JP2004171289A - Road traffic measuring device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a controller, a control program and a control method for making it unnecessary to synchronize the time, and flexibly executing the control of equipment, and easily controlling and managing the equipment by simultaneously inputting a plurality of camera videos. <P>SOLUTION: The video of a camera which monitors a signal and the video of a camera which measures a traffic flow are simultaneously fetched, and the video of the signal is compounded with the measured image of the traffic flow, and outputted so that a video synchronizing the signal can be prepared. Also, the compounded image is processed so that the current condition of the signal can be automatically recognized, and simultaneously the traffic flow is measured so that the traffic flow synchronizing with the signal can be measured. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a monitoring system using a television camera applied to a road or the like, and more particularly to a traffic measurement device that recognizes a lighting state of a signal with a camera and measures a traffic flow in synchronization with the state.
[0002]
[Prior art]
For conducting traffic flow simulations for investigating the causes of traffic congestion and traffic planning, the Traffic Engineering Study Group H7.5 published a manual on road traffic capacity research manuals, Vol. 2 (PP.14-PP.22), the saturated traffic flow rate at the intersection (the number of stops passing through the blue line per hour), the traffic volume, and the headway time (for the first vehicle The measurement of the transit time from when the head passes the stop line to when the second vehicle passes the stop line) is manually measured at the site, or once VTR shooting is performed at the site, and the playback screen is displayed later. I read the timer and hit it on my PC.
[0003]
In any of the methods, since it is necessary to grasp the passage of the vehicle in conjunction with a change in a traffic light or a stopwatch, a lot of time and labor is required for the measurement work. In addition, there is a possibility that human error may intervene due to the long-time monotonous work. As a result, the reliability of data may be reduced.
[0004]
Recently, a tentative name of Koshi Watcher Proto (http: //www.trpt.cst.nihon-u), which performs a measurement work by pressing a twelve buttons with a built-in timer and in which measurement items are given in advance by a researcher, is used. .Ac.jp / ROADDATA / research / koshi-watcher / koshi-w0.html), an attempt has been made to upgrade the measuring device, but there remains a problem that practice is required until the user fully uses the device, and a mistake is likely to occur due to manual operation.
[0005]
Therefore, it is conceivable to perform measurement by using a system that measures a traffic flow such as an intersection using image processing. The technology for obtaining the number of vehicles, the speed, and the like by image processing has advanced considerably technically, and it is possible to measure in real time with an accuracy of 90% or more. For details of the tracking process of a vehicle or the like, a method using template matching is described in JP-A-6-30417.
[0006]
[Patent Document 1]
JP-A-6-30417 [Non-Patent Document 1]
http: // www. trpt. cst. nihon-u. ac. jp / ROADDATA / research / koshi-watcher / koshi-w0. html
[0007]
[Problems to be solved by the invention]
To measure the saturated traffic flow rate, measure the number of vehicles that can stay through the stop line during the red lighting time of the traffic light and pass during the blue lighting time for each signal cycle, and total the blue lighting time It is necessary to repeatedly measure until the cumulative value corresponds to one hour. In general, it is conceivable that one camera shoots the traffic light and the flow of the vehicle at the same time so that the flow of the traffic light and the vehicle in the video is processed using this video.
[0008]
However, depending on the location, it is often difficult to automatically recognize the lighting state because the flow of the traffic light is far from the traffic signal or the signal is very small even if it is reflected. For this reason, it was difficult to measure traffic data in synchronization with a traffic light.
[0009]
[Means for Solving the Problems]
In order to solve the above-mentioned problems, the present invention is to input an image of a camera (signal camera) for photographing a traffic light and an image of a camera (traffic flow camera) for measuring a traffic flow to an image processing device, and to convert the image of the traffic flow camera into an image. The image of the traffic light captured by the traffic light camera is pasted, and the result of the image synthesis processing is output as a video, and the recognition of the lighting state of the traffic light and the recognition of the traffic flow are executed. In addition, various data are measured and totaled according to the present state of the traffic light.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
Embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram illustrating a configuration of the traffic measurement device. The respective camera images are input from a camera 1 for photographing the traffic flow measurement area and a camera 2 for photographing the traffic signal, and an image processing device 3 for performing image processing is connected to a central processing device 5 via a network 6. .
