JP3007960B2 - Traveling vehicle detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a traveling vehicle detecting device for detecting a vehicle traveling on a road in a wide area and grasping the condition of the vehicle and the road.

[0002]

2. Description of the Related Art Conventionally, as a wide area monitoring device for monitoring a vehicle traveling on a wide area road using a plurality of image sensors,
One example is a “distributed monitoring system” in which one subsystem is provided for each image sensor, and the subsystems are connected.
An example of this is the 1993 IEICE Autumn Meeting (D
-157) and a "distributed monitoring system" constructed by the subsystem and main system. FIG. 6 shows the configuration of this device. In FIG. 6, 801
Is a main system having a function of monitoring and controlling the entire system, and 802 and 803 are subsystems. The main system 801 has a hierarchical structure in which an upper layer and the subsystems 802 and 803 are a lower layer. Although only two subsystems are shown in FIG. 6, these subsystems can be added and their operations are common.

[0003] The operation of the wide area monitoring device configured as described above will be described below. Each subsystem 80
2,803 divides the monitoring area in charge of each system into multiple overlapping areas and combines the monitoring information obtained in each area, using the results of monitoring target extraction and its characteristics. Then, it creates recognition parameters describing the image features and distributes them to other subsystems. The delivered recognition parameters are registered as tracking data in the respective subsystems 802 and 803, and are registered in the respective subsystems 802 and 803.
In the 2,803 monitored areas, objects having similar characteristics are searched. If the target exists, the detection information is transferred to the main system 801 which is the upper layer, and a request to confirm the target is made. The main system 801 includes a subsystem 802,
Based on the request of 803, the target is confirmed, and image information and the like are fetched from the subsystems 802 and 803 as necessary, thereby tracking, searching, and identifying the target.

An example of a vehicle detection device described in the Institute of Electrical Engineers of Japan Road Traffic Research Group (RTA-96-9) corresponding to the above-mentioned "subsystem" conventionally used for detecting a running vehicle. There is. FIG. 7 shows the configuration of this vehicle detection device, and its operation will be described. Vehicle detection
A vehicle 902 individually detects each vehicle from an image input from the camera 901 and generates rectangular data (enclosed by a circumscribed rectangle) representing the detected vehicle. In this method, a background image obtained by removing all moving objects from an input image is generated, and a moving region is extracted based on a difference between the background image and the input image. An area that can be determined to be a vehicle is detected from the moving area, and is set as an area of the traveling vehicle. At this time, if the vehicles are detected overlapping, the underline detection position of the vehicle indicates an individual vehicle area, and that area is defined as rectangular data. Next, in order to identify the vehicle, a label inherited in the time direction is given to each rectangular data by the label 903.

[0005] The inheritance of the label follows the rectangular data of the previous frame having an overlap rate equal to or greater than the threshold value. When two rectangular data share the same label, they are replaced with their circumscribed rectangles. Then, a new label is given to the rectangular data whose label could not be inherited. Next, in the 3D conversion 904, the position, width, and height of the vehicle in the real world are calculated from the rectangular size on the image plane, and rectangular data of an inappropriate size or position for the vehicle is deleted. Finally, the average value of the actual width and height of the vehicle is obtained by the traffic measurement 905, and the average speed of the vehicle is calculated from the detected position and detection time by the "least square method".

[0006]

However, the "distributed monitoring system" shown in FIG. 6 employs a hierarchical system in which an upper system and a lower system are combined, so that the number of lower systems remains unchanged. This leads to an increase in the load on the host system. Further, there is a problem in the expandability and flexibility of the system, such as the need for backup of the upper system or the difficulty in independently changing the lower system and the upper system. In addition, in the case of the traveling vehicle detection device shown in FIG. 7, when an obstacle exists due to detection by a single single camera, or when a plurality of objects are close to each other, a part of the objects is Sometimes shooting is not possible. In this case, it becomes difficult to detect and track the target object, and the reliability of the monitoring result decreases. Also, in monitoring with a plurality of cameras, various combinations of camera installation positions are possible, and the monitoring accuracy greatly depends on the camera installation positions, so that there is a problem that reliability also decreases. .

