JPH10253649A - Congestion monitoring system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a congestion monitoring system in which the load of an image processing operation for a congestion judgment is reduced, by which the costs of an image processing apparatus are lowered and which can miniaturize the image processing apparatus. SOLUTION: While a measuring line which is set on a road image is used as a reference, a distance is measured on the basis of a movement distance at a prescribed cycle regarding the feature amount of a vehicle which is closest to the measuring line. (S6). Then, a plurality of vehicle speeds which are obtained by a speed measuring means are averaged, and an average speed is computed. (S8). Then, the average speed is compared with a prescribed reference value, and a congestion judgment is made. (S 10). Since the speed is measured by the speed measuring means regarding the feature amount of the vehicle which is closest to the measuring line, it is not required to correlate individual vehicles on two road images at a prescribed cycle when a plurality of vehicles exist on a road image, and the load of an image processing operation is reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a traffic jam monitoring system, and more particularly to a system for monitoring traffic jam from road images.

[0002]

2. Description of the Related Art Conventionally, when monitoring traffic congestion on a road, an ultrasonic or loop coil type traffic counter is installed on the road to measure the number of vehicles passing and the speed. Also,
Recently, a video image of a road image is taken by a video camera, and the number of vehicles passing, the speed, and the like are calculated by image processing.

[0003]

With a conventional ultrasonic or loop coil type traffic counter, it is possible to know only the traffic condition of the cross section of the road where the counter is installed, and the two-dimensional traffic condition of the road can be obtained. There is a problem of not knowing. In addition, in a monitoring system using conventional image processing, a two-dimensional traffic condition of a road can be known. However, since the speed of each individual vehicle is calculated from the input road image, the individual vehicle must be recognized in real time in each image frame, and the individual vehicle must be associated between adjacent video frames. In addition, there is a problem that a high-speed processing capability is required and the processing apparatus becomes expensive and large.

[0004] The present invention has been made in view of the above points,
It is an object of the present invention to provide a traffic jam monitoring system that can reduce the load of image processing for traffic jam determination, can reduce the cost and size of an image processing device, and can execute processing other than traffic jam determination.

[0005]

According to the first aspect of the present invention, a feature amount of a vehicle is extracted from a road image obtained by imaging a road, and the characteristic amount of the vehicle is determined based on the measurement line set on the road image. Speed measuring means for measuring a speed from a moving distance in a predetermined cycle for a characteristic amount of a nearby vehicle; average speed calculating means for calculating an average speed by averaging a plurality of vehicles obtained by the speed measuring means; Traffic congestion determining means for comparing traffic congestion by comparing the average speed with a predetermined reference value;

As described above, the speed measuring means measures the speed of the characteristic amount of the vehicle closest to the measurement line and does not measure the speed of the characteristic amount of another vehicle far from the measurement line. When present, it is not necessary to associate individual vehicles with two road images having a predetermined cycle, and since the association is unnecessary, a simple feature such as a horizontal edge projection value can be used as a vehicle feature amount. Only the amount needs to be extracted, and the load of image processing is reduced.

According to a second aspect of the present invention, in the traffic congestion monitoring system according to the first aspect, a measurement line used by the speed measuring means is additionally set on the road image based on a result of the judgment by the congestion judging means. And measurement line setting means for measuring the speed of the characteristic amount of the vehicle closest to each measurement line. As described above, by setting additional measurement lines when a traffic jam is likely to occur or when a traffic jam occurs based on the traffic jam determination result, the load of image processing is somewhat heavy, but the traffic jam situation in the vicinity of each of the plurality of measurement lines is slightly increased. Can be monitored in detail.

[0008]

FIG. 2 is a block diagram showing an embodiment of the system of the present invention. In FIG. 1, a camera 10 captures an image of a road 12 in a direction in which a vehicle 14 moves away. The image signal of this road is supplied to the image processing device 16. The image processing device 16 extracts the feature amount of the vehicle from each image frame, measures the speed of the vehicle from the feature amount, obtains the average speed of the plurality of vehicles, and determines whether the road is congested based on the average speed, or Determine whether it is crowded or natural. The traffic condition determination result is supplied from the communication device 18 to the monitoring center 22 through the network 20 together with the image signal obtained by the camera 10.