[0011]
The image processing device 3 includes an image input unit 301 that inputs the images of the cameras 1 and 2 and combines the two images, a signal recognition unit 302 that processes an image of a traffic signal in the image based on the combined image, and , A traffic flow recognition unit 303 that measures an image of a traffic flow in the image, and a communication unit 304 that communicates a recognition result and parameters required for recognition with the central processing unit 5. The image processing device 3 can output an image obtained by synthesizing the images of the cameras 1 and 2, and by connecting this image to a recording device 4 such as a VTR, the synthesized image can be recorded.
[0012]
The central processing unit 5 is used for setting parameters and totalizing the measurement results, but a notebook PC or PDA may be used so that the work can be easily performed outdoors. If the image processing apparatus 3 is provided with a medium for storing the measurement results, it is not always necessary to connect the tally results to the network 6. However, it is necessary to connect a keyboard or a display to the image processing apparatus 3 for parameter setting, but it is desirable to connect a PDA or a notebook PC for portable measurement. This is because the device configuration of the image processing apparatus 3 can be minimized. The network 6 may be a wired or wireless LAN.
[0013]
FIG. 2 is a diagram for explaining the arrangement and the field of view of the camera. The camera 1 has a field of view 8 for measuring a required traffic flow, and the camera 2 has a field of view 7 near a traffic light. Generally, the field of view of a traffic flow measurement camera captures an image at a depth of about 100 m to 150 m. For this reason, the image of the traffic signal is often not displayed at the same time. Therefore, only the traffic signal is photographed by the camera 2. By doing so, the images of the two cameras are combined in advance and recorded on a VTR or the like using the image processing device, and the combined images are processed to measure the traffic flow in accordance with the state of the traffic light. Is no longer necessary.
[0014]
Otherwise, it is necessary to record each video on a VTR or the like, then process each video offline, write the GPS time data to the video in advance, and adjust the time between the videos.
[0015]
FIG. 3 shows a configuration in a case where the distance between the camera 1 (2) and the camera 2 (1) is large and one camera is wirelessly transmitted. The video of the camera 2 (1) is temporarily input to the wireless video transmission device 9, the video from the wireless video transmission device 9 is received by the wireless video reception device 10, and the video is input to the image processing device 3. By doing so, it is possible to create a composite video even when the traffic signal and the traffic flow measurement area are far apart. As the wireless video transmission / reception devices 9 and 10, a commercially available 2.5 GHz simple transmission device or the like can be used.
[0016]
FIG. 4 is a diagram for explaining a composite state of video. The image of the camera 1 is an image of a traffic flow measurement area, and the field of view is relatively wide, about 100 m11. On the other hand, the image of the camera 2 shooting the traffic light is taken so that the traffic light can be seen clearly, and the field of view is relatively narrow 12. A video obtained by synthesizing a signal area image with the measurement image 11 (for example, cutting out and pasting a traffic light portion) becomes a synthesized image 13. The composite image 13 is output from a video output unit of the image processing device 3 or a video output unit connected to the image processing device 3 (any video output unit is not shown), and the traffic flow can be measured in real time. It is. In addition, the video can be recorded on the VTR 4, so that the video can be taken back to an office or the like and processed again off-line.
[0017]
FIG. 5 is a diagram illustrating a processing flow of the entire traffic flow measurement system. First, an initialization process 15 is performed. In this processing, there are a traffic light area, a traffic light writing area setting, a detection area setting for traffic flow measurement, a lane setting, and the like. When the central processing unit 5 is connected, the central processing unit 5 receives the parameters from the central processing unit 5 as parameters.
[0018]
Next, the images of the cameras 1 and 2 are inputted 16 and these inputted images are synthesized (for example, a traffic light area is cut out from a traffic light photographed image and pasted on a measurement image) 17. Output the synthesized video 18. The vehicle detection processing 19 is executed for the detection area, and when the detection is performed, the vehicle is tracked 20. This process is repeated 21 for the number of lanes. Next, the lighting state of the traffic light is recognized 22 and the measurement result is output 23. Step 16 and subsequent steps are repeated until the processing is completed. There may be a case where the processing from the processing 19 to the processing 23 is not performed and only the synthetic image is created.
[0019]
FIG. 6 is an example of a user interface screen for setting a combination parameter when combining images. When the image input button is pressed, the image of the camera connected to the image processing device 3 is obtained via the LAN, and the measurement image is displayed on the upper left and the traffic light image is displayed on the upper right. Using the mouse, specify the signal area on the signal image and press the signal area setting button. Further, a writing area is specified on the measurement screen using a mouse, and a writing area setting button is pressed. When the confirmation button is pressed, the composite image is displayed below. The area is set by such a mechanism.
