JP5275059B2 - Vehicle detection device, vehicle detection system, vehicle detection method for vehicle detection device, and vehicle detection program - Google Patents

Description

  The present invention provides, for example, a vehicle detection device, a vehicle detection system, and a vehicle of the vehicle detection device that detect a traveling vehicle including a vehicle hidden behind another vehicle using an image sensor in a DSSS (safety driving support system). The present invention relates to a detection method and a vehicle detection program.

  A vehicle is detected using an image sensor in DSSS, and the vehicle type, vehicle speed, vehicle position (distance from the stop line of the intersection to the point where the vehicle is located), and vehicle ID (given for each vehicle) for each detected vehicle There is a way to identify

  In the current system, once the detected vehicle is hidden by the shadow due to some factor (such as another vehicle interrupting in front), the vehicle detection is interrupted, and the vehicle is not detected, and there is no actual vehicle. It will be. In this case, the vehicle detection information is ineligible as information provided to the oncoming vehicle. As a result, it prevents the improvement of safe driving support functions such as preventing traffic accidents.

JP-A-11-284997

  Conventional vehicle detection using image sensors detects ordinary vehicles, small vehicles and two-wheeled vehicles that are hidden behind large vehicles, small vehicles and two-wheeled vehicles, and ordinary vehicles that are hidden beside large vehicles, and small vehicles and two-wheeled vehicles. There is a problem that it cannot be done.

  Furthermore, in conventional vehicle detection using an image sensor, ordinary vehicles, small vehicles, and two-wheeled vehicles that follow (or run parallel) behind a large vehicle are detected by changing the lane of the large vehicle. However, there is a problem that it is not treated as a detection vehicle and remains undetected.

  Furthermore, in conventional vehicle detection using an image sensor, ordinary vehicles, small vehicles, and two-wheeled vehicles that have not been detected by the image sensor because they are surrounded by a large vehicle are in a state where they are detected by their own lane change. However, there is a problem that it is not treated as a detection vehicle and remains undetected.

  An object of the present invention is to make it possible to detect, for example, a vehicle hidden behind another vehicle when viewed from an image sensor.

  The vehicle detection device of the present invention inputs a plurality of road images showing a road, and detects the vehicle shown in the road image and the position of the vehicle based on each of the inputted road images using a CPU. A road image processing unit that outputs information including a vehicle ID for identifying the detected vehicle and a vehicle position indicating the detected position of the vehicle as vehicle information for each vehicle, and each of the road image processing units output by the road image processing unit A vehicle information setting unit that sets vehicle information in a predetermined vehicle information table, and a vehicle information output that outputs the vehicle information set in the vehicle information table to the output device as information on a vehicle traveling on the road. A part.

  According to the present invention, for example, a vehicle hidden behind another vehicle can be detected by repeatedly updating the vehicle information table.

1 is a configuration diagram of a vehicle detection tracking system 100 according to Embodiment 1. FIG. FIG. 3 is a diagram showing imaging ranges of the first image sensor 110 and the second image sensor 111 in the first embodiment. FIG. 3 is a diagram showing imaging ranges of the first image sensor 110 and the second image sensor 111 in the first embodiment. 2 is a functional configuration diagram of an image sensor data processing device 200 and an information processing determination device 300 according to Embodiment 1. FIG. 2 is a diagram illustrating an example of hardware resources of an image sensor data processing device 200 and an information processing determination device 300 according to Embodiment 1. FIG. 3 is a flowchart illustrating a vehicle detection tracking method of the vehicle detection tracking system 100 according to the first embodiment. FIG. 4 shows a vehicle information table 391 in the first embodiment. FIG. 4 shows a mesh map in the first embodiment. The figure which shows the record of the detection-missing vehicle in the vehicle information table 391 in Embodiment 1, and the record of a detection range passage vehicle. FIG. 3 is a diagram showing a tracking example of the vehicle in the first embodiment.

Embodiment 1 FIG.
A vehicle detection system for detecting a traveling vehicle using an image sensor will be described.

FIG. 1 is a configuration diagram of a vehicle detection tracking system 100 according to the first embodiment.
A vehicle detection tracking system 100 (an example of a vehicle detection system) in the first embodiment will be described below with reference to FIG.

  The vehicle detection tracking system 100 includes a first image sensor 110, a second image sensor 111, an image sensor data processing device 200, and an information processing determination device 300 (an example of a vehicle detection device).

  If the vehicle is equipped with an on-board device for DSRC (Dedicated Short Range Communication), it communicates with the on-board device to acquire vehicle information (vehicle type, speed, etc.) and communicates with the on-vehicle device (not shown). The travel position of the vehicle (lane, distance from the intersection, etc.) can be specified based on the installation location of the vehicle. However, when the vehicle-mounted device is not mounted on the vehicle, the vehicle information and the travel position cannot be obtained.