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems of the conventional traveling vehicle detection apparatus, to make it easy to expand the detection area and to change the function, to have a simple structure, and to provide a simple structure. An object of the present invention is to provide a traveling vehicle detection device that can stably extract and track an object even when objects and objects are close to each other.

[0008]

According to a first aspect of the present invention, there is provided a vehicle detecting apparatus for monitoring a road in a wide area and detecting a running vehicle, wherein the monitored area is divided into a plurality of small monitored areas. In each of the small monitoring areas, a plurality of small monitoring apparatuses that enable a plurality of image sensors arranged at different positions to monitor the small monitoring areas redundantly, and which of the small monitoring apparatuses among the plurality of small monitoring apparatuses It is possible to set in advance whether to manage the device, and a plurality or a single management device that collectively manages monitoring results of vehicle information from each small monitoring device, and a single device that connects this management device and each small monitoring device. A configuration in which each of the small monitoring devices independently monitors the area in charge and each management device obtains the monitoring result from the small monitoring device as necessary. By taking It has the effect of realizing the traveling vehicle detection device which is rich in extensibility by a single structure.

According to a second aspect of the present invention, in the first aspect of the present invention, the management of the small monitoring device by the management device on a single network is dynamically set according to the load of the management device. The feature is that it is possible.

[0010] The invention described in claim 3 is the invention according to claim 1 or 2.
In the invention described in (1), the small monitoring device has N cameras, and N / 2 cameras face each other so as to face forward and backward on the road, and the cameras facing in the same direction have substantially the same shooting area. In addition, the cameras facing each other are installed so as to have overlapping shooting areas, and the two cameras are directed to the same direction, are processed by the target monitoring device, and two primary integrated processing devices are configured to perform integrated processing of the results. It is characterized by having a secondary integrated processing device that integrates the result of the primary integrated processing device, and has an effect of improving the vehicle detection accuracy in each small monitoring device.

The invention described in claim 4 is the first or second invention.
In the invention described in the sub-monitoring device, the target monitoring device includes a detection unit of a moving region, a labeling unit, a position conversion unit of a lower end position and a real space on the screen of the detected moving region,
The result is integrated by a primary integrated processing device to obtain the position of one moving area, and the motion detection of the lower end position photographed from before and after one moving area and the size of the moving area are obtained. Has the function of realizing the function.

The invention described in claim 5 is the invention according to claim 3 or 4.
In the invention described in the above, in the secondary integrated processing device, the small monitoring device uses the position L _f detected from the front of one moving region and the position L _r of the same moving region similarly photographed from the rear, and The length L = | L _f −L _r | is calculated, and a process of estimating the number of target objects existing in the moving region from the length L is performed. Has the effect of determining the number of vehicles existing in the vehicle.

The invention described in claim 6 is the third or fourth invention.
In the invention according to, a small monitoring device secondary consolidated processing device, the speed V _f and the same moving region taken from behind similarly determined according to a continuous process from the position L _f detected from the front of one moving region those positions using the velocity V _r determined according to a continuous process from L _r, compares the V _f and V _r, which comprises treating a group of a plurality of target objects as long as substantially the same speed This has the effect of defining the properties of one moving area and simplifying the method of vehicle detection.

According to a seventh aspect of the present invention, in the invention according to any one of the third to sixth aspects, the small monitoring device uses a group of target objects detected by the secondary integrated processing device to set a small monitoring area. The corresponding device transmits the length L and speed V of the moving area of each group to a higher-level management device as group information of a plurality of target objects in the section, and performs wide area monitoring. Has the effect of improving the detection accuracy of the target object in.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. FIG. 1 shows an embodiment of a device for detecting a traveling vehicle in a wide area according to the present invention. In FIG. 1, 111, 112, 113, and 114 are small monitoring devices, 101 is a management device, and 130 is a network. The number of management devices can be increased according to the number of small monitoring devices. The monitoring area is divided into a plurality of small monitoring areas, and one small monitoring device monitors one small monitoring area. In FIG. 1, the monitoring area is divided into four small monitoring areas, and monitoring is performed by four small monitoring devices.