The monitoring center 22 is supplied with road condition determination signals and image signals from a plurality of other image processing devices to centrally monitor the traffic conditions at each point. FIG. 1 shows a flowchart of a process executed by the image processing device 16. In the figure,
In the initial setting of step S2, a distance table for associating the vertical position on the screen with the distance on the road in the screen is created, and the processing area is set for each lane so that processing is transmitted in units of road lanes. I do. In step S4, the video frame is always compared with the immediately preceding video frame. If there is a difference in the luminance value at the same pixel position, the luminance value at the same pixel position on the average background is + a (for example, a = 1 the luminance of the current video frame). (When the value is large) or -a (for example, when a = 1 the luminance value of the current video frame is small), and the average background is created and updated.

Next, step S corresponding to the speed measuring means
In step 6, a speed measurement process is performed to measure the speed of a specific vehicle on the screen. Next, step S corresponding to the average speed calculation means
In step 8, an average speed calculation process is performed to calculate an average speed of a plurality of vehicles. Thereafter, step S corresponding to the congestion determination means
At 10, traffic jam determination processing is performed, and the determination result is notified to the monitoring center. Step S corresponding to the measurement line setting means
At 12, a measurement line setting process is performed, and one or more measurement lines are set on the screen according to the result of the traffic jam determination.
Thereafter, the process proceeds to step S4, and steps S4 to S12 are repeated.

FIG. 3 shows a flowchart of the speed measurement process. In the figure, in step S20, a predetermined image processing cycle Δ
At t, a road image is captured, and at step S22, a feature amount of the vehicle is extracted. Here, a difference between the captured road image and the average background is obtained, and a horizontal edge is extracted from the difference image. Then, in order to project the extracted horizontal edge on the vertical axis and remove noise, a predetermined threshold value (this threshold value is predetermined based on the width of the vehicle passing on the projection surface) among the obtained projection values is set. A projection value exceeding the value is extracted as a feature value. For example, FIG. 4 (A), FIG. 5 (A), FIG. 6 (A)
In the road images shown respectively, FIGS. 4B, 5B, and 6
(B) The horizontal edge and the feature amount shown respectively are extracted.

Next, in step S24, it is determined whether or not the characteristic amount of the vehicle has been extracted. If not, the elapsed time from the start of the first speed measurement processing in step S26, or the characteristic amount of the previous time. It is determined whether or not the elapsed time from the extraction is equal to or longer than a predetermined time M (M is, for example, 10 sec). Here, when the elapsed time <M, the processing is terminated. Elapsed time ≧
In the case of M, the process proceeds to step S28, and since there is no passing vehicle, the speed is determined to be a natural flow (no congestion), and the process ends.

If it is determined in step S24 that the characteristic amount has been extracted, the process proceeds to step S30, where the vehicle travels from the predetermined measurement line L (upper side of the vertical axis).
The feature amount _Pn closest to the measurement line L is extracted. Next, in step S32, the distance _Xn from the measurement line L to the extracted feature value _Pn is calculated using the distance table. Thereafter, it is determined whether the feature quantity P _n-1 extracted in image frames processed last time in step S34, the feature amount P _n-1
If there is no, the process ends, and if there is the feature amount P _n−1 , the process proceeds to step S36.

[0014] Step S36 characteristic amount P _n-1 in which the distance X _n of feature amounts P _n obtained in this process was obtained by the previous process
It is determined whether the distance is _{equal to} or more than the distance X _n-1 . If X _n <X _n−1, the process is terminated _assuming that the feature amount P _n is a different vehicle from the feature amount P _n−1 . For example, when each of the road images shown in FIGS. 4A, 5A, and 6A is processed in this order, a case where the state of FIG. 5A has advanced to the state of FIG. In the figure, the vehicle indicated by A moves away from the vehicle indicated by B, which corresponds to the above case.