[0020]
When the display screen is small like a PDA, the measurement image, the traffic light image, and the composite image are simultaneously displayed on the screen, and the enlarged screen is called up and displayed by tapping each screen. By doing so, the operability can be improved.
[0021]
FIG. 7 is an example of a user interface screen for associating camera coordinates (camera parameters) with world coordinates (world parameters), in which camera parameters (depth information) can be set. By specifying in advance four locations whose distances from the camera are known, the relationship between the camera coordinate system and the world coordinate system is defined as perspective transformation.
[0022]
FIG. 8 is an example of an interface screen for setting a detection area and a lane boundary. A vehicle detection area required for traffic flow measurement can be set, and a lane boundary can be set. Image processing is executed according to the contents set on this screen.
[0023]
FIG. 9 is a diagram showing an outline of the tracking processing, and shows a state of vehicle detection and vehicle tracking in the left lane in the figure. Reference numeral 25 denotes a detection area on the left lane in the figure, and reference numeral 26 denotes a detection area on the right lane in the figure. When the vehicle passes through the 25 detection areas, the end of the vehicle is detected, and the video is stored as a template image as 27. Since the vehicle should have moved to the vicinity of the stored template image 27 at the next time, a search area 28 is set, and a place most similar to the template image is searched in that area. As a result, the position of 29 is obtained. At the same time as storing the position 29, the image near this position is updated as a template image. By sequentially repeating such processing, the vehicle tracking processing can be executed. In the case of the right lane, the setting is for traveling from the back to the front in the figure. Similar processing is executed by detecting an image of the front part of the vehicle and registering the image.
[0024]
FIG. 10 is a diagram illustrating a recognition flow of signal presentation. It is determined whether there is a signal area setting, and if so, the area setting is performed 31. First, a projection distribution of density is obtained for the area 32. Among these distributions, the horizontal position of the maximum frequency distribution is determined 33, and the determination is made on the left, middle, and right sides as "blue", "yellow", and "red" 34 to 37. This processing focuses on the fact that the lit lamp is the brightest, instead of the color processing, but may be executed by the color processing.
[0025]
FIG. 11 is a diagram illustrating a measurement result output flow. When the vehicle is tracked, the number of vehicles is counted at a timing when the vehicle has passed a predetermined position. At this timing, the time is first output 39, and if the state of the previous signal has changed from the state of the current signal, the number and the average speed that have passed in the signal cycle up to that time are calculated and output 41. It also outputs the state of the signal 42. After that, the speed of the individual vehicle and the inter-vehicle time with the preceding vehicle are output. FIG. 12 shows an example of this output.
[0026]
FIG. 13 is a diagram illustrating details of the image processing apparatus 3. The image processing device 3 includes A / D converters 52 and 53 for inputting the images of the cameras 1 and 2, a plurality of image memories 54 for storing the images, and a dedicated image processor 55 for image-processing the data in the image memories. A D / A converter 56 for outputting data of the image memory as a video, a CPU 59 for operating these hardware, a RAM 60 for storing a program for describing the operation, and a communication unit 58 for communicating with an external device. It is comprised including. Note that the image memory 54 and the RAM 60 may be a common memory. The A / D converter of the camera input can simultaneously input to the image memory.
[0027]
As described above, a video image of a camera that captures a road with a wide field of view and a video camera that captures a traffic light with a predetermined size is input, and a composite image is created by combining the traffic signal image and the image to be measured, and the video is created. And outputs it to a VTR, etc., measures traffic flow and recognizes traffic signals, and outputs the measurement results in accordance with the status of the traffic signals, making it possible to measure traffic conditions at intersections in a portable manner. .
[0028]
Next, another embodiment will be described.
[0029]
Road traffic parameters include the signal cycle (time when the signal goes from blue to yellow to red to blue), split (main road side and sub road side, time distribution of blue, yellow and red, respectively), offset (two adjacent When calculating the offset of the traffic signal at the intersections, the time difference from the start of the green signal of the upstream signal A to the start of the green signal of the downstream signal B is an important value. In particular, the offset is an important parameter for the traveling vehicle to flow smoothly. Such parameters have setting information in the traffic light itself, but it is not usually possible to obtain this information.
[0030]
For this reason, as shown in FIG. 14, the offset from the signal at the distant intersection is currently measured manually using a stopwatch or the like. However, there is a problem that the error is large and it is difficult to measure for a long time. FIG. 14 is a diagram illustrating a process of synthesizing images of a plurality of traffic signals.