Each of the small monitoring devices functions independently, and monitors the assigned small monitoring area by extracting and tracking a monitoring target. In each of the small monitoring devices, for example, an image of the inside of the monitoring area is captured by a camera, an object is extracted based on an inter-frame difference, a background difference, or the like, and the position of the traveling vehicle is tracked. The management device obtains, via the network 130, the number and attributes of individual vehicles and a plurality of vehicles running at a constant speed (defined as a group of vehicles in the present invention) detected from the small monitoring device, as necessary. The monitoring result of an arbitrary small monitoring device is displayed, and the monitoring results of a plurality of small monitoring devices are collectively displayed. When the number of small monitoring devices is large, a plurality of management devices are provided. For example, the management device 101 manages the results of the small monitoring devices 111 and 112, and the other small monitoring devices 113 and 114 newly A small monitoring device to be managed by each management device is set in advance, such as managing the results by another management device provided. As a result, the small monitoring device and the management device are handled in exactly the same manner on the network, and communication between the devices is realized by an appropriate communication protocol.

Although FIG. 1 shows a configuration in which there are four small monitoring devices and one management device, the number of these devices can be changed. Alternatively, a small monitoring device to be in charge of each management device may be dynamically assigned to an appropriate management device while monitoring the load of each management device without previously setting the small monitoring device to be in charge of the management device.

FIG. 2 shows an embodiment of the small monitoring apparatus according to the present invention. In this embodiment, four visible cameras are shown. The number of cameras can be increased according to the detection accuracy. The cameras 201 and 211 are installed so as to face the same direction along the road and have substantially the same shooting range. The cameras 221 and 231 are also installed so as to face the same direction along the road and have substantially the same shooting range. In addition, cameras 201 and 211 and a camera
The cameras 221 and 231 face each other, and the cameras facing each other are installed so as to have overlapping shooting ranges (FIG. 3).
reference).

Reference numerals 202 and 212 denote object monitoring devices, for example, using images input from the cameras 201 and 211, for example,
The target is extracted using the inter-frame difference and the moving region extraction.
The result is output to the primary integrated processing device 203. 222, 23
Reference numeral 2 denotes a target monitoring device that extracts a target vehicle by using images input from the cameras 221 and 231 respectively. As in the present embodiment, one set of the target monitoring device corresponds to one camera. However, if the capability of the target device is improved, a plurality of cameras can be supported. Next, the output of the target monitoring device that has processed the video captured from the same direction is output to the primary integrated processing device 223.

In the primary integrated processing devices 203 and 204, the monitoring results input from the two target monitoring devices respectively in charge, for example, attributes such as the number of objects and their positions, are integrated, and the results are subjected to the secondary processing. Output to the integrated processing device 204. 2
The secondary integrated processing unit 204 integrates the monitoring results input from the primary integrated processing units 203 and 223 to obtain the final tracking result of the traveling vehicle. Then, based on the attribute of the target obtained through the network 130 and the attribute of the target obtained by the secondary integrated processing device 204, the target grasped by the other small monitoring device is associated with the subject currently grasped. . By installing a plurality of the small monitoring devices so that the monitoring areas are adjacent to each other, an arbitrary section on a road can be monitored.

Also, in the primary integrated processing device, the processing of the video image in which the cameras are installed on both sides of the road is processed by the target monitoring device, and the result is used to accurately capture the position of the target. As a method of this processing, the position and height of the traveling vehicle can be estimated using multi-view processing using a triangulation method.