The vehicle detection and tracking system 100 includes image sensors (first image sensor 110 and second image sensor 111) (also referred to as a camera, a video camera, and an imaging device), and an image of the road 101 (hereinafter referred to as “the image sensor”). Based on "road image"), it detects both vehicles with onboard equipment and vehicles without onboard equipment, and acquires vehicle information and travel location (hereinafter referred to as "vehicle information"). To do.
In addition, the vehicle detection and tracking system 100 includes a first image sensor 110 and a second image sensor 111 that capture images of the road 101 from different directions in order to reduce the number of vehicles hidden behind other vehicles and not appearing in the road image.
Furthermore, the vehicle detection and tracking system 100 detects a vehicle that is hidden by another vehicle and is not shown in either the road image of the first image sensor 110 or the road image of the second image sensor 111 based on the past road image. At the same time, the vehicle information of the vehicle is specified based on the past road image.
As a result, the vehicle detection and tracking system 100 detects a detection omission of a vehicle traveling in a specific range of the road 101 shown in the road image (here, a certain range before the intersection) (hereinafter referred to as “detection range”). Reduce omissions in vehicle information acquisition.

  For example, the vehicle information is provided to a vehicle equipped with a vehicle-mounted device in DSSS (Safety Driving Support System), and is used to prevent a traffic accident.

The first image sensor 110 is installed above the intersection in the direction opposite to the traveling direction of the vehicle, images the road 101, and acquires a road image in which each vehicle traveling toward the intersection is captured from the front.
The road image acquired by the first image sensor 110 includes the large vehicle 121 and the ordinary vehicle 124 running in parallel with the large vehicle 121.
Hereinafter, the imaging range (detection range) of the first image sensor 110 is indicated by a solid line.

The second image sensor 111 is installed toward the road 101 above the road 101 near the intersection. The second image sensor 111 captures the road 101 and acquires a road image of each vehicle taken from the side. The second image sensor 111 includes a wide-angle lens, and the imaging range is substantially the same as the first image sensor 110.
The road image acquired by the second image sensor 111 includes the large vehicle 121 and the ordinary vehicle 122 and the two-wheeled vehicle 123 that are traveling behind the large vehicle 121.
Hereinafter, the imaging range (detection range) of the second image sensor 111 is indicated by a dotted line.

  The first image sensor 110 and the second image sensor 111 capture the road 101 while mutually compensating for the blind spots hidden in the vehicle. Thereby, the possibility that the vehicle located at the blind spot appears in at least one of the road image of the first image sensor 110 and the road image of the second image sensor 111 increases.

2 and 3 are diagrams illustrating imaging ranges of the first image sensor 110 and the second image sensor 111 according to the first embodiment.
In FIG. 2, the ordinary vehicle 122 and the two-wheeled vehicle 123 traveling behind the large vehicle 121 are located within the imaging range (solid line) of the first image sensor 110, but in the line-of-sight direction of the first image sensor 110. The large vehicle 121 is shielded. For this reason, the ordinary vehicle 122 and the two-wheeled vehicle 123 are not shown in the road image of the first image sensor 110. However, the ordinary vehicle 122 and the two-wheeled vehicle 123 are not shielded by the large vehicle 121 in the line-of-sight direction of the second image sensor 111. For this reason, the road image of the second image sensor 111 shows the ordinary vehicle 122 and the two-wheeled vehicle 123 together with the large vehicle 121.
In FIG. 3, the ordinary vehicle 124 running in parallel with the large vehicle 121 is located within the imaging range (dotted line) of the second image sensor 111, but the large vehicle 121 is in the line-of-sight direction of the second image sensor 111. It will be shielded. For this reason, the ordinary vehicle 124 does not appear in the road image of the second image sensor 111. However, the ordinary vehicle 124 is not shielded by the large vehicle 121 in the line-of-sight direction of the first image sensor 110. For this reason, the ordinary vehicle 124 is shown together with the large vehicle 121 in the road image of the first image sensor 110.

  Returning to FIG. 1, the description of the configuration of the vehicle detection and tracking system 100 will be continued.

  The first image sensor 110 and the second image sensor 111 acquire a predetermined number of road images per unit time (for example, 30 images per second) by continuous shooting (also referred to as moving image shooting), and obtain the acquired road images. The images are sequentially output to the image sensor data processing device 200.

  The image sensor data processing device 200 detects a vehicle in the road image of the first image sensor 110 and the second image sensor 111, generates vehicle information 211 of the detected vehicle, and processes the generated vehicle information 211 as information processing. The data is output to the determination device 300.

  Based on each vehicle information 211 generated by the image sensor data processing device 200, the information processing determination device 300 includes information about each vehicle currently traveling in the detection range (a “vehicle information table 391” and “mesh” described later). Map data 392 ").

  In addition to the first image sensor 110 and the second image sensor 111, the second image sensor is sandwiched between the third image sensor and the road 101 that are located downstream of the intersection (on the side far from the intersection) and opposed to the first image sensor 110. There may be a fourth image sensor installed facing the image sensor 111.

FIG. 4 is a functional configuration diagram of the image sensor data processing device 200 and the information processing determination device 300 according to the first embodiment.
The functional configurations of the image sensor data processing device 200 and the information processing determination device 300 according to Embodiment 1 will be described below with reference to FIG.

  The image sensor data processing device 200 includes an image sensor data processing unit 210 (an example of a road image processing unit) and a road image storage unit 290.

The image sensor data processing unit 210 inputs road images sequentially output from the first image sensor 110 and the second image sensor 111.
The image sensor data processing unit 210 performs image processing on the input road image using a CPU (Central Processing Unit), and detects a vehicle shown in the road image.
The image sensor data processing unit 210 identifies the vehicle information 211 (vehicle type, vehicle speed, vehicle position, lane, vehicle ID, etc.) of the detected vehicle for each road image using the CPU, and processes the identified vehicle information 211. The data is output to the determination device 300.