On the other hand, if X _n ≧ X _n−1 , it is determined that the feature amounts P _n and P _n−1 are the same vehicle, and the process proceeds to step S38. The case where the state has progressed from the state of FIG. 4A to the state of FIG. 5A corresponds to the above case. Step S38 using an image processing cycle Δt in the processing ends by calculating the velocity V _n of the vehicle by the following equation.

[0016]

(Equation 1)

As described above, in the speed measurement process S6, the speed is measured for the characteristic amount of the vehicle closest to the measurement line L.
Since the speed is not measured for the feature amount of another vehicle far from the measurement line L, there is no need to associate individual vehicles with two road images having a predetermined cycle when a plurality of vehicles exist in the road image. Further, since the association is unnecessary, it is only necessary to extract a simple feature amount such as a projection value of a horizontal edge as a feature amount of the vehicle, and the load of image processing is reduced.

FIG. 7 shows a flowchart of the average speed calculation process. In the figure, a time T _n at which the speed V _n was calculated in step S40 and a time T _{n-1 at} which the speed V _n-1 was calculated before the time T _n
Is used to determine whether or not the elapsed time is less than or equal to a predetermined time M. T _n -T _n-1> M multiple times within a predetermined time M case, since the vehicle speed is not obtained, the counter C at step S42 for the first vehicle speed for the moving average vehicle speed V _n 1 sets, and ends the process assigns the value of a counter C to the speed V _n.

Meanwhile, in the case of T _n -T _n-1 ≦ M assigned the value of the counter C to the vehicle speed V _n proceeds to step S44.
Next, in step S46, it is determined whether or not the value of the counter C is equal to or more than a predetermined value m (m is, for example, 49). If C <m, the value of the counter C is incremented by 1 in step S48, and the process ends. . If C ≧ m, the process proceeds to step S50, and an average speed V _av is calculated from the following formula from m + 1 speeds from the speed V _n to the speed V _nm .

[0020]

(Equation 2)

Thereafter, in step S52, the value of the counter C is decremented by a predetermined value α (α is, for example, 10), and the process is terminated. That is, the moving average is obtained by the equation (2). In this step S52, the value of the counter C may be set to 1. FIG. 8 shows a flowchart of the traffic congestion determination process. In the figure, in step S60, it is determined whether or not the average speed V _av is equal to or lower than the speed V _A (V _A is, for example, 20 km / h). If V _av ≦ V _A , it is determined in step S62 that there is heavy traffic. If V _av > V _A, the process proceeds to step S64, where the average speed V _av is equal to the speed V _B (V _B is, for example, 40 km /
h) It is determined whether or not the following conditions are satisfied, and if V _av ≦ V _B , it is determined in step S66 that there is light traffic. If V _av > V _B, the process proceeds to step S68, and it is determined whether or not the average speed V _av is equal to or lower than the speed V _C (V _C is, for example, 60 km / h). If V _av ≦ V _C , congestion is performed in step S70. Is determined. Also, V _av > V _C
In step S72, it is determined that the flow is natural (no congestion). When any of the above steps S62, S66, S70, and S72 is determined, the process ends.

FIG. 9 shows a flowchart of the measurement line setting process. In the figure, at step S80, it is determined whether or not heavy traffic or light traffic is determined, and in the case of traffic, measurement lines L 'and L "are set in addition to the measurement line L at step S82. ', L' are respectively set at predetermined positions on the vertical axis above the measurement line L as shown in FIG. If there is no congestion in step S80, step S
It is determined at 84 whether or not congestion has occurred. If congestion has occurred, at step S86, the measurement line L "as well as the measurement line L" is determined.
Set. If it is not congested, only the established line L is set in step S88, and the process ends.

