JPH11185194A - Vehicle detector - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a vehicle detector capable of correctly detecting a vehicle approaching an intersection by using an image processing technique. SOLUTION: This vehicle detector is provided with a TV camera 1 for picking up the image of a road, a moving body extraction device 2 for extracting a moving body from the image obtained by the TV camera 1, a vehicle judgement device 3 for judging whether or not the moving body obtained by the moving body extraction device 2 is a vehicle, an advance judgement device 4 for judging whether or not the moving body judged as the vehicle by the vehicle judgement device 3 reaches an approach judgement position set beforehand, a vehicle tracking device 5 for tracking the vehicle judged as reaching the approach judgement position by the advance judgement device 4 and a passing judgement device 6 for judging whether or not the vehicle tracked by the vehicle tracking device 5 reaches a passing judgement position set beforehand.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle detecting device suitable for detecting a vehicle entering an intersection.

[0002]

2. Description of the Related Art Heretofore, counting the number of vehicles entering an intersection and identifying vehicles that ignore signals have been performed manually, and it has been difficult to permanently monitor vehicles at an intersection. In recent years, image processing technology has been used for vehicle detection, but it is often used on straight roads, and there are multiple vehicles such as right-turn and left-turn vehicles, and it detects vehicles entering an intersection where pedestrians also exist. The method is still not enough.

[0003] On the other hand, if a vehicle entering an intersection can be detected at the time of entry, the number of vehicles entering the intersection can be accurately counted, which is a measure of traffic congestion at the intersection. In addition, if a traffic signal can detect a vehicle that has entered an intersection with a red light, the vehicle that has ignored the signal can be identified, and the image at that time can be recorded as evidence.

[0004]

SUMMARY OF THE INVENTION In view of the above background, an object of the present invention is to provide a vehicle detection device that can accurately detect a vehicle entering an intersection by using an image processing technique. It is another object of the present invention to provide a vehicle detection device capable of counting the number of vehicles entering an intersection based on detected vehicle information and detecting a vehicle ignoring a signal.

[0005]

SUMMARY OF THE INVENTION The present invention provides an image pickup means for picking up an image of a road, a moving object extracting means for extracting a moving object from an image obtained by the image pickup means, and a moving object obtained by the moving object extracting means. Vehicle determination means for determining whether or not the vehicle is a vehicle; approach determination means for determining whether the moving object determined to be a vehicle by the vehicle determination means has reached a preset approach determination position; Means for tracking a vehicle determined to have reached the entry determination position by the means, and a pass for determining whether the vehicle tracked by the vehicle tracking means has reached a preset pass determination position The above problem has been solved by providing the determination means.

In the present invention, the number of vehicles entering the intersection is counted by providing the number of vehicles entering the intersection by outputting the passage judgment output from the passage judgment means and counting the number of vehicles. it can. In addition, if a signal for determining a vehicle that ignores a signal is provided based on the output of the determination signal and the state of the traffic signal, a vehicle that ignores the signal can be detected.

Hereinafter, the present invention will be described in detail. A conventionally known TV camera may be used as an imaging unit for imaging the road. The imaging range may include only the lane in which the vehicle is to be detected, or the entire intersection when the vehicle is detected at the intersection. The advantage of setting the entire intersection as the imaging range will be described in the section of the embodiment described later.

[0008] If there is an object that moves based on the image obtained by the imaging means (referred to as "moving object" in this specification), it is extracted by the moving object extracting means. As a method for extracting a moving object, there is a method of comparing the brightness of an image captured based on the brightness of a road surface and, when there is a difference of a certain value or more, using that portion as a moving object, A method of extracting a vehicle edge and an outline by performing processing, a method of extracting a difference in brightness between two images at a predetermined time interval, and performing a temporal differentiation of the brightness to extract an edge and an outline of the vehicle are known. However, any method can be applied.

The moving objects extracted by the moving object extracting means include moving objects other than vehicles. Therefore, the vehicle determining unit determines whether the moving object is a vehicle so that only the vehicle is detected from the various moving objects by the vehicle determining unit. Here, in the present invention, a motorcycle, a bicycle with a motor, a bicycle, and a human (pedestrian) are not detected. In other words, a motorcycle such as a motorcycle or a pedestrian is extracted as a moving object by the moving object extracting means, but is not judged as a vehicle by the vehicle judging means and is excluded from the detection target.

In order to determine whether a moving object is a vehicle, the size of the moving object may be used as a criterion. That is,
If the moving object has an area equal to or larger than the predetermined area, the moving object is determined to be a vehicle. Conversely, if the moving object has an area smaller than the predetermined area, the moving object is not determined to be a vehicle. . As the size determination target, not only the area but also the length, width, diagonal length, and the like of the moving object can be used.

