs, traffic lights, and structures around the road (electric poles, street lights, guardrails, etc.). The type of the peripheral object can be grasped by identifying the learned model described above.

  For example, A in FIG. 8 shows the image data in the frame N, and B in FIG. 8 shows the image data in the frame N + 1. In FIGS. 8A and 8B, the areas C3 to C6 corresponding to the fixedly installed peripheral objects move rearward as the vehicle C moves, but the area C1 of the moving truck The position has hardly changed, and the position of the area C2 of the passenger vehicle traveling in the oncoming lane toward the vehicle C has largely changed.

  By using the SFM, the analysis unit 102, for example, moves the position of the vehicle C in the second frame 2 m forward and 1 m left as compared with the position of the first frame, and determines the position of the vehicle C in the third frame. It is possible to calculate information indicating the relative position of the vehicle C based on the first frame of the video data, such as moving 2.5 m forward and 0.5 m left as compared to the second frame. .

  Subsequently, the analysis unit 102 selects the position information corresponding to the time when the first frame of the video data was captured from the position information indicating the absolute position of the vehicle C acquired by the GPS sensor. The position information acquired by the GPS sensor includes time information, and the video data according to the present embodiment also includes time information indicating the time of recording. Therefore, the analysis unit 102 can select the GPS position information corresponding to the first frame of the video data by matching the time included in the video data with the time included in the position information.

  Then, the analysis unit 102 uses the position information of the GPS corresponding to the first frame of the video data to identify the absolute position of the vehicle C in the second and subsequent frames of the video data. As described above, the analysis unit 102 calculates the information indicating the relative movement between frames. In addition, the vehicle data in the present embodiment includes azimuth data indicating in which direction the front direction of the image is facing, and it is possible to grasp which direction the front is pointing from the azimuth data. . Therefore, the analysis unit 102 sets the selected GPS position information as the absolute position of the vehicle C in the first frame, and thus the vehicle C corresponding to the relative position of the vehicle C in the second frame (relative position obtained by SFM). The absolute position (latitude and longitude) of can be calculated. For example, when the latitude of the first frame acquired by the GPS sensor is 134.45 degrees and the longitude is 32.85 degrees, the latitude and the longitude are used as the starting points to move 2 m forward in the north direction and 1 m in the west direction. The latitude and longitude corresponding to the determined position become the absolute position of the vehicle C in the second frame. The analysis unit 102 repeats the processing for each frame to calculate the absolute position (latitude and longitude) of the vehicle C based on the SFM for all the frames.

[Estimation of absolute position of vehicle C]
As shown in FIG. 9, the analysis unit 102 estimates the absolute position of the vehicle C for each frame using the absolute position of the vehicle C based on the SFM and the absolute position of the vehicle C based on GPS. It is assumed that the points f11 to f16 shown on the left side of FIG. 9 are absolute positions of the vehicle C obtained by analyzing 1 frame to 6 frames of the video data, respectively. Further, points f21 to f26 shown in the center diagram of FIG. 9 are absolute positions of the vehicle C at times corresponding to 1 frame to 6 frames of the video data among the absolute positions of the vehicle C obtained by GPS. I assume. Points f31 to f36 shown on the right side of FIG. 9 indicate the estimated absolute positions of the vehicle C. As described above, the point f11 indicating the absolute position of the vehicle C corresponding to one frame of the video data is the same point as the point f21.

  The analysis unit 102 may estimate the absolute position of the vehicle C by simply averaging the absolute position of the vehicle C based on the SFM and the absolute position of the vehicle C by GPS, for example. For example, at a speed of 30 km / h or more, the absolute value of the vehicle C may be a value obtained by dividing the total value by 2. If the vehicle speed of the vehicle C at the points f13 and f23 is 30 km or more, the latitude of the absolute position of the vehicle C (point f33) is calculated by (latitude of the point f13 + latitude of the point f23) / 2. be able to. Similarly, the longitude of the absolute position (point f33) of the vehicle C can be calculated by (longitude of point f13 + longitude of point f23) / 2.

  In addition, the analysis unit 102 determines, as the absolute position of the vehicle C, the absolute position obtained by averaging the absolute position of the vehicle C based on the SFM and the absolute position of the vehicle C based on the GPS with a predetermined weight according to the vehicle speed of the vehicle C. It may be one that does. For example, for speeds of less than 30 km / h (for example, 5 km), the absolute position of the vehicle C based on SFM is considered to have higher accuracy, and the absolute position of the vehicle C based on SFM and the absolute position of the vehicle C based on GPS are calculated. For example, the absolute position averaged at a ratio of 4: 1 may be set as the absolute position of the vehicle C. If the vehicle speed of the vehicle C at the points f16 and f26 is 5 km or more, the latitude of the absolute position of the vehicle C (point f36) is (latitude of the point f16 × 4 + latitude of the point f26 × 1) / 5. Can be calculated by Similarly, the longitude of the absolute position (point f36) of the vehicle C can be calculated by (longitude of point f16 × 4 + longitude of point f26 × 1) / 5.