  The road image storage unit 290 is a storage unit (storage device) that stores a road image and vehicle information using a storage medium.

  The information processing determination apparatus 300 includes a coordinate map unit 310 (an example of a vehicle information setting unit), a tracking processing unit 320 (an example of a non-detected vehicle specifying unit and a vehicle information estimation setting unit), a tracking comparison processing unit 330, and a mesh map information unit. 340 (an example of a vehicle information output unit), an intersection passing vehicle management unit 350, and a determination device storage unit 390 are provided.

The determination device storage unit 390 is a storage unit that stores the vehicle information table 391, the mesh map data 392, and the intersection passing vehicle management table 393 using a storage medium.
The vehicle information table 391 is data in which vehicle information of each vehicle currently traveling in the detection range is set.
The mesh map data 392 indicates information of a map (hereinafter referred to as “mesh map”) in which the detection range is divided into a mesh shape (lattice shape) of the traveling direction of the road 101 and the width direction of the road 101. The location of each vehicle currently traveling is represented on the mesh map.
The intersection passing vehicle management table 393 is data used for managing vehicle IDs, and for each of a plurality of vehicle IDs, a vehicle ID and a use state (“used” or “unused”) of the vehicle ID are associated with each other. Is set.

The coordinate map unit 310 inputs vehicle information output from the image sensor data processing device 200, and sets the input vehicle information in the vehicle information table 391 using the CPU.
Furthermore, the coordinate map unit 310 generates mesh map data 392 indicating the position where each vehicle is located on the mesh map using the CPU based on the vehicle information table 391.

The tracking processing unit 320 is included in each vehicle information newly output by the image sensor data processing device 200 among the vehicle IDs included in each vehicle information set in the vehicle information table 391 by the coordinate map unit 310. A vehicle ID that does not match any of the vehicle IDs to be identified is specified using the CPU.
Further, the tracking processing unit 320 determines the current vehicle position of the vehicle (hereinafter referred to as “detected omission vehicle”) identified by the vehicle ID based on the vehicle position included in the vehicle information including the specified vehicle ID. Estimate using. When the estimated vehicle position is included in the detection range, the tracking processing unit 320 sets vehicle information including the vehicle ID and the estimated vehicle position in the vehicle information table 391.
Furthermore, the tracking processing unit 320 specifies the vehicle ID of a vehicle that has passed the detection range (hereinafter referred to as “detection range passing vehicle”) based on the estimated vehicle position, and includes the specified vehicle ID. The vehicle information is deleted from the vehicle information table 391.
Further, the tracking processing unit 320 corrects the mesh map data 392 using the CPU based on the estimated vehicle position.

  The tracking comparison processing unit 330 calculates the number of vehicles currently traveling in the detection range using the CPU based on the vehicle information table 391.

  The mesh map information unit 340 outputs the vehicle information table 391, the mesh map data 392, and the number of vehicles to an output device (for example, a display device).

  The intersection passing vehicle management unit 350 updates the intersection passing vehicle management table 393 with the vehicle ID of the detection range passing vehicle specified by the tracking processing unit 320 as an unused ID.

FIG. 5 is a diagram illustrating an example of hardware resources of the image sensor data processing device 200 and the information processing determination device 300 according to the first embodiment.
In FIG. 5, the image sensor data processing device 200 and the information processing determination device 300 include a CPU 911 (also referred to as a central processing unit, a microprocessor, or a microcomputer) that executes a program. The CPU 911 is connected to the ROM 913, the RAM 914, the communication board 915, and the magnetic disk device 920 via the bus 912, and controls these hardware devices. Instead of the magnetic disk device 920, another storage device (for example, a semiconductor memory such as a RAM or a flash memory) may be used.

  The communication board 915 is connected to a communication network such as a LAN (Local Area Network), the Internet, and a telephone line by wire or wireless.

  The magnetic disk device 920 stores an OS 921 (operating system), a program group 923, and a file group 924. The programs in the program group 923 are executed by the CPU 911 and the OS 921.

  The program group 923 stores a program for executing a function described as “˜unit” in the embodiment. The program is read and executed by the CPU 911.

In the file group 924, in the embodiment, result data such as “determination result”, “calculation result of”, “processing result of” when executing the function of “to part”, “to part” The data to be passed between programs that execute the function “,” other information, data, signal values, variable values, and parameters are stored as items “˜file” and “˜database”. The vehicle information table 391, the mesh map data 392, and the intersection passing vehicle management table 393 are examples included in the file group 924.
The “˜file” and “˜database” are stored in a recording medium such as a disk or a memory. Information, data, signal values, variable values, and parameters stored in a storage medium such as a disk or memory are read out to the main memory or cache memory by the CPU 911 via a read / write circuit, and extracted, searched, referenced, compared, and calculated. Used for CPU operations such as calculation, processing, output, printing, and display. During these CPU operations, information, data, signal values, variable values, and parameters are temporarily stored in the main memory, cache memory, and buffer memory.
In addition, arrows in the flowcharts described in the embodiments mainly indicate input / output of data and signals, and the data and signal values are recorded in the memory of the RAM 914, the magnetic disk of the magnetic disk device 920, and other recording media. . Data and signal values are transmitted online via a bus 912, signal lines, cables, or other transmission media.