The moving body determined to be a vehicle by the vehicle determination means determines whether or not it has reached a preset approach determination position by the approach determination means, and stores the information if it has reached. This entry determination position is set on the road in front of the intersection in the vehicle traveling direction. However, if a crosswalk is provided at the intersection, it is set in front of the crosswalk in the vehicle traveling direction. The reason will be described later.

The vehicle tracking means tracks the vehicle stored as having reached the preset entry determination position by the entry determination means, and then determines whether the tracked vehicle has reached the preset pass determination position. Is determined by the passage determination means. The passage determination position is set at a position closer to the intersection than the entry determination position. Therefore, the vehicle that passes through the intersection travels and passes through the entry determination position, the passage determination position, and the intersection in this order, but when the vehicle reaches the passage determination position, it is determined that the vehicle has entered the intersection.

[0013] As described above, by performing vehicle passage determination at two locations, the entry determination position and the passage determination position, accurate vehicle detection can be performed by eliminating the effects of objects to be excluded from detection targets such as pedestrians. Can be. In other words, the two-wheeled vehicle and the pedestrian are not individually determined to be vehicles by the vehicle determination unit, but, for example, a group of pedestrians may be determined to be a moving object by the moving object extractor, and are determined to be a moving object. If the group has a size equal to or larger than the size of the vehicle, the vehicle is erroneously determined to be a vehicle by the vehicle determination unit. Therefore, in order to truly detect only a vehicle, a moving object erroneously determined to be a vehicle must be excluded from detection targets.

For example, if an attempt is made to determine the approach to an intersection only at the position immediately before the intersection, and if there is a pedestrian crossing there, a group of pedestrians at or near the pedestrian crossing will be sequentially moved by a moving body and a vehicle It is determined that the vehicle has entered the intersection. However, it is not usually possible for a group of pedestrians to exist at a position that is somewhat distant from the intersection, so if it is determined at such a position (the entry determination position of the present invention), the group of pedestrians A moving object erroneously determined as a vehicle based on the vehicle can be excluded from detection targets. Therefore, when a pedestrian crossing is provided at the intersection, the entry determination position is set to the vehicle traveling direction before the pedestrian crossing.

The vehicle that has reached the entry determination position is tracked, and when the tracked vehicle reaches a passage determination position set closer to the intersection, it is determined that the vehicle has entered the intersection. It is possible to accurately detect a vehicle entering an intersection without being affected by a person or the like.

If the passage determining means detects that the vehicle has entered the intersection and outputs this information, and this output is counted by the number counting means,
The number of vehicles that have entered the intersection can be ascertained.
Further, a signal state monitoring means for monitoring the state of the traffic light (red, blue, yellow) is provided, and if the output from the passage determination means is compared with the state of the signal, a vehicle ignoring the signal can be detected.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below based on embodiments with reference to the drawings. FIG. 1 is a block diagram showing a basic configuration of an approaching vehicle detection device according to the present invention, and FIG.
It comprises a camera 1, an I / F unit 10, an image processing unit 20, and a control unit 30.

The TV camera 1 captures an image of the entire intersection, and
The F unit 10 inputs the image. Image processing unit 20
Based on the image data from the / F section 10, a series of processes from extraction of the moving object to determination of passage are performed. The control unit 30 controls the entire vehicle detection device. When the control unit 30 causes the present device to function as a device for counting the number of vehicles entering an intersection, the control unit 30 obtains information on vehicles entering the intersection from the image processing unit 20 and counts the information. When functioning as a signal-ignoring vehicle detection device, the state of the signal and the image processing unit 2
By comparing the result with 0, it is determined that the vehicle ignores the signal.

FIG. 2 is a block diagram showing an embodiment of a vehicle detecting device for counting the number of vehicles entering an intersection. As shown in the figure, the vehicle detecting device of the present embodiment includes a TV camera (imaging means) 1, a moving object extractor 2, a vehicle determiner 3, an approach determiner 4, a vehicle tracker 5, and a pass determiner 6.
The moving object extractor 2, the vehicle judging device 3, the approach judging device 4, the vehicle tracking device 5, and the passage judging device 6 are provided with the image processing section 20 of FIG. The device 7 constitutes the control unit 30. The TV camera 1 is for capturing an image of a vehicle entering the intersection from the rear in the traveling direction, including the entire intersection. The image obtained here is sent to the moving object extractor 2.

The moving body extractor 2 calculates a difference between an image sequentially transmitted from the TV camera 1 and a background image of an intersection imaged and stored by the TV camera 1 in a state where no vehicle exists in advance. Find the part where the brightness changes. Since the portion where the brightness changes significantly is the image portion of the moving object and the portion where the brightness does not change much is the background image, the portion where the brightness changes greatly is extracted as the moving object region A2. The concept of the moving object area A2 is shown in FIG. 4, and is extracted as an area larger than the area occupied by actual vehicles or people.