  The analysis unit 102 estimates the absolute position of the vehicle C for each frame by using the absolute position of the vehicle C based on the SFM that has been subjected to the noise removal process and the absolute position of the vehicle C that has been subjected to the noise removal by GPS. You may do it. For removing noise, for example, a Kalman filter may be used. The Kalman filter used for removing the noise of the absolute position of the vehicle C based on the SFM and the Kalman filter used for removing the noise of the absolute position of the vehicle C by GPS may be different Kalman filters. It is possible to further improve the accuracy of the estimated absolute position of the vehicle C.

[Estimate the absolute position of surrounding objects]
The analysis unit 102 calculates the absolute position of the peripheral object for each frame based on the estimated absolute position of the vehicle C for each frame and the relative positional relationship between the vehicle C and the peripheral object for each frame. As described above, the relative positional relationship between the vehicle C and the peripheral objects is, as shown in FIG. 7B, the distance between the vehicle C and the peripheral objects and the traveling direction of the vehicle C (the central direction of the image). It is indicated by the angle in the left-right direction where the peripheral object is present, where is the reference (0 degree). The analysis unit 102 sets the moving direction of the absolute position of the vehicle C, which is obtained by taking the difference between the absolute positions of the vehicle C for each frame, as the traveling direction of the vehicle C, and uses the estimated traveling direction as a reference to determine the surroundings. By calculating the latitude and longitude corresponding to the relative position of the object, the absolute position (latitude, longitude) of the surrounding object can be obtained.

(Generation of image showing accident situation)
The generation unit 103 generates an image indicating an accident situation by mapping the absolute position of the vehicle C and the absolute position of surrounding objects for each frame estimated by the above-described processing procedure on the road map. The process corresponds to the process procedure of step S15 of FIG.

  FIG. 10 is a diagram showing an example of an image generated by the accident analysis device 10. The Y axis and the X axis in FIG. 10 correspond to latitude and longitude, respectively. In the roads shown in FIGS. 10A to 10E, the center line indicates the median strip. The two lanes on the right side are lanes traveling from bottom to top, and the two lanes on the left side are lanes traveling from top to bottom. It is assumed that A to E in FIG. 10 correspond to the frame X, the frame X + 1, the frame X + 2, the frame X + 3, and the frame X + 4, respectively, but they are merely examples and the present invention is not limited thereto. The situation of the accident can be reproduced by mapping the absolute position of the vehicle C and the absolute position of surrounding objects for each frame on the road map. For example, at the time point A in FIG. 10, the passenger vehicle V2 is traveling from the oncoming lane of the vehicle C, and at the time point B in FIG. 10, the state in which the vehicle C and the passenger vehicle V2 are approaching each other and collided is illustrated. For the presence / absence of a collision and the display of the collision site, it is possible to use the result estimated by the processing procedure regarding the “estimation of the collision position” described later.

  In addition, in C to E of FIG. 10, the passenger vehicle V2 does not exist. This is because the passenger vehicle V2 is not captured by the camera D because the passenger vehicle V2 has moved to the rear of the vehicle C. If the camera D is also capable of capturing the rear of the vehicle, the motion data of the rear of the vehicle C is analyzed to estimate the movement of the passenger vehicle V2 after the collision and reflect it in the image showing the accident situation. It is possible.

(Estimation of collision direction)
The analysis unit 102 estimates the direction in which the vehicle C has collided with another vehicle, an obstacle, or the like, based on the acceleration generated in the vehicle C at the moment of the collision. The process corresponds to the process procedure of step S16 of FIG.

  More specifically, the analysis unit 102 determines the direction of the collision based on the acceleration pattern generated in the vehicle C at the moment of the collision. When a collision occurs, a negative acceleration is generated in the vehicle C in the direction in which the collision target exists, and the reaction recoil causes the acceleration in the direction opposite to the collision direction. That is, it is general that acceleration is generated in the vehicle C so as to reciprocate on the axis connecting the direction in which the collision object is present and the direction 180 degrees opposite to the direction.

  Therefore, the analysis unit 102 sums the output values from the acceleration sensor for each direction within a predetermined period from the time of the collision time, and estimates the direction with the larger total value as the collision direction. More specifically, as shown in A of FIG. 11, when the acceleration direction is 0 degree in the front direction of the vehicle C, the front direction (D1: 337.5 degrees to 22.5 degrees), the right front direction ( D2: 22.5 to 67.5), right (D3 direction: 67.5 to 112.5), right rear (D4: 112.5 to 157.5), rear (D5: 157). .5 degrees to 202.5 degrees), left rear (D6: 202.5 degrees to 247.5 degrees), left (D7: 247.5 degrees to 292.5 degrees), and left front (D8: 292.5 degrees). (Degrees to 337.5 degrees), and the direction with the larger total value is regarded as the collision direction. The acceleration measured by the acceleration sensor is an acceleration that follows the law of inertia. Therefore, at the time of a collision, a positive acceleration is measured in the collision direction.