  In addition, what is described as “˜unit” in the embodiment may be “˜circuit”, “˜device”, “˜device”, and “˜step”, “˜procedure”, “˜”. Processing ". That is, what is described as “˜unit” may be realized by firmware stored in the ROM 913. Alternatively, it may be implemented only by software, or only by hardware such as elements, devices, substrates, and wirings, by a combination of software and hardware, or by a combination of firmware. Firmware and software are stored as programs on a magnetic disk or other recording medium. The program is read by the CPU 911 and executed by the CPU 911. That is, the program causes the computer to function as “to part”. Alternatively, the procedure or method of “to part” is executed by a computer.

FIG. 6 is a flowchart showing a vehicle detection tracking method of vehicle detection tracking system 100 in the first embodiment.
A vehicle detection tracking method (an example of a vehicle detection method) of the vehicle detection tracking system 100 according to the first embodiment will be described below with reference to FIG.
Each "~ part" of the vehicle detection tracking system 100 performs each process (S110-S160) demonstrated below using CPU.
Each process (S110 to S160) described below is repeatedly executed each time the road image 119 is captured and output by the first image sensor 110 and the second image sensor 111.

The image sensor data processing unit 210 inputs the road image 119 output from the first image sensor 110 and the second image sensor 111, and generates vehicle information 211 based on the input road image 119 (S110).
The coordinate map unit 310 updates the vehicle information table 391 based on the vehicle information 211 (S120), and updates the mesh map data 392 based on the road image 119 (S121).
The tracking processing unit 320 detects the detection omission vehicle and the detection range passing vehicle based on the vehicle information 211 and the old vehicle information table 391b (S130), and the vehicle information table 391 and the mesh map for the detection omission vehicle and the detection range passage vehicle. The data 392 is updated (S131).
The tracking comparison processing unit 330 specifies the number of vehicles based on the vehicle information table 391 (S140).
The mesh map information unit 340 displays the mesh map data 392 and the number of vehicles on the display device (S150).
The intersection passing vehicle management unit 350 changes the use state of the vehicle ID of the vehicle passing the detection range in the intersection passing vehicle management table 393 to “unused” (S160).
Below, the detail of each process (S110-S160) is demonstrated.

<S110>
The image sensor data processing unit 210 inputs the road image 119 output from the first image sensor 110 and the second image sensor 111, and stores the input road image 119 in the road image storage unit 290.
The image sensor data processing unit 210 generates and generates vehicle information 211 (vehicle type, vehicle speed, vehicle position, lane, vehicle ID, etc.) of each vehicle shown in the road image 119 based on the input road image 119. The vehicle information 211 of each vehicle is stored in the road image storage unit 290.
The image sensor data processing unit 210 outputs the road image 119 and the vehicle information 211 of each vehicle generated based on the road image 119 to the information processing determination device 300. The determination device storage unit 390 of the information processing determination device 300 stores the road image 119 output from the image sensor data processing unit 210 and the vehicle information 211 of each vehicle.

  For example, the image sensor data processing unit 210 performs an edging process on the road image 119 based on the luminance (color information) of each pixel to identify a pixel area in which the vehicle is shown, and the size and shape of the identified pixel area The vehicle type (large vehicle, ordinary vehicle, two-wheeled vehicle, etc.) is specified based on the above.

Further, for example, the image sensor data processing unit 210 determines the vehicle speed based on the moving amount of the pixel region of the vehicle and the imaging time interval in two consecutive road images 119 (the nth image and the n + 1th image). Is calculated.
In the road image 119 of the second image sensor 111 (an image obtained by imaging each vehicle from the upper side), the pixel area of the vehicle moves “2 pixels” in the traveling direction, and the range that appears in one pixel is “0.2 m” The distance traveled by the vehicle during the time from when the first road image 119 is captured until the second road image 119 is captured (imaging time interval) is “0.4 m (= 0.2 [m / Pixel] × 2 [pixel]) ”. When the imaging time interval is “1/30 second”, the vehicle speed is “43.2 km / h (= 12 [m / second] = 0.4 [m] ÷ (1/30 [second])). "
The image sensor data processing unit 210 may calculate the vehicle speed using the road image 119 of the first image sensor 110 (an image obtained by imaging each vehicle from the front upper side). For example, when the leading vehicle is traveling at a point away from the intersection to some extent, the road image 119 of the first image sensor 110 shows the entire front of the leading vehicle. The road image 119 is obtained by projecting a three-dimensional imaging range determined based on the installation height (height from the road surface), the line-of-sight direction (tilt), and the focus of the first image sensor 110 onto a two-dimensional plane. Therefore, the pixel on the road image 119 on which each point on the road surface is projected is determined. That is, for each pixel of the road image 119, the point on the road surface that appears in the pixel can be specified. The image sensor data processing unit 210 identifies, in each of the two road image vehicle information 211, a pixel in which a point on the road surface is captured in the pixel area of the leading vehicle, and based on the amount of movement of the identified pixel, the leading vehicle And the vehicle speed is calculated based on the movement distance and the imaging time interval.