[0031]
Therefore, referring to FIG. 15, which is a diagram for explaining an interface for synthesizing a plurality of traffic lights, a synthesized image 68 obtained by synthesizing the video of the traffic light 1 image 66 and the traffic light 2 image 67 is obtained, and the lighting of each signal is performed. By measuring the state by image processing, the cycle, split, and offset of each signal can be measured.
[0032]
FIG. 16 shows an example of an interface for setting a traffic light area, in which a traffic light area of an image of each camera is designated and synthesized. Although there are various distances at the intersection, when the next signal can be seen from one place, it can be enlarged and photographed by zooming of the camera. When it cannot be seen, the camera 62 and 63 as shown in FIG. What is necessary is to take a picture in the vicinity of the point, send the video by wireless transmission or the like, and combine them.
[0033]
Accordingly, a traffic flow measuring device capable of automatically measuring a vehicle flow such as a saturated traffic flow rate at an intersection or collecting images is provided.
[0034]
As described above, the traffic flow measurement area and the traffic light are photographed by the two camera images, and these are combined, and the images are processed to match the present state of the traffic light, and the traffic flow (saturated traffic flow rate) , Number, inter-vehicle distance, inter-vehicle time, etc.), which is very useful for investigating the causes of traffic congestion. Further, by photographing adjacent traffic signals and synthesizing them, it is possible to easily measure the signal state of the intersection, that is, cycle, split, and offset.
[0035]
【The invention's effect】
According to the present invention, it is easy to measure traffic data in synchronization with a traffic light.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration of a traffic measurement device.
FIG. 2 is a diagram illustrating a field of view of a camera.
FIG. 3 is a diagram illustrating a configuration in a case where one camera is wirelessly transmitted.
FIG. 4 is a diagram illustrating a state of synthesizing a video.
FIG. 5 is a diagram showing an overall processing flow of the traffic flow measuring device.
FIG. 6 is a diagram showing an interface when synthesizing images.
FIG. 7 is a diagram illustrating an interface that associates camera coordinates with world coordinates.
FIG. 8 is a diagram showing an interface for designating a detection area and a lane boundary.
FIG. 9 is a diagram illustrating an outline of a tracking process.
FIG. 10 is a diagram illustrating a recognition flow of signal presentation.
FIG. 11 is a diagram illustrating a measurement result output flow.
FIG. 12 is a diagram illustrating an example of a measurement result output.
FIG. 13 is a diagram illustrating hardware details of an image processing apparatus.
FIG. 14 is a diagram illustrating a process of synthesizing images of a plurality of traffic signals.
FIG. 15 is a diagram illustrating an interface for combining a plurality of traffic lights.
FIG. 16 is a diagram illustrating an example of an interface for setting a traffic signal area.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Camera for traffic flow measurement, 2 ... Camera for traffic light photography, 3 ... Image processing apparatus, 5 ... Central processing apparatus.

Claims (4)

From a traffic light monitoring camera that captures a traffic light and a traffic flow monitoring camera that measures traffic flow, an image processing device to which an image is input is provided,
The image processing device includes:
A signal recognition unit that processes the traffic light monitoring camera image and measures the present state of the traffic light,
A traffic flow measuring unit that processes a video of the traffic flow monitoring camera to measure a traffic flow,
A traffic measurement device having a communication unit that outputs a traffic flow measurement result in accordance with the present state of the traffic light. In claim 1,
The traffic measurement device is characterized in that the image processing device includes an image processing unit that cuts out a traffic light area from an image of a traffic light monitoring camera and writes the signal area into a predetermined area of an image of a traffic flow measurement camera. Equipped with a plurality of traffic light monitoring cameras for shooting a plurality of traffic lights, and an image processing device for simultaneously inputting a plurality of videos,
The image processing device
A signal recognition unit for recognizing a signal area of a plurality of signal monitoring camera images and measuring the present state of the signal,
A traffic measurement device comprising: a signal parameter calculation unit that calculates at least three parameters of a cycle, a split, and an offset from a present state of a traffic signal. In claim 3,
The image processing device
An image synthesizing unit that cuts out each of the traffic light areas of the plurality of traffic light monitoring camera images and synthesizes them into one image,
The traffic measuring device, wherein the signal recognizing unit performs image processing on an image synthesized by the image synthesizing unit to measure the present state of the traffic light.

Images