  The analysis unit 102 may regard the time when the acceleration sensor detects the largest acceleration as the time at the time of the collision. When the camera D is a drive recorder, the analysis unit 102 may regard the time at the collision time recorded in the drive recorder as the time at the collision time.

  As shown in FIG. 11B, when the vehicle C contacts the vehicle V on the right front side, the acceleration sensor indicates, for example, time 1 (45 degree direction, 3.0 G), time 2 (47 degree direction, 1. 0G), time 3 (225 degree direction, 2.5G), time 4 (227 degree direction, 0.5G), and time 5 (40 degree direction, 2.0G) are assumed to be acquired in chronological order. . That is, the vehicle C collides with the vehicle V at the time point of time 1, so that the vehicle C is rapidly decelerated, and thereafter, the acceleration that reciprocates between the right front side and the left rear side is generated.

  In this case, the total acceleration corresponding to the front right (D2) is 3.0G + 1.0G + 2.0G = 6.0G, and the total acceleration corresponding to the rear left (D6) is 2.5G + 0.5G = 3. It is 0.0G. Therefore, the analysis unit 102 estimates the right front (D2) direction as the collision direction.

  Moreover, the analysis unit 102 may estimate the direction in which the vehicle C collides with a pedestrian by analyzing the video data. For example, in the above-described processing of detecting a surrounding object, it is estimated that the vehicle C and the pedestrian have collided in the front (D1) direction when the pedestrian is captured in a predetermined size or more in the image. You may When the vehicle C and the pedestrian collide, the acceleration detected by the acceleration sensor is smaller than when the vehicle C collides with another vehicle, and when the vehicle C and the pedestrian collide, a frontal collision may occur. Mostly. Therefore, when detecting the collision between the vehicle C and the pedestrian, it is preferable to analyze the video data.

(Output of information indicating the situation of the accident)
The analysis unit 102 outputs information indicating the situation of the accident caused by the vehicle C. The process corresponds to the process procedure of step S17 of FIG. As described above, the analysis unit 102 outputs each item shown in FIG. 3 as the information indicating the situation of the accident. Hereinafter, a processing procedure for specifying each item shown in FIG. 3 will be described in detail. In the following description, "another vehicle" means an opponent with which the vehicle C collides.

[A: Target of contact (cars, obstacles, etc.)]
At the time when the vehicle C collides, the analysis unit 102 sets the peripheral object closest to the vehicle C among the peripheral objects existing in a predetermined range (for example, a range of 45 degrees) around the collision direction of the vehicle C as the vehicle C. It is specified that it is the contacted object. More specifically, the analysis unit 102 causes the absolute position of the surrounding object at the time when the vehicle C collides (or the absolute position of the surrounding object obtained by searching the image of the frame closest to the time when the vehicle C collides). ), A peripheral object existing in a predetermined range centered on the collision direction of the vehicle C and at a position closest to the vehicle C is extracted, and the extracted peripheral object is an object contacting the vehicle C. Specify that there is.

  A specific example will be described with reference to FIG. In FIG. 12, it is assumed that the vehicle C has collided with a surrounding object at the front right. In the case of A of FIG. 12, the object at the position closest to the front right direction of the vehicle C is the passenger vehicle V1. Therefore, the analysis unit 102 identifies that the vehicle C has collided with the passenger vehicle V1. Similarly, in the case of B of FIG. 12, the object at the position closest to the front right direction of the vehicle C is the passenger vehicle V5. Therefore, the analysis unit 102 identifies that the vehicle C has collided with the passenger vehicle V5.

  If the collision direction of the vehicle C is a direction not captured by the camera D (that is, only the front of the vehicle C is captured), it is not possible to estimate the absolute value of the surrounding objects in that direction. Therefore, the contacted object cannot be specified. In this case, the analysis unit 102 may consider that some kind of vehicle has collided.

[B: contact area]
The analysis unit 102 identifies the contact portion according to the collision position of the vehicle C. Specifically, when the collision direction is the front surface (D1) in FIG. 11, the contact portion is the front bumper shown in FIG. When the collision direction is the front right (D2), the contact portion is the front right bumper shown in FIG. When the collision direction is right (D3), the contact portion is the right side surface shown in FIG. When the collision direction is the right rear (D4), the contact portion is the right rear bumper shown in FIG. When the collision direction is rearward (D5), the contact portion is the rear bumper shown in FIG. When the collision direction is rear left (D6), the contact portion is the rear left bumper shown in FIG. When the collision direction is left (D7), the contact portion is the left side surface shown in FIG. When the collision direction is the front left (D8), the contact portion is the front left bumper shown in FIG.