  For example, the image sensor data processing unit 210 calculates the vehicle position based on the pixel area of the vehicle. In the road image 119 of the second image sensor 111, the pixel region where the stop line of the intersection is shown and the pixel region of the vehicle are shifted by “100 pixels” in the traveling direction, and the range shown in one pixel is “0.2 m”. In some cases, the vehicle position is a position that is “20 m (= 0.2 [m / pixel] × 100 [pixel])” from the stop line at the intersection. The image sensor data processing unit 210 may calculate the vehicle position using the road image 119 of the first image sensor 110. For example, the image sensor data processing unit 210 calculates the vehicle position of the leading vehicle based on the shift amount of the pixel area of the leading vehicle.

  Further, for example, the image sensor data processing unit 210 identifies a lane based on a pixel region of the vehicle (right side, center, left side, etc. in the road image 119). For example, when the leading vehicle is traveling at a point away from the intersection to some extent, the road image 119 of the first image sensor 110 shows the entire front of the leading vehicle. Further, as described above, for each pixel of the road image 119, a point on the road surface that appears in the pixel can be specified. The image sensor data processing unit 210 identifies a pixel in the road image 119 of the first image sensor 110 where a spot on the road surface is shown in the pixel area of the leading vehicle, and the leading vehicle travels based on the identified pixel. Identify the lane you are in.

The image sensor data processing unit 210 assigns a different vehicle ID to each vehicle one by one.
The image sensor data processing unit 210 appears in a road image 119 (hereinafter referred to as “first image”) of the first image sensor 110 and a road image 119 (hereinafter referred to as “second image”) of the second image sensor 111. In order not to give two vehicle IDs to the same vehicle, it is determined whether or not the vehicle shown in the first image and the image shown in the second image are the same vehicle. First, the image sensor data processing unit 210 compares the first vehicle information (vehicle type, vehicle speed, and vehicle position) acquired based on the first image with the second vehicle information acquired based on the second image. . Then, when the first vehicle information and the second vehicle information match, the image sensor data processing unit 210 is the same vehicle as the vehicle shown in the first image and the vehicle shown in the second image. If it is determined that there is no match between the first vehicle information and the second vehicle information, it is determined that the vehicle shown in the first image is different from the vehicle shown in the second image.
In addition, the image sensor data processing unit 210 has two consecutive road images so as not to assign two vehicle IDs to the same vehicle shown in the two consecutive road images 119 (first image or second image). It is determined whether or not the vehicles shown in 119 are the same vehicle. First, the image sensor data processing unit 210 compares the pixel area of the vehicle in the first road image 119 with the pixel area of the vehicle in the second road image 119. Then, when the movement amount of the pixel area is within a predetermined range (for example, 1 to 3 pixels), the image sensor data processing unit 210 and the vehicle and the second road image that are reflected in the first road image 119 are displayed. If the vehicle shown in 119 is determined to be the same vehicle, and the amount of movement of the pixel area is outside the predetermined range, the vehicle shown in the first road image 119 and the second road image 119 are shown. It is determined that the vehicle is a different vehicle.
The image sensor data processing unit 210 assigns an unused vehicle ID to the newly detected vehicle.

  After S110, the process proceeds to S120.

<S120>
The coordinate map unit 310 stores a copy of the vehicle information table 391 in the determination device storage unit 390 as an old vehicle information table 391b.
And the coordinate map part 310 updates the vehicle information table 391 based on the vehicle information 211 of each vehicle memorize | stored in the determination apparatus memory | storage part 390 in S110.

FIG. 7 is a diagram showing the vehicle information table 391 in the first embodiment.
The vehicle information table 391 updated in S120 (FIG. 6) will be described below based on FIG.

In the vehicle information table 391, a vehicle type, a vehicle speed, a vehicle position, a lane, and the like are set in association with the vehicle ID.
In S120 (FIG. 6), the coordinate map unit 310 searches the vehicle information table 391 using the vehicle ID included in the vehicle information 211 as a search key, and sets the set value of the record (row) with the matching vehicle ID in the vehicle information 211. Change to each included value. When there is no record in the vehicle information table 391 that matches the vehicle ID, the coordinate map unit 310 generates a record in which each value included in the vehicle information 211 is set, and adds the generated record to the vehicle information table 391.
For example, the coordinate map unit 310 updates and updates the vehicle information table 391 by changing and adding records as shown in FIG.

  Returning to FIG. 6, the description of the vehicle detection tracking method will be continued.

  After S120, the process proceeds to S121.

<S121>
The coordinate map unit 310 updates the mesh map data 392 based on the road image 119 stored in the determination device storage unit 390 in S110. The mesh map data 392 is information representing a mesh map described below.

FIG. 8 is a diagram showing a mesh map in the first embodiment.
The mesh map data 392 updated in S121 (FIG. 6) will be described below based on FIG.

The mesh map is illustrated by dividing the detection range into a mesh shape with a predetermined width (for example, 1 meter in length and width). Each intersection of the mesh map is identified by xy coordinates (x, y), where the traveling direction of the vehicle is “x axis” and the width direction of the road 101 is “y axis”.
FIG. 8 shows that the vehicle identified by the vehicle ID “n” is located in an area surrounded by four points (5, 5) (5, 7) (8, 5) (8, 7). Is shown.
In the mesh map data 392, four xy coordinates indicating an area where the vehicle is located are set as a record for each vehicle ID.