FIG. 3 shows an example of installation of the camera in FIG. The cameras 201, 211, 221, and 231 capture images from before and after a vehicle traveling in a monitoring area defined by the small monitoring device. In the example in FIG. 3, the cameras 201 and 211 capture images from the front, and the cameras 221 and 231 capture images from the rear. Camera 201 and camera 211
The cameras 221 and 231 are installed at the same position on the road from both sides of the road. The camera 201 and the camera 211 have a purpose of detecting the moving region of the traveling vehicle with a small blind spot by photographing the detected moving region from both sides,
The cameras 201 and 211 and the cameras 221 and 231 have the purpose of accurately detecting the length, width, and speed of an individual vehicle or vehicle group by detecting the traveling vehicle from front and rear.

FIG. 4 shows an example of detecting an individual vehicle or vehicle group detected from front, rear, left and right. In this example, detection of a vehicle group will be particularly described. In other words, the information that can be detected from the video ahead is
A vehicle speed V _f can be calculated from the change in the position L _f and temporal position of the leading vehicle of the vehicle group. Similarly, from the rear of the video, the position L _h and the vehicle speed V _h of the last Oguruma it is required. In this case, when the relationship between L _f , L _h , V _f , and V _h is arranged, the estimated number of vehicles included in the observation section is estimated from the distance L between the tail position and the head position, and the speed of the tail vehicle is determined. the leading car speed comparing | V _f -V _h | is one region of the state, for example, can decide whether to a constant speed collective, the one feature quantity predicting subsequent state changes.

Based on the above, as a method of judging one moving region, L ≦ L _t (L _t is a threshold suitable for one vehicle length), and | V _f −V _h | ≒ 0 in is capable of determining as one _{car, L max ≧ L ≧ L t} (L max is the maximum length of the vehicle group), | V _f -V _h | same plurality of vehicles when the ≒ 0 It can be seen that a group of vehicles, which is one aggregate running at a speed, is formed. The vehicle group detected as described above can be handled as a tracking target similar to one vehicle in grasping the traveling vehicle. This group of vehicles can be used for a practical system for grasping the state of traffic flow (natural flow, congestion, sudden events).

Next, FIG. 5 shows a configuration example of a small monitoring device for detecting a vehicle group. In FIG. 5, target monitoring devices 501, 502, 503, and 504 for each camera (camera image), primary integrated processing devices 505 and 506 for integrating the processing results of the target monitoring devices,
It comprises a secondary integrated processing device 507 for integrating the processing results of the primary integrated processing device. The target monitoring devices 501 and 502 process camera images taken from the front, and
At 3,504, a video taken from behind is processed. Each target monitoring device detects a moving region from each camera image based on an inter-frame accumulated difference or a background difference, performs a labeling process on the moving region, and obtains a lower end position of each moving region.

Since each lower end position is almost near the ground in the real space, if the installation position of each camera in the real space is known, the lower end portion of the vehicle area in the video is displayed on the real space of the vehicle. The position can be determined from the linear mapping function. Then, the primary integrated processing units 505 and 506 shape the positions of the traveling vehicles obtained from the two front and rear camera images and extract highly accurate vehicle positions. The secondary integrated processing device 507 detects a vehicle group from the position and speed calculated from the front and rear moving regions shown in FIG. As described above, it is assumed that a plurality of vehicles are included in the moving area. The information that can be detected from this area is as shown in FIG.
As shown in FIG. 7, vehicles traveling between narrow vehicles form one group.

[0027]

The present invention is as described above. According to a first aspect of the present invention, there is provided a vehicle detecting apparatus for monitoring a road in a wide area and detecting a running vehicle, wherein a monitoring area includes a plurality of small areas. A plurality of small monitoring devices that are divided into monitoring areas, and in each of the small monitoring areas, a plurality of image sensors arranged at different positions can monitor the small monitoring areas in an overlapping manner; It is possible to set in advance which of the small monitoring devices to manage, and a plurality of or single management devices that collectively manage the monitoring results of the vehicle information from each of the small monitoring devices; A plurality of independently functioning small monitoring devices, characterized by comprising a single network coupling the monitoring devices;
By providing a plurality or a single management device, it is possible to realize a traveling vehicle detection device in a wide area where the monitoring area can be easily expanded and the configuration is simple, and a traveling vehicle detection device having a large scale and a scalability is provided. This can be achieved, and there is an effect that an object moving across a small monitoring area can be tracked with only the small monitoring device.