[C: Road type]
The analysis unit 102 identifies the road type of the road on which the vehicle C was traveling at the time of the collision by collating the absolute position of the vehicle C at the time when the vehicle C collided with the map data. The road type identified by the analysis unit 102 may include information such as straight roads, curves, intersections, and T-junctions, in addition to information on ordinary roads and highways.

[D: Signal color of own vehicle and other vehicles]
The analysis unit 102 specifies the signal color by using the color of the traffic light and the arrow signal that are specified by performing image analysis of the video data for each frame. For example, when the road type of the road on which the vehicle C was traveling at the time of the collision is an intersection, the analysis unit 102 determines that the color of the traffic signal specified by analyzing the video data is the highest at the time of the collision. It is an image of a frame at a close time, and the color of the traffic signal specified before the vehicle C enters the intersection is specified as the signal color of the vehicle C (either blue, yellow, or red). . Whether or not the vehicle C has entered the intersection can be determined by matching the absolute position of the vehicle C with the map data.

  In addition, it is assumed that the signal color on the crossing side at the intersection is not reflected in the video data, or is reflected only for a very short time, and the determination is difficult. Therefore, when the signal color on the vehicle C side is blue, the analysis unit 102 may specify that the signal color on the intersecting side is red. If another vehicle that collided with vehicle C is traveling on the road crossing the intersection, the traffic light of the other vehicle was red (that is, the other vehicle entered the intersection at a red traffic light). Will be specified. Similarly, when the signal color on the vehicle C side is red, the analysis unit 102 may specify that the signal color on the intersecting side is blue. If another vehicle that collided with the vehicle C is traveling on the road on the side crossing the intersection, the signal on the side of the other vehicle was green (that is, the other vehicle entered the intersection at a green light). Will be specified.

[E: Positional relationship between pedestrians and pedestrian crossings / safety zones]
When the target that has come into contact with the vehicle C is a pedestrian, the analysis unit 102 uses the absolute position of the pedestrian and the absolute position of the pedestrian crossing (or safety zone) that are specified by performing image analysis of the image data for each frame. , Specify the positional relationship between pedestrians and pedestrian crossings / safety zones. Note that the analysis unit 102 may acquire the absolute position of the pedestrian crossing from the map data.

  The analysis unit 102 obtains an absolute position of the vehicle C by analyzing an image of a frame before the time when the vehicle C collides with a pedestrian and finally a frame image showing a pedestrian crossing or a safety zone. Then, the point where the line connecting the contacting pedestrian and the absolute position of the pedestrian collides with the farthest side from the vehicle C in the area of the pedestrian crossing or the safety zone is calculated. Then, when the distance between the absolute position of the person and the point is within a predetermined distance (for example, 7 m), the analysis unit 102 specifies that the pedestrian who suffered the accident was on the pedestrian crossing or the safety zone. Further, when the distance between the absolute position of the person and the point exceeds a predetermined distance (for example, 7 m), the analysis unit 102 specifies that the pedestrian who suffered the accident was not on the pedestrian crossing or the safety zone.

  FIG. 14 is a diagram for explaining the process of identifying the positional relationship between a pedestrian and a pedestrian crossing / safety zone. For example, in FIG. 14, the analysis unit 102 causes the line L connecting the absolute position P1 of the vehicle C (the absolute position of the center of the vehicle C) and the absolute position P3 of the pedestrian to collide with the side of the pedestrian farthest from the vehicle C. When the point P2 is calculated and the distance between the pedestrian's absolute position P3 and the point P2 is a predetermined distance, it is considered that the pedestrian R was on a pedestrian crossing.

  In the example of A of FIG. 14, the pedestrian is on the pedestrian crossing, but in the example of B of FIG. 14, the pedestrian is not on the pedestrian crossing. However, in both examples, the analysis unit 102 considers the pedestrian to be on a pedestrian crossing if the distance between the absolute value P3 of the pedestrian and the point P2 is within a predetermined distance. In this case, it is not so important whether the pedestrian who was in the accident was walking on the pedestrian crossing surely, and it was judged that he was walking on the pedestrian crossing even if it was slightly protruding from the pedestrian crossing. This is because it is often done.

[F: Trajectory of own vehicle / other vehicle at intersection]
When the accident occurs near the intersection (when the absolute position of the vehicle C at the time when the vehicle C collides is within a predetermined range from the center position of the intersection), the analysis unit 102 determines the absolute position of the vehicle C for each frame. Are arranged on the map data of the intersection to specify the traveling locus of the vehicle C. In addition, the analysis unit 102 specifies the traveling locus of the other vehicle by arranging the absolute position of the other vehicle for each frame on the map data of the intersection. More specifically, the analysis unit 102 determines whether the vehicle C and the other vehicle have proceeded straight at the intersection, made a right turn, made a left turn, or made a quick right turn based on the magnitude and direction of the curvature of the traveling locus. Specify.