  In S121 (FIG. 6), the coordinate map unit 310 is similar to the image sensor data processing unit 210 (see S110), and the position of each vehicle shown in the road image 119 based on the road image 119 (or the position of each vehicle). The point on the road surface) is specified in the xy coordinate system of the mesh map. The coordinate map unit 310 identifies the corresponding vehicle ID with reference to the vehicle information table 391 (vehicle position, lane, etc.), and sets (changes / adds) the xy coordinates in the mesh map data 392 in association with the vehicle ID. . At this time, the coordinate map unit 310 searches the mesh map data 392 using the vehicle ID as a search key, and changes the set value (xy coordinate) of the record with the matching vehicle ID to the specified xy coordinate. If there is no record in the mesh map data 392 that matches the vehicle ID, the coordinate map unit 310 generates a record in which the vehicle ID and xy coordinates are set, and adds the generated record to the mesh map data 392. The coordinate map unit 310 updates the vehicle information table 391 by changing / adding records.

  Returning to FIG. 6, the description of the vehicle detection tracking method will be continued.

  After S121, the process proceeds to S130.

<S130>
The tracking processing unit 320 detects the detection omission vehicle and the detection range passing vehicle based on the vehicle information 211 of each vehicle stored in S110 and the old vehicle information table 391b stored in S120.
A detection-missing vehicle is a vehicle that is located within the detection range but has not been detected by the image sensor data processing device 200 due to reasons such as being hidden by another vehicle.
A detection range passing vehicle is a vehicle that has not been detected by the image sensor data processing device 200 because it has passed the detection range.
At this time, the tracking processing unit 320 extracts a record whose vehicle ID does not match any of the vehicle information 211 of each vehicle among the records set in the old vehicle information table 391b.
Further, the tracking processing unit 320 compares the vehicle position (distance from the intersection) set in the extracted record with a predetermined distance (for example, 1 meter). The tracking processing unit 320 determines that the record of the corresponding vehicle information table 391 is a record of a vehicle that is not detected if the vehicle position is equal to or greater than a predetermined distance, and the corresponding vehicle information table 391 if the vehicle position is less than the predetermined distance. Is determined to be a record of a vehicle passing through the detection range.

FIG. 9 is a diagram showing a record of a vehicle that is not detected and a record of a vehicle that has passed the detection range in the vehicle information table 391 in the first embodiment.
For example, as shown in FIG. 9, there is vehicle information 211 of a vehicle identified by a vehicle ID “2”, and three records (vehicle ID: “1” “2” “3”) are stored in the old vehicle information table 391b. Suppose that it is set. In S <b> 130 (FIG. 6), the tracking processing unit 320 excludes the record of the vehicle ID “2” from the three records set in the old vehicle information table 391 b and the record of the vehicle ID “1” and the vehicle ID “ 3 ”is extracted.
Further, the tracking processing unit 320 compares the vehicle position “0.5 m” set in the extracted record of the vehicle ID “1” with a predetermined distance “1 m”, and the vehicle position is less than the predetermined distance. Therefore, the record of the vehicle ID “1” in the vehicle information table 391 is determined as the record of the vehicle passing through the detection range. Further, the tracking processing unit 320 compares the vehicle position “10 m” set in the extracted record of the vehicle ID “3” with the predetermined distance “1 m”, and the vehicle position is equal to or greater than the predetermined distance. It is determined that the record of the vehicle ID “3” in the vehicle information table 391 is a record of a detection-missing vehicle.

  Returning to FIG. 6, the description of the vehicle detection tracking method will be continued.

  After S130, the process proceeds to S131.

<S131>
The tracking processing unit 320 updates the vehicle information table 391 and the mesh map data 392 for the detection-missing vehicle and the detection range passing vehicle detected in S130.
For example, the tracking processing unit 320 updates the vehicle information table 391 on the assumption that the detection-missing vehicle has traveled while maintaining the vehicle speed. When the record of the vehicle ID “3” in the vehicle information table 391 shown in FIG. 9 is updated, the tracking processing unit 320 sets the vehicle speed “54 km / h” and the imaging time interval (referred to as “1/30 seconds”). Based on this, the travel distance “0.5 m” of the detected vehicle is calculated, and the vehicle position “10 m” is changed to “9.5 m (= 10 m−0.5 m)”.
Further, for example, the tracking processing unit 320 updates the coordinates of the detection-missing vehicle included in the mesh map data 392 to the coordinates closer to the intersection by the travel distance.
In addition, the tracking processing unit 320 deletes the record of the vehicle passing the detection range from the vehicle information table 391 and the mesh map data 392.
After S131, the process proceeds to S140.

<S140>
The tracking comparison processing unit 330 counts the number of records included in the vehicle information table 391 and specifies the number of vehicles traveling in the detection range. The number of records included in the vehicle information table 391 corresponds to the number of vehicles traveling in the detection range.
After S140, the process proceeds to S150.

<S150>
The mesh map information unit 340 displays a mesh map as shown in FIG. 8 on a specific display device based on the mesh map data 392 updated in S131. The mesh map information unit 340 also displays the number of vehicles specified in S140 on the display device. The mesh map information unit 340 may further display a vehicle information table 391.
After S150, the process proceeds to S160.