According to a second aspect of the present invention, in the first aspect, the management of the small monitoring device by the management device on a single network is dynamically set according to the load of the management device. Since it is possible, there is an effect that it is possible to easily increase or decrease each device and change the relationship between the small monitoring device and the management device.

The invention described in claim 3 is the invention according to claim 1 or 2
In the invention described in (1), the small monitoring apparatus has N cameras, and these cameras are arranged so that N / 2 cameras face each other toward the front and rear of the road, and the cameras facing the same direction have substantially the same shooting area. In addition, the cameras facing each other are installed so as to have overlapping shooting areas, and the two cameras are directed to the same direction, are processed by the target monitoring device, and two primary integrated processing devices are configured to perform integrated processing of the results. Since there is a secondary integrated processing device that integrates the result of the primary integrated processing device, it is possible to stably monitor the target without being affected by the presence of an obstacle or the approach between the targets. In addition, there is an effect that the vehicle detection accuracy in the small monitoring device can be improved, and an object moving across the small monitoring area can be tracked only by the small monitoring device.

The invention described in claim 4 is the first or second invention.
In the invention described in the sub-monitoring device, the target monitoring device includes a detection unit of a moving region, a labeling unit, a position conversion unit of a lower end position and a real space on the screen of the detected moving region,
Since the result is integrated by the primary integrated processing device to determine the position of one moving area, it is easy to expand the monitoring range and change the function, and a large-scale monitoring system can be realized with a simple configuration, and obstacles can be reduced. There is an effect that a system for stably detecting and tracking a target can be realized even if the target exists or the targets are close to each other.

The fifth aspect of the present invention provides the third or fourth aspect.
In the invention described in the above, in the secondary integrated processing device, the small monitoring device uses the position L _f detected from the front of one moving region and the position L _r of the same moving region similarly photographed from the rear, and length L = | L _f -L _r | is calculated, since the processing for estimating the number of target objects present than this length L in the moving area, facilitate the number of vehicles on the within the single moving region And a system for stably detecting and tracking an object can be realized.

[0032] The invention according to claim 6, continuous in the invention of claim 1, 3 or 4, in a small monitoring device secondary integration processing unit from the position L _f detected from the front of one moving region and the speed V _f determined according to treatment, using the velocity V _r determined according to the process is continuous from the position L _r of the same motion area taken from behind as well, compared with the V _f and V _r,
If the speed is almost the same, it is processed as a group of multiple target objects, so that stable road monitoring can be realized,
By detecting a plurality of vehicles running at a constant speed as a group of vehicles in addition to individually detecting traveling vehicles, there is an effect that a large-scale monitoring system can be realized with a simple configuration.

According to a seventh aspect of the present invention, in the invention of any one of the third to sixth aspects, the small monitoring device uses a group of target objects detected by the secondary integrated processing device to set a small monitoring area. The corresponding device transmits the length L and velocity V of the moving area of each group as group information of a plurality of target objects in the section to a higher-level management device, and performs monitoring over a wide area, thereby improving detection accuracy. In addition, the monitoring range can be expanded and the function can be easily changed, and a large-scale monitoring system can be realized with a simple configuration. In addition, even if obstacles are present or objects are close to each other, a system for stably detecting and tracking the objects can be realized. There is an effect that effective information of a group of vehicles can be detected collectively.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an embodiment of a traveling vehicle detection device of the present invention.

FIG. 2 is a configuration diagram of an embodiment of a small monitoring device used in the traveling vehicle detection device of the present invention.

FIG. 3 is a camera layout diagram in the embodiment of the small monitoring device of the present invention.