The extraction of the moving body region A2 is performed in a preset approaching vehicle processing range A1. As shown in FIG. 3, the approaching vehicle processing range A1 is set near an intersection of a lane in which a vehicle is to be detected. Incoming vehicle processing range A
The reason for setting 1 is to reduce the computational load on the hardware of the detection device by performing processing only on the lane to be detected, and to improve the accuracy of detection by limiting the vehicles to be detected. is there. In addition, the moving body area A2
It is desirable that the entire vehicle be located within the approach vehicle processing range A1, but the present invention is not limited to this.

The vehicle judging device 3 judges whether the moving object area A2 extracted by the moving object extracting device 2 is based on a vehicle or not (for example, a bicycle or a person). If so, the moving body area A2 is determined to be the vehicle area A3. Whether the vehicle is a vehicle or not is determined by the moving body area A2
Is determined based on whether or not the size (area) is equal to or larger than a predetermined value A set in advance. That is, when the area A2 of the extracted moving body region is A ≦ A2, it is determined that the vehicle region is the vehicle region A3,
If A> A2, it is not determined to be the vehicle area A3. In addition,
The determination criterion for the vehicle region A3 may be based not only on the area of the moving body region A2 but also on the length, width, diagonal length, and the like as described above.

Since the size of an ordinary car and that of a minicar are different from each other, it is necessary to take this point into account when setting a criterion. That is, the criterion is set so that the moving object region based on the mini vehicle can be determined as the vehicle region. More specifically, when the determination is made based on the area, as shown in FIG.
If about / 3 is within the approaching vehicle processing range A1 (shaded area in the figure), a criterion is set so that the moving body area A2 is determined to be the vehicle area A3.

The approach determiner 4 determines whether or not the vehicle area determined by the vehicle determiner 3 is a vehicle approaching an intersection, and stores it. In other words, the vehicle entering the intersection is shown in FIG.
Is entered from the a side of the approaching vehicle processing range and passes to the b side. If the vehicle area determined by the vehicle determiner 3 includes the approach determination line L1 as shown in FIG. And memorize.

For example, a group of pedestrians is determined by the moving body extractor 2 to be a moving body area A2. If the moving body area A2 has a predetermined size, the vehicle deciding section 3 determines whether the moving body area A2 has a vehicle area A3. It is determined, but this is erroneously determined. However, if a group of pedestrians appears, it is on or near the pedestrian crossing of the intersection. Therefore, the moving object area A2 based on the group of pedestrians is determined by the vehicle determiner 3 in the vehicle area A.
Even if it is determined that the vehicle is 3, the vehicle does not normally include the approach determination line L <b> 1, and is not stored as an approaching vehicle by the approach determiner 4.

The vehicle tracker 5 tracks the vehicle stored in the approach determiner 4 by the following method. That is, FIG.
As shown in the figure, the vehicle area A3 stored in the entry determiner 4
If the overlap (shaded area in the figure) of the vehicle area A32 extracted by the vehicle determiner 3 after 1 and the sampling time Δt is a certain ratio or more and the movement vector V of the vehicle areas A31 and A32 is not downward, the same. The storage of the vehicle area A31 is updated to A32 assuming that the vehicle is traveling. Thereafter, the same process is repeated at intervals of Δt between the stored vehicle region and the extracted vehicle region to track the vehicle.

If the sampling time Δt is set to 0.1 second, a speed of about 140 km / h is required for the vehicle to travel the length (about 4 m). When approaching, the vehicle area always overlaps. Therefore, by examining the overlap of the vehicle regions, it is possible to confirm whether or not the vehicle regions are based on the same vehicle.

The pass determination unit 6 determines whether or not the vehicle area A3 tracked by the vehicle tracker 5 has reached the pass determination line L2 on the approaching vehicle processing range A1. In FIG. 7, if the vehicle area A3 has reached the passage determination line L2 on the approaching vehicle processing range A1, it is determined that the vehicle has entered the intersection and an approach signal is output, and the tracking is terminated.

In the present embodiment, the pass judgment unit 6
If the vehicle area A3 has reached the passage determination line L2 of the approaching vehicle processing range A1, it is determined that the vehicle has entered the intersection, and the tracking ends, that is, the processing ends.
When it is desired to obtain information indicating in which direction of right or left turn the vehicle has passed and passed the intersection, the TV camera 1 has taken an image of the entire intersection, so that the processing may be continued as appropriate.

The number counter 7 counts the number of vehicles entering and exiting from the passage determiner 6, and counts the number of vehicles entering the intersection.