  FIG. 15 is a diagram for explaining traveling directions of the vehicle C and another vehicle. For example, when the curvature of the traveling locus in the predetermined area including the intersection is less than the predetermined value, the analysis unit 102 determines that the vehicle C or another vehicle has proceeded straight at the intersection, and the curvature of the traveling locus in the predetermined area. When is greater than or equal to a predetermined value, it may be determined that the vehicle C or another vehicle turns left or right at the intersection. In addition, the analysis unit 102 may determine that the vehicle makes a right turn when the traveling loci of the own vehicle and the other vehicle making a right turn pass inside the center of the intersection. Further, the analysis unit 102 may determine that the vehicle C and the other vehicle have changed lanes when the curvatures are reversed in a predetermined time (for example, 2 seconds).

"G: Lane position on the highway"
When the accident occurs on the highway (when the absolute position of the vehicle C at the time when the vehicle C collides is on the highway), the analysis unit 102 analyzes the video data for each frame, or By collating the absolute position of the vehicle C and the map data, it is specified whether the vehicle C is traveling on the main lane or the overtaking lane. For example, if the vehicle is traveling in the rightmost lane, it is specified that the vehicle is traveling in the overtaking lane, and if the vehicle is traveling in a lane other than the rightmost, it is specified that the vehicle is traveling in the main lane.

"H: Priority at intersections"
When the accident occurs near the intersection, the analysis unit 102 extracts the presence or absence of a stop sign at the intersection and the road width from the map data to determine whether the vehicle C can preferentially enter the intersection. Determine whether or not. More specifically, of the roads that cross the intersection, the vehicle running on the side without the stop sign is specified as the priority. Further, when there is a difference of twice or more in the width of the road crossing the intersection, the vehicle traveling on the wider road is specified as the priority. Also, if there is no stop sign and there is no difference in road width, it is determined that there is no priority relationship.

"I: Pause, signal ignore violation"
The analysis unit 102 matches the traveling locus of the vehicle C obtained by arranging the absolute positions of the vehicle C and the other vehicle for each frame with the map data so that the vehicle C and the other vehicle do not stop the stop sign. Specify whether the vehicle passed. In addition, when the vehicle C or the other vehicle passes through the stop sign without stopping, the analysis unit 102 determines that the vehicle speed of the vehicle C or the other vehicle at the time of the passage is a predetermined section before and after the stop sign (for example, 3 m before and after the stop sign). Etc.), if the speed is equal to or higher than a predetermined speed (for example, 5 km), the suspension violation is specified. On the other hand, when the vehicle speed of the vehicle C or the other vehicle at the time of passing is less than the predetermined speed, the analysis unit 102 specifies that the suspension violation has not occurred.

  Further, the analysis unit 102 detects whether or not the vehicle C has passed the signal by colliding the traveling locus of the vehicle C obtained by arranging the absolute positions of the vehicle C for each frame with the map data. When passing the signal, the analysis unit 102 obtains the color of the traffic light obtained by analyzing the image of the frame before the time when the signal passed, and the image of the frame in which the traffic light finally appears. To do. When the color of the signal is red, the analysis unit 102 identifies that the vehicle C is ignoring the signal when passing the signal.

"J: Whether or not the speed limit is observed"
The analysis unit 102 identifies the speed limit set on the road on which the vehicle C has traveled, by matching the traveling path of the vehicle C obtained by arranging the absolute positions of the vehicle C for each frame with the map data. . Further, the analysis unit 102 compares the vehicle speed data included in the vehicle data of the vehicle C with the vehicle speed and the speed limit at a timing before a predetermined time (for example, 5 seconds) before the time when the vehicle C collides. When the vehicle speed is higher than the speed limit, the analysis unit 102 identifies that the vehicle C is in speed violation. When the vehicle speed is less than or equal to the speed limit, the analysis unit 102 identifies that the vehicle C did not violate the speed.

"K: Speed before collision of own vehicle / other vehicle"
The analysis unit 102 identifies the speed of the vehicle C before the collision from the vehicle speed data included in the vehicle data of the vehicle C. The analysis unit 102 also calculates the vehicle speed of the other vehicle from the change in the absolute position of the other vehicle estimated for each frame of the video data. For example, when the moving distance of the other vehicle per frame is 1 m and the frame rate of the video data is 15 frames per second, it can be specified that the other vehicle was traveling at 60 km / h.

"L: Presence or absence of obstacles on the road"
The analysis unit 102 analyzes the image of the video data for each frame to identify the presence or absence of an obstacle on the road. The analysis unit 102 has an object that cannot be determined as a road on the lane road on which the vehicle C is traveling, the object is not a person, a car, a bicycle, or a motorcycle, and the absolute position between frames is If it is not moving, it may be considered as an obstacle.

"M: Opening and closing the door of another vehicle that collided"
The analysis unit 102 performs image analysis of the video data for each frame to identify whether the door of the other vehicle is open or closed.

"N: Presence or absence of winker action before collision"
The analysis unit 102 identifies whether or not the vehicle C was operating the winker before the collision from the winker operation information included in the vehicle data of the vehicle C. The analysis unit 102 also performs image analysis of the video data on a frame-by-frame basis to identify whether or not the winker was operating before the collision.