<S160>
The intersection passing vehicle management unit 350 changes the use state from “in use” to “unused” for the vehicle ID of the detection range passing vehicle detected in S130 among the vehicle IDs set in the intersection passing vehicle management table 393. change.
When the image sensor data processing unit 210 detects a new vehicle to which no vehicle ID is assigned in S110, the image sensor data processing unit 210 acquires a vehicle ID whose use state is “unused” from the intersection passing vehicle management table 393, and uses the acquired vehicle ID. Assign to a new vehicle and change the usage state of the vehicle ID assigned to the new vehicle from “unused” to “in use”.
The intersection passing vehicle management unit 350 may record information such as the time of passing through the detection range and the lane in which the vehicle has traveled for each detection range passing vehicle.
After S160, the vehicle detection tracking method processing (S110 to S160) ends. The vehicle detection tracking method processing (S110 to S160) is repeatedly executed every time the road image 119 is captured and output by the first image sensor 110 and the second image sensor 111.

In the first embodiment, for example, the following vehicle detection tracking system 100 has been described.
In the vehicle detection and tracking system 100, image sensors (first image sensor 110 and second image sensor 111) are arranged in front of the road 101 (in the direction opposite to the traveling direction) and next to the road 101, for a total of two image sensors. Detect the vehicle with
The vehicle detection tracking system 100 includes an image sensor data processing device 200 and an information processing determination device 300.
The image sensor data processing device 200 determines the vehicle type identification information, vehicle speed, intersection stop line for each vehicle traveling on the road 101 based on the road image 119 captured by the first image sensor 110 and the second image sensor 111. Information such as the distance from the vehicle is acquired, a unique vehicle ID is assigned to each vehicle, and the acquired and assigned information (vehicle information 211) is communicated to the information processing determination device 300.
The information processing determination apparatus 300 includes a coordinate map unit 310 and a tracking processing unit 320.
The coordinate map unit 310 calculates the two-dimensional map information including the XY axes based on the data sent from the image sensor data processing device 200 based on the mounting position of the image sensor and the vehicle detection angle. The acquired vehicle information 211 and each vehicle are plotted on a mesh map.
The tracking processing unit 320 dynamically tracks each vehicle plotted on the mesh map on the map up to the intersection entrance (detection range) based on the vehicle information 211. The tracking processing unit 320 compares the latest vehicle information 211 of each vehicle sent from the image sensor data processing device 200 with the vehicle information 211 of each vehicle that is already on the mesh map and is dynamically tracked. New vehicles other than already tracked vehicles, vehicles that suddenly become undetectable due to any factors, and vehicles that can be detected again after being undetectable continue to be tracked dynamically.

FIG. 10 is a diagram illustrating a tracking example of the vehicle in the first embodiment.
The image sensor data processing device 200 assigns one vehicle ID to one preceding vehicle 125 (for example, a two-wheeled vehicle) detected by the first image sensor 110 and the second image sensor 111 at the same time. Further, when the image sensor data processing device 200 detects the subsequent vehicle 126 (for example, a large vehicle or a normal vehicle) following the preceding vehicle 125 by at least one of the first image sensor 110 and the second image sensor 111, A vehicle ID different from that of the preceding vehicle 125 is assigned to the succeeding vehicle 126 (FIG. 10A).
As illustrated in FIG. 10B, the information processing determination apparatus 300 detects that the following vehicle 126 has caught up with the preceding vehicle 125 and the preceding vehicle 125 has run in parallel with the following vehicle 126 (seeing from the second image sensor 111). When the vehicle 125 is hidden behind the following vehicle 126), the preceding vehicle 125 and the following vehicle 126 can be dynamically tracked by using the road image 119 of the first image sensor 110.
In addition, as illustrated in FIG. 10C, the information processing determination device 300 is configured so that the following vehicle 126 passes the preceding vehicle 125 and the preceding vehicle 125 follows the following vehicle 126 (as viewed from the first image sensor 110). When the vehicle 125 is hidden behind the succeeding vehicle 126), the preceding vehicle 125 and the following vehicle 126 can be dynamically tracked by using the road image 119 of the second image sensor 111.
In addition, as illustrated in FIG. 10D, the information processing determination device 300, when the preceding vehicle 125 runs in parallel with the subsequent vehicle 126, the road image 119 of the first image sensor 110 and the road image of the second image sensor 111. By using at least one of 119, the preceding vehicle 125 and the following vehicle 126 can be dynamically tracked.
In the case of FIG. 10D, both the road image 119 of the first image sensor 110 (hereinafter referred to as “first image”) and the road image 119 of the second image sensor 111 (hereinafter referred to as “second image”). The preceding vehicle 125 is shown in FIG. The image sensor data processing device 200 matches the first vehicle information (vehicle type, vehicle speed, and vehicle position) acquired from the first image with the second vehicle information acquired from the second image. The reflected preceding vehicle 125 and the preceding vehicle 125 reflected in the second image are recognized as the same vehicle. Similarly, the image sensor data processing device 200 recognizes the following vehicle 126 shown in the first image and the following vehicle 126 shown in the second image as the same vehicle.

  As a result, the vehicle detection tracking system 100 can be used for temporary vehicles such as a vehicle (ordinary vehicle, two-wheeled vehicle) hidden behind a large vehicle, a vehicle running in parallel behind a large vehicle, or a vehicle surrounded by a plurality of large vehicles. In particular, it is possible to track (detect) a vehicle that is not shown in the road image 119 and reduce detection omission of the vehicle.