According to the vehicle detection device described above, the number of vehicles can be detected and counted when the vehicle enters the intersection without being affected by pedestrians or the like. Can be used as a guide. In addition, since the approaching vehicle processing range is set, the detection target can be narrowed down to only the lane in which the detection target is to be detected, so that the calculation load on the hardware of the detection device is reduced, and the target vehicle is accurately detected. Can be detected.

FIG. 8 is a block diagram showing a basic configuration of a vehicle detecting device for detecting a vehicle ignoring a signal at an intersection. As shown in the figure, the vehicle detection device of the present embodiment has a T
It comprises a V camera (imaging means) 1, a moving object extractor 2, a vehicle determiner 3, an entry determiner 4, a vehicle tracker 5, a passage determiner 6, a signal state monitor 8 and a signal disregard determiner 9. The moving object extractor 2, the vehicle determiner 3, the approach determiner 4, the vehicle tracker 5, and the pass determiner 6 are the image processing unit 20 of FIG.
And the signal state monitor 8 and the signal disregard determiner 9 constitute the control unit 30. This vehicle detecting device includes a signal state monitor 8 and a signal ignoring determiner 9 instead of the number counter 7 included in the vehicle detecting device that counts the number of vehicles entering the intersection.

Since the configuration and operation of the vehicle detection device from the TV camera 1 to the passage determination device 6 are the same as those of the above-described vehicle detection device, only the operation after the input / output from the passage determination device 6 will be described. . The signal state monitor 8 monitors the state of the traffic signal S, that is, the states of blue, yellow, and red, and outputs to the signal ignoring determiner 9 which state it is in.

The signal ignoring judgment unit 9 judges a vehicle ignoring the signal based on the traveling input / output from the passage judging unit 6 and the signal state output from the signal state monitoring unit 8. That is, if the output from the signal state monitor 8 indicates that the state of the traffic light is red when the traffic input / output is received from the passage determiner 6, the vehicle is determined to be a signal-ignoring vehicle. . Since the TV camera 1 captures an image of the entire intersection, the vehicle determined to be a signal-ignoring vehicle can be recorded as a still image, and the offending vehicle can be specified. Further, the determination result is output as a signal-ignoring vehicle entering signal, and a warning can be given to the signal-ignoring vehicle based on the output.

As described above, according to the present embodiment, by monitoring the state of the traffic light and comparing the state with the approaching vehicle signal, it is possible to detect a vehicle ignoring the signal and to specify the vehicle. This makes it easier to warn and control the offending vehicle.

[0036]

As described above, according to the present invention, the number of vehicles entering an intersection can be counted because the vehicles entering the intersection can be accurately detected. Therefore, it can be used as a measure of the traffic congestion at the intersection. Further, by detecting a vehicle entering an intersection and comparing the detected signal with the signal state, it is possible to specify a vehicle ignoring a signal, thereby facilitating warning and control of a violating vehicle.

[Brief description of the drawings]

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

FIG. 2 is a block diagram illustrating an embodiment of a vehicle detection device for counting the number of vehicles that have entered an intersection.

FIG. 3 is a diagram showing an approach vehicle processing range.

FIG. 4 is a diagram illustrating the concept of a moving object area.

FIG. 5 is a diagram showing a relationship between a vehicle area and an approaching vehicle processing range when the vehicle is determined to be an approaching vehicle.

FIG. 6 is a diagram illustrating a criterion for determining whether or not a vehicle is the same vehicle.

FIG. 7 is a diagram illustrating a relationship between a vehicle area and an approaching vehicle processing range when the vehicle is determined to be an intersection approaching vehicle.

FIG. 8 is a block diagram illustrating an embodiment of a vehicle detection device for detecting a vehicle ignoring a signal.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 TV camera 2 Moving object extractor 3 Vehicle judging device 4 Entry judging device 5 Vehicle tracking device 6 Passing judging device 7 Number counter 8 Signal status monitor 9 Signal disregard judging device A1 Incoming vehicle processing range A2 Moving object area A3 Vehicle area L1 approach judgment line L2 pass judgment line

Claims (3)

[Claims] An image pickup means for picking up an image of a road, a moving object extracting means for extracting a moving object from an image obtained by the image pickup means, and determining whether or not the moving object obtained by the moving object extracting means is a vehicle. Vehicle determination means for performing the determination; determining whether or not the moving object determined to be a vehicle by the vehicle determination means has reached a preset approach determination position; and reaching the approach determination position by the approach determination means. Vehicle tracking means for tracking a vehicle determined to be traveling, and pass determination means for determining whether the vehicle tracked by the vehicle tracking means has reached a preset pass determination position. Vehicle detection device. 2. The vehicle detection device according to claim 1, further comprising a number-of-entries counting unit that counts a passage determination output from the passage determination unit. 3. The vehicle detection device according to claim 1, further comprising signal ignoring determination means for determining a vehicle ignoring a signal based on a passage determination output from the passage determining means and a state of the traffic light.