  FIG. 16 shows the contents specified by the processing procedure described above for each item included in the information indicating the situation of the accident.

(Search for corresponding accident case)
FIG. 17 is a diagram showing an example of the accident case DB. As shown in FIG. 17, the accident case includes “search conditions” for searching the accident case, “accident details” indicating the contents of the accident, and “negligence ratio (basic)” indicating a basic negligence ratio. It includes a "correction factor" that is a condition that requires correction of the negligence rate.

  In the example of FIG. 17, the target of the collision of the vehicle C is the vehicle (corresponding to “passenger car or truck” in item A of FIG. 16), and the vehicle C collides with the vehicle at the intersection (item C). "Corresponding to" in the intersection "), and it is an accident of a straight ahead vehicle and a right turn vehicle (in item F, vehicle C goes straight and another vehicle turns right, or another vehicle goes straight and vehicle C turns right) Accident example 1 when a straight-ahead vehicle has a red traffic light and a right-turning vehicle has a yellow color (in item D, vehicle C is red and another vehicle is yellow, or vehicle C is yellow and another vehicle is red) And that the ratio of negligence between a straight-ahead vehicle and a right-turn vehicle is 70:30.

  In addition, as a correction factor, when a straight ahead vehicle is a speed violation of 15 km or more (a speed violation of 15 km or more in items J and K), the negligence ratio of a straight ahead vehicle and a right turn vehicle is 75:25. Has been done. Similarly, when a straight ahead vehicle has a speed violation of 30 km or more (a speed violation of 30 km or more in items J and K), the negligence rate of a straight ahead vehicle and a right turn vehicle is shown to be 80:20. . Further, when the right-turn vehicle turns right without a turn signal (no turn signal in item N), it is shown that the negligence rate of a straight-ahead vehicle and a right-turn vehicle is 60:40.

  The correction factor may include a correction factor that is difficult to determine by the accident analysis device 10, such as "other significant negligence" in the fault rate.

  The search unit 104 searches for an accident case that matches the specified content by matching the content specified for each item of the information indicating the situation of the accident with the accident case DB. The search unit 104 also acquires the negligence rate corresponding to the searched accident case from the accident case DB. Further, the search unit 104 searches for the presence or absence of the corresponding correction element by matching the searched correction element of the accident case with each item indicating the situation of the accident. When the corresponding correction element exists, the search unit 104 corrects the negligence rate according to the correction rate of the corresponding correction element.

  Further, the output unit 105 outputs the accident content and the error rate of the accident case retrieved by the retrieval unit 104 to the terminal 20. If the correction element includes a correction element that is difficult to determine by the accident analysis device 10, the output unit 105 outputs the number of the accident case, the determined error rate, and the correction element that is difficult to determine. The wording may be output to the terminal 20.

  When searching the accident case DB, the search unit 104 may search not only the accident cases that meet all of the search conditions but also the accident cases that meet part of the search conditions. Further, when searching for an accident case that corresponds to part of the search conditions, the search unit 104 ranks the accident cases according to the number of items that do not match, and the output unit 105 outputs the accident content in order of rank. You may do it.

  Further, the accident cases may be ranked in descending order of the number of occurrences, and the search unit 104 may search the accident case DB in descending order of the rank.

<Summary>
According to the embodiment described above, the accident analysis device 10 analyzes the information indicating the situation of the accident based on the vehicle data and the video data acquired from the vehicle C. This made it possible to quickly match the accident situation with the accident case.

  The embodiments described above are for facilitating the understanding of the present invention and are not for limiting the interpretation of the present invention. The flowcharts and sequences described in the embodiment, the components included in the embodiment, and the arrangement, material, condition, shape, size, and the like of the embodiment are not limited to those illustrated but can be appropriately changed. Further, the configurations shown in different embodiments can be partially replaced or combined.

  10 ... Accident analysis device, 11 ... Processor, 12 ... Storage device, 13 ... Communication IF, 14 ... Input device, 15 ... Output device, 20 ... Terminal, 100 ... Storage unit, 101 ... Acquisition unit, 102 ... Analysis unit, 103 ... generation unit, 104 ... search unit, 105 ... output unit

Claims (16)