When the measurement lines L 'and L "are set in addition to the measurement line L, steps S6, S8, S
In each process of 10, the speed measurement, the average speed calculation, and the traffic congestion determination are performed for each of the measurement lines L 'and L "similarly to the measurement line L. Thus, for example, when the traffic congestion occurs, the end position of the traffic congestion is measured by the measurement line L' or L. "Can be determined in detail. Of course, measurement lines L ', L "
Although the load of image processing is slightly heavier if the additional setting is made, it is important to know the detailed situation such as a congested state where traffic congestion is likely to occur and where the end position of the traffic congestion is when traffic congestion is occurring is there.

As described above, in the present embodiment, the load of image processing for judging traffic congestion can be reduced, so that the cost and size of the image processing apparatus 16 can be reduced. Further, the image processing device 16 can execute other processes such as a falling object detection process, a stopped vehicle detection process, a vehicle fire detection process, and an abnormally running vehicle detection process.

[0025]

As described above, the first aspect of the present invention provides
Speed measurement for extracting a feature amount of a vehicle from a road image obtained by imaging a road and measuring a speed from a moving distance in a predetermined cycle with respect to a feature amount of the vehicle closest to the measurement line based on a measurement line set on the road image. Means, average speed calculating means for calculating an average speed by averaging the speeds of the plurality of vehicles obtained by the speed measuring means, and traffic congestion determining means for performing traffic congestion judgment by comparing the average speed with a predetermined reference value And

As described above, the speed measuring means measures the speed of the characteristic amount of the vehicle closest to the measurement line, and does not measure the speed of the characteristic amount of another vehicle far from the measurement line. When present, it is not necessary to associate individual vehicles with two road images having a predetermined cycle, and since the association is unnecessary, a simple feature such as a horizontal edge projection value can be used as a vehicle feature amount. Only the amount needs to be extracted, and the load of image processing is reduced.

[0027] The invention described in claim 2 is the same as that in claim 1.
In the traffic congestion monitoring system described above, a measurement line used by the speed measurement means is additionally set on the road image based on the determination result of the traffic congestion determination means, and the speed is determined for the characteristic amount of the vehicle closest to each measurement line. It has a measurement line setting means for measuring. As described above, by setting additional measurement lines when a traffic jam is likely to occur or when a traffic jam occurs based on the traffic jam determination result, the load of image processing is somewhat heavy, but the traffic jam situation in the vicinity of each of the plurality of measurement lines is slightly increased. Can be monitored in detail.

[Brief description of the drawings]

FIG. 1 is a flowchart of a main process of the system of the present invention.

FIG. 2 is a block diagram of the system of the present invention.

FIG. 3 is a flowchart of a speed measurement process.

FIG. 4 is a diagram for explaining the present invention.

FIG. 5 is a diagram for explaining the present invention.

FIG. 6 is a diagram for explaining the present invention.

FIG. 7 is a flowchart of an average speed calculation process.

FIG. 8 is a flowchart of a congestion determination process.

FIG. 9 is a flowchart of a measurement line setting process.

FIG. 10 is a diagram for explaining the present invention.

[Explanation of symbols]

 Reference Signs List 10 camera 12 road 14 vehicle 16 image processing device 18 communication device 20 network 22 monitoring center S6 speed measurement process

Claims (2)

[Claims] 1. A feature amount of a vehicle is extracted from a road image obtained by imaging a road, and a speed of a feature amount of a vehicle closest to the measurement line is calculated based on a measurement line set on the road image from a moving distance in a predetermined cycle. Speed measuring means for measuring the speed of the vehicle, average speed calculating means for calculating an average speed by averaging the speeds of the plurality of vehicles obtained by the speed measuring means, and determining the congestion by comparing the average speed with a predetermined reference value. And a traffic congestion determining means for performing traffic congestion. 2. The traffic congestion monitoring system according to claim 1, wherein a measurement line used by the speed measuring means is additionally set on the road image based on a result of the judgment by the traffic congestion judging means. A traffic congestion monitoring system comprising a measurement line setting means for measuring a speed of a characteristic amount of a nearby vehicle.