FIG. 4 is an explanatory diagram of a method of detecting a vehicle group in the small monitoring device of the present invention.

FIG. 5 is a configuration diagram of an embodiment of detecting a vehicle group in the small monitoring device of the present invention.

FIG. 6 is a configuration diagram of a conventional example similar to the present invention.

FIG. 7 is a configuration diagram of a conventional example of the same type as the present invention.

[Explanation of symbols]

 101 Management device 111-114 Small monitoring device 130 Network 201, 211, 221, 231, 901 Camera 202 Target monitoring device 203 Primary integrated processing device 204 Secondary integrated processing device 212, 222, 232 Target monitoring device 223 Primary integrated processing Equipment 501-504 Object monitoring equipment 505,506 Primary integrated processing equipment 507 Secondary integrated processing equipment 801 Main system 802,803 Subsystem 902 Vehicle detection 903 Label 904 3D exchange 905 Traffic measurement

────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kazufumi Mizusawa 1006 Kazuma Kadoma, Osaka Prefecture Matsushita Electric Industrial Co., Ltd. (56) References JP-A-9-330494 (JP, A) JP-A-6 −333193 (JP, A) Japanese Utility Model 2-67498 (JP, U) (58) Fields surveyed (Int. Cl. ⁷ , DB name) G08G 1/00-1/16

Claims (7)

(57) [Claims] In a vehicle detection device for monitoring a road in a wide area and detecting a running vehicle, the monitoring area is divided into a plurality of small monitoring areas, and in each of the small monitoring areas,
A plurality of small monitoring devices that allow a plurality of image sensors arranged at different positions to monitor the small monitoring area redundantly,
Which of the plurality of small monitoring devices is to be managed.
Can be set in advance, and the
Multiple or independent management of vehicle information monitoring results
Management device, and this management device and each small monitoring device are connected.
A traveling vehicle detection device comprising a single network . 2. A management device on a single network.
The management of small monitoring devices by equipment depends on the load of the management device.
The traveling vehicle detection device according to claim 1, wherein the traveling vehicle detection device can be dynamically set . 3. The small surveillance device has N cameras, and N cameras are arranged so as to face each other forward and backward on the road.
/ 2 cameras are installed so that the cameras facing in the same direction have almost the same shooting area, and the cameras facing each other have overlapping shooting areas, and processing is performed by the target monitoring device for each camera facing in the same direction. 3. The apparatus according to claim 1, further comprising: two primary integrated processing devices for performing the integrated processing of the result and a secondary integrated processing device for integrating the result of the primary integrated processing device. Traveling vehicle detection device. 4. The small monitoring device includes a moving area detecting section, a labeling section, a lower end position of the detected moving area on the screen, and a real space position converting section as a target monitoring apparatus. The traveling vehicle detection apparatus according to claim 3, wherein the position of one moving area is obtained by integrating the moving area by a next integration processing apparatus. 5. The secondary integrated processing apparatus uses a position L _f detected from the front of one moving region and a position L _r of the same moving region similarly photographed from the rear, to obtain a length L = | L _f −
5. The traveling vehicle detection device according to claim 3, wherein a process of calculating L _r | and estimating the number of target objects existing in the moving region from the length L is performed. In 6. The secondary integration processing apparatus, the speed V _f determined according to a continuous process from the position L _f detected from the front of one moving region from the position L _r of the same motion area taken from behind as well using the velocity V _r determined according to a continuous process, compared with the V _f and V _r, in claim 3 or 4, characterized in that process as a group of a plurality of target objects as long as substantially the same speed The traveling vehicle detection device as described in the above. 7. The small monitoring device uses a group of target objects detected by the secondary integrated processing device, and the small monitoring area correspondence device uses group information of a plurality of target objects in a section as group information of a moving region of each group. 7. The traveling vehicle detection device according to any of 3 to 6, wherein the length L and the speed V are transmitted to a higher-level management device to monitor a wide area.