An acquisition unit that acquires vehicle data measured by a sensor included in the accident vehicle and image data captured by a camera included in the accident vehicle,
An analysis unit that analyzes the situation of an accident caused by the accident vehicle based on the acquired vehicle data and the video data,
Accidents caused by the accident vehicle by comparing the situation of the accident caused by the accident vehicle analyzed by the analysis unit and the accident case data in which the situation of the accident and information about past accident cases are associated with each other. A search unit that searches for information on past accident cases corresponding to the situation of
Have
The analysis unit is
By analyzing the video data, with respect to a plurality of frames including a predetermined frame, the relative position of the accident vehicle between the frames including the predetermined frame is calculated at one frame intervals,
Based on the position information measured by the GPS device included in the accident vehicle at the time corresponding to the time of the predetermined frame, and the relative position of the accident vehicle between the frames including the predetermined frame, among the plurality of frames Estimating a first absolute position of the accident vehicle for each of the frames other than the predetermined frame;
A first absolute position of the accident vehicle for each of the plurality of frames other than the predetermined frame, and the accident vehicle of the accident vehicle measured for each of the plurality of frames other than the predetermined frame by the GPS device. Information indicating the absolute position of the accident vehicle is estimated by averaging the second absolute position based on a predetermined weight.
Accident analysis device. An acquisition unit that acquires vehicle data measured by a sensor included in the accident vehicle and image data captured by a camera included in the accident vehicle,
An analysis unit that analyzes the situation of an accident caused by the accident vehicle based on the acquired vehicle data and the video data,
Accidents caused by the accident vehicle by comparing the situation of the accident caused by the accident vehicle analyzed by the analysis unit and the accident case data in which the situation of the accident and information about past accident cases are associated with each other. A search unit that searches for information on past accident cases corresponding to the situation of
Have
The analysis unit is
Based on information indicating the absolute position of the accident vehicle and information indicating the relative positional relationship between the accident vehicle and the other vehicle, which is obtained by analyzing the image of the other vehicle shown in the image data. And estimate the absolute position of the other vehicle,
The accident vehicle and the other vehicle are compared with each other by comparing the difference between the coordinates of the location of the other vehicle in the image data and the center coordinate of the image data with the data indicating the correspondence between the coordinate and the angle. An accident analysis device that estimates an angular difference between a traveling direction of the accident vehicle and a direction in which the other vehicle exists, out of information indicating a relative positional relationship with the vehicle. The analysis unit obtains information indicating the absolute position of the accident vehicle and a relative positional relationship between the accident vehicle and the other vehicle, which is obtained by analyzing an image of the other vehicle shown in the image data. Estimating the absolute position of the other vehicle based on the information shown,
Accident analysis apparatus according to claim 1 or 2. By mapping the absolute position of the accident vehicle and the absolute position of the other vehicle to map data, a generation unit that generates an image showing the situation of the accident caused by the accident vehicle,
The accident analysis device according to claim 3 . The analysis unit compares the image size of a portion of the video data in which the other vehicle is seen with data indicating the correspondence relationship between the image size and the distance, and compares the accident vehicle and the other vehicle with each other. Analyzing the situation of the accident by estimating the distance between the accident vehicle and the other vehicle showing a different positional relationship,
The accident analysis device according to claim 3 or 4 . The analysis unit compares the difference between the coordinates of the location of the other vehicle in the video data and the center coordinates of the video data with the data indicating the correspondence between the coordinates and the angle. Analyzing the situation of the accident by estimating the angle difference between the traveling direction of the accident vehicle and the direction in which the other vehicle is present, which indicates the relative positional relationship between the accident vehicle and the other vehicle.
Accident analysis device according to any one of claims 3-5. By mapping the absolute position of the accident vehicle and the absolute position of the other vehicle to map data, a generation unit that generates an image showing the situation of the accident caused by the accident vehicle,
The accident analysis device according to claim 2. The analysis unit is
By analyzing the video data, with respect to a plurality of frames including a predetermined frame, the relative position of the accident vehicle between the frames including the predetermined frame is calculated at one frame intervals ,
Based on the position information measured by the GPS device included in the accident vehicle at the time corresponding to the time of the predetermined frame, and the relative position of the accident vehicle between the frames including the predetermined frame, among the plurality of frames Estimating a first absolute position of the accident vehicle for each of the frames other than the predetermined frame ;
A first absolute position of the accident vehicle for each of the plurality of frames other than the predetermined frame, and the accident vehicle of the accident vehicle measured for each of the plurality of frames other than the predetermined frame by the GPS device. Information indicating the absolute position of the accident vehicle is estimated by averaging the second absolute position based on a predetermined weight.
The accident analysis device according to claim 7 . The analysis unit compares the image size of a portion of the video data in which the other vehicle is seen with data indicating the correspondence relationship between the image size and the distance, and compares the accident vehicle and the other vehicle with each other. Analyzing the situation of the accident by estimating the distance between the accident vehicle and the other vehicle showing a different positional relationship,
The accident analysis device according to claim 7 or 8 . The information on the past accident cases includes the fault rate,
The search unit searches for a fault rate of the accident vehicle in an accident caused by the accident vehicle, as information about an accident case corresponding to an accident situation caused by the accident vehicle.