  DESCRIPTION OF SYMBOLS 100 Vehicle detection tracking system, 101 Road, 102 Intersection, 103 Traveling direction, 110 1st image sensor, 111 2nd image sensor, 118 Imaging range, 119 Road image, 121 Large vehicle, 122 Normal vehicle, 123 Two-wheeled vehicle, 124 Normal vehicle , 125 preceding vehicle, 126 following vehicle, 190 image sensor data, 200 image sensor data processing device, 210 image sensor data processing unit, 211 vehicle information, 290 road image storage unit, 300 information processing determination device, 310 coordinate map unit, 320 Tracking processing unit, 330 Tracking comparison processing unit, 340 mesh map information unit, 350 intersection passing vehicle management unit, 390 determination device storage unit, 391 vehicle information table, 391b old vehicle information table, 392 mesh map data, 393 intersection passing vehicle management Te Bull, 911 CPU, 912 Bus, 913 ROM, 914 RAM, 915 communication board, 920 a magnetic disk device, 921 OS, 923 programs, 924 files.

Claims (4)

A plurality of road images taken of the road is continuously inputted in sequence, one at a time, the one or more with one or more of the vehicle reflected in the road image on the basis of the road image for each road image input The position of each of the vehicles is detected using the CPU, and one or more vehicles detected from the newly input road image are not used for vehicles different from any of the vehicles detected from the previously input road image Of one or more vehicles detected from the newly input road image, the same vehicle ID is assigned to the same vehicle as one of the vehicles detected from the previously input road image, and newly a road image processing section for outputting information including the vehicle position indicating the a information for each vehicle detected from the input road image position of the vehicle ID and the vehicle was applied to the vehicle as a vehicle both information,
Of the vehicle information output by the road image processing unit, vehicle information including a vehicle ID different from any vehicle information set in the vehicle information table is selected as additional vehicle information, and the additional vehicle Information is additionally set in the vehicle information table, and vehicle information including the same vehicle ID as any one of the vehicle information set in the vehicle information table is selected as the vehicle information for update. A vehicle information setting unit for updating the vehicle position of the vehicle information including the same vehicle ID as the vehicle information for update among the vehicle information set in the vehicle information table to the vehicle position of the vehicle information for update ; ,
Using the CPU, a vehicle ID that does not match any of the vehicle IDs included in each of the vehicle information output by the road image processing unit among the vehicle IDs included in each of the vehicle information set in the vehicle information table. An undetected vehicle identifying unit that identifies the undetected vehicle ID;
Using the CPU, the current vehicle position of the vehicle identified by the undetected vehicle ID based on the vehicle position included in the vehicle information including the undetected vehicle ID identified by the undetected vehicle identification unit. Vehicle information estimation setting for estimating and updating the vehicle position included in the vehicle information including the undetected vehicle ID to the estimated vehicle position when the estimated vehicle position is included in the range shown in the road image And
A vehicle detection device comprising: a vehicle information output unit configured to output each vehicle information set in the vehicle information table to an output device as information on a vehicle traveling on the road. Is installed in the traveling direction and opposite of the road around the road, a first image sensor for capturing a road image,
Disposed toward the road to the side of the road, and the second image sensor for imaging a road image,
And a vehicle detection device according to claim 1 Symbol placement,
Road image processing unit of the vehicle detection device, the first image vehicle for each vehicle reflected in at least one of the road image captured by the a road image captured second image sensor by a sensor A vehicle detection system characterized by outputting information. Road image processing unit, a plurality of copies in succession road road image type in sequence, one at a time, one or more that appearing in the road image on the basis of the road image for each road image input Any vehicle detected from the previously input road image among the one or more vehicles detected from the newly input road image by detecting the vehicle and the respective positions of the one or more vehicles using the CPU. An unused vehicle ID is assigned to a different vehicle, and among the one or more vehicles detected from the newly input road image, the same vehicle ID is assigned to the same vehicle as one of the vehicles detected from the previously input road image. the grant, and outputs the information including the vehicle position indicating the position of the vehicle ID and the vehicle was given to the vehicle to a detected information of each vehicle from the newly input road image as car both information road image Process,
The vehicle information setting unit selects vehicle information including a vehicle ID different from any vehicle information set in the vehicle information table among the vehicle information output by the road image processing unit as additional vehicle information. The additional vehicle information is additionally set in the vehicle information table, and the vehicle information including the same vehicle ID as any one of the vehicle information set in the vehicle information table among the vehicle information is updated. The vehicle position of the vehicle information including the same vehicle ID as the vehicle information for update among the vehicle information set in the vehicle information table is updated to the vehicle position of the vehicle information for update. Vehicle information setting process to
The undetected vehicle specifying unit does not match any of the vehicle IDs included in each of the vehicle information output by the road image processing unit among the vehicle IDs included in each of the vehicle information set in the vehicle information table. An undetected vehicle identification process is performed for identifying the vehicle ID as an undetected vehicle ID using the CPU.
The vehicle information estimation setting unit has a current vehicle current identified by the undetected vehicle ID based on the vehicle position included in the vehicle information including the undetected vehicle ID identified by the undetected vehicle identification unit. When the vehicle position is estimated using a CPU and the estimated vehicle position is included in the range shown in the road image, the estimated vehicle position is included in the vehicle information including the undetected vehicle ID. Vehicle information estimation setting process to be updated to
A vehicle information output unit performs a vehicle information output process for outputting each vehicle information set in the vehicle information table to an output device as information on a vehicle traveling on the road. Vehicle detection method. A vehicle detection program for causing a computer to execute the vehicle detection method according to claim 3 .

Images