Accident analysis device according to any one of claims 1-9. The information about the past accident case includes one or more correction information for correcting the error rate,
If the correction information corresponding to the situation of the accident caused by the accident vehicle exists in the one or more correction information, the search unit corrects the fault rate of the accident vehicle according to the correction information.
The accident analysis device according to claim 10 . The situation of the accident includes at least the color of a signal when the accident vehicle passes an intersection, a priority relationship when passing the intersection, and whether or not the vehicle speed of the accident vehicle exceeds a speed limit. ,
Accident analysis device according to any one of claims 1 to 11. An accident analysis method performed by an accident analysis device,
Acquiring vehicle data measured by a sensor included in the accident vehicle and image data captured by a camera included in the accident vehicle,
Analyzing the situation of an accident caused by the accident vehicle based on the acquired vehicle data and the video data;
The accident situation caused by the accident vehicle analyzed by the analyzing step is compared with accident case data in which the accident situation and information about past accident cases are associated with each other. Searching for information about past accident cases corresponding to the situation of the accident,
Including,
The analyzing step includes
By analyzing the video data, with respect to a plurality of frames including a predetermined frame, the relative position of the accident vehicle between the frames including the predetermined frame is calculated at one frame intervals,
Based on the position information measured by the GPS device included in the accident vehicle at the time corresponding to the time of the predetermined frame, and the relative position of the accident vehicle between the frames including the predetermined frame, among the plurality of frames Estimating a first absolute position of the accident vehicle for each of the frames other than the predetermined frame;
A first absolute position of the accident vehicle for each of the plurality of frames other than the predetermined frame, and the accident vehicle of the accident vehicle measured for each of the plurality of frames other than the predetermined frame by the GPS device. Information indicating the absolute position of the accident vehicle is estimated by averaging the second absolute position based on a predetermined weight.
Accident analysis method. An accident analysis method performed by an accident analysis device,
Acquiring vehicle data measured by a sensor included in the accident vehicle and image data captured by a camera included in the accident vehicle,
Analyzing the situation of an accident caused by the accident vehicle based on the acquired vehicle data and the video data;
The accident situation caused by the accident vehicle analyzed by the analyzing step is compared with accident case data in which the accident situation and information about past accident cases are associated with each other. Searching for information about past accident cases corresponding to the situation of the accident,
Including,
The analyzing step includes
Based on information indicating the absolute position of the accident vehicle and information indicating the relative positional relationship between the accident vehicle and the other vehicle, which is obtained by analyzing the image of the other vehicle shown in the image data. And estimate the absolute position of the other vehicle,
The accident vehicle and the other vehicle are compared with each other by comparing the difference between the coordinates of the location of the other vehicle in the image data and the center coordinate of the image data with the data indicating the correspondence between the coordinate and the angle. An accident analysis method for estimating an angular difference between a traveling direction of the accident vehicle and a direction in which the other vehicle exists, out of information indicating a relative positional relationship with the vehicle. On the computer,
Acquiring vehicle data measured by a sensor included in the accident vehicle and image data captured by a camera included in the accident vehicle,
Analyzing the situation of an accident caused by the accident vehicle based on the acquired vehicle data and the video data;
The accident situation caused by the accident vehicle analyzed by the analyzing step is compared with the accident case data in which the accident situation and the information about the past accident case are associated with each other. Searching for information about past accident cases corresponding to the situation of the accident,
Run
The analyzing step includes
By analyzing the video data, with respect to a plurality of frames including a predetermined frame, the relative position of the accident vehicle between the frames including the predetermined frame is calculated at one frame intervals,
Based on the position information measured by the GPS device included in the accident vehicle at a time corresponding to the time of the predetermined frame, and the relative position of the accident vehicle between the frames including the predetermined frame, among the plurality of frames, Estimating a first absolute position of the accident vehicle for each of the frames other than the predetermined frame;
A first absolute position of the accident vehicle for each of the frames other than the predetermined frame among the plurality of frames; and for the accident vehicle measured by the GPS device for each of the frames other than the predetermined frame of the plurality of frames. Information indicating the absolute position of the accident vehicle is estimated by averaging the second absolute position based on a predetermined weight.
program. On the computer,
Acquiring vehicle data measured by a sensor included in the accident vehicle and image data captured by a camera included in the accident vehicle,
Analyzing the situation of an accident caused by the accident vehicle based on the acquired vehicle data and the video data;
The accident situation caused by the accident vehicle analyzed by the analyzing step is compared with accident case data in which the accident situation and information about past accident cases are associated with each other. Searching for information about past accident cases corresponding to the situation of the accident,
Run
The analyzing step includes
Based on information indicating the absolute position of the accident vehicle and information indicating the relative positional relationship between the accident vehicle and the other vehicle, which is obtained by analyzing the image of the other vehicle shown in the image data. And estimate the absolute position of the other vehicle,
The accident vehicle and the other vehicle are compared with each other by comparing the difference between the coordinates of the location where the other vehicle is shown in the image data and the center coordinate of the image data with the data indicating the correspondence between the coordinate and the angle. A program for estimating an angle difference between a traveling direction of the accident vehicle and a direction in which the other vehicle exists, out of information indicating a relative positional relationship with the vehicle.