JP2001126183A - Device for detecting traffic flow - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a traffic flow detecting device by which a small vehicle hidden behind a large vehicle is adequately discriminated and detected and also the vehicle detecting precision can be improved. SOLUTION: The device is provided with a plurality of IR cameras which are disposed in left and right at a prescribed interval along the sides of a traveling road, a plurality of vehicle detecting parts for inputting image data of the plurality of traveling vehicles, which are photographed by the IR cameras and an integration judging part for inputting a plurality kind of vehicle information which are processed by the vehicle detecting parts. The vehicle information processed by the judging part are integrated and also a cross-correlation arithmetic processing is executed concerning the respective kinds of data so that the vehicle is synthetically judged.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a traffic flow detecting device for discriminating and detecting a plurality of vehicles traveling on a road.

[0002]

2. Description of the Related Art FIG. 9 (a) shows a schematic configuration diagram of a conventional traffic flow detecting device. That is, as shown in FIG. 9 (a), conventionally, one visible camera 2 is arranged along the road side of the traveling road 1, and a plurality of vehicles traveling on the road are photographed by the visible camera 2, and the photographed image is taken. The vehicle is determined and the traffic volume is grasped by determining the data by a predetermined algorithm. In this case, when the ordinary vehicle 4 (small vehicle) is running in parallel with the large vehicle 3, it is difficult to distinguish between the two vehicles 3 and 4 by the so-called "separation method" algorithm determination. I was Briefly describing the algorithm determination of the “separation method”,
First, the captured background image data is generated, the background difference is binarized and shaped to detect the shape of the vehicle area, and then these are detected as a plurality of rectangular data (individual vehicle data), and then merged or deleted. The determination is made by image processing according to.

[0003]

However, the conventional traffic flow detecting device has the following problems. That is, when the ordinary vehicle 4 is completely hidden from the large vehicle 3 as shown in FIG. 9B, the vehicle type cannot be accurately and accurately grasped by the photographing conditions of the visible camera 2. There is a problem that the detection accuracy is reduced. In order to solve such a problem, it has been proposed to use two visible cameras instead of one visible camera. In this case, the detection accuracy can be improved more than one visible camera. Although it is possible, the error may not be reduced due to the influence of the shadow or the like due to the vehicle. Furthermore, since a visible camera is used as an imaging means of the vehicle, there is a risk that erroneous detection may occur due to a shadow of the vehicle itself or a building, and illumination is required at night. Further, in recent years, as the number of types of vehicles has increased along with the increase in traffic volume, it has been desired to realize a traffic flow detection device capable of further improving the accuracy of discriminating and detecting these vehicles.

Accordingly, an object of the present invention is to solve the problems of the conventional traffic flow detecting device, to accurately determine and detect even a vehicle concealed by a large vehicle, and to detect a plurality of vehicles traveling on a road. An object of the present invention is to provide a traffic flow detection device capable of further improving detection accuracy.

[0005]

According to the present invention, in order to achieve the above object, the invention according to claim 1 comprises a plurality of light emitting devices which are arranged on the left and right at predetermined intervals along a road side of a traveling road. A plurality of infrared cameras, a plurality of vehicle detection units to which image data of a plurality of traveling vehicles captured by the plurality of infrared cameras are input, and a plurality of vehicle information processed by each of the vehicle detection units. And an integrated determination unit that integrates the vehicle information processed by the integrated determination unit and performs a cross-correlation calculation process on each image data to perform a comprehensive determination of the vehicle. Things.

According to a second aspect of the present invention, in the first aspect of the invention, the vehicle detecting section has an A / D conversion section, a frame memory, a CPU, and a communication I / F, and the integrated determining section has a communication I / F. / F, a vehicle position determination unit, and an output I / F, wherein an infrared camera is used as an imaging unit.

According to a third aspect of the present invention, in the first aspect of the present invention, the integration determination processing unit processes two image data respectively photographed by two infrared cameras separately, and A magnification conversion process and a left-right inversion process are performed on one of the image data after the vehicle contour extraction process, and a cross-correlation calculation process is performed on the two image data to determine the identity of the vehicle. It is characterized in that it is determined.

According to a fourth aspect of the present invention, in the third aspect of the present invention, the image data is compared with image data processed based on vehicle speed and vehicle width information as vehicle information obtained by the integrated determination processing section. When these pieces of information do not match, the vehicle is not determined.

[0009]

1 and 2 show an embodiment of a road to which the present invention is applied. FIG. 1 shows a front view of a vehicle in a traveling direction, and FIG. 2 shows a plan view. It is assumed that the ordinary vehicle 4 adjacent to the large vehicle 3 is running in parallel. Further, as shown in FIG. 2, in the present invention, a predetermined interval (about 100 m) is provided along the road side of the traveling road.
(Four) infrared cameras 6 arranged on both sides
(Front left), 7 (front right), 8 (rear left) and 9 (rear right), and the plurality of infrared cameras 6 to 9 respectively control a plurality of vehicles traveling on the road from a position above the road. It is arranged so that you can shoot. The visual field position with respect to the vehicle is almost the same. In this example, an embodiment using four infrared cameras is shown, but it is assumed that at least two infrared cameras are required in the present invention.

FIG. 3 is an overall block diagram of a functional configuration according to an embodiment of the present invention. That is, as shown in FIG. 3, the traffic flow detecting device according to the present invention comprises four infrared cameras 6.
9. Image data of a plurality of traveling vehicles photographed from 9 to ９9 are input, and a single vehicle detection unit 10〜13 and a plurality of vehicle information processed by each of these vehicle detection units 10〜13 are integrated. And a determination unit 14. Among these, each of the vehicle detection units 10 to 13 includes an A / D conversion unit 16, a frame memory 17,
A PU 18 and a communication I / F 19;
/ F20, a vehicle position determination unit 21, and an output I / F22.

The outline of each function of this overall block diagram will be described. An image (image data) photographed by the infrared camera 6 is input to a vehicle detecting section 10 and is digitally converted by an A / D converting section 16 therein. The converted image data is temporarily stored in the frame memory 17. Then, the CPU 18 calculates each vehicle data such as the detection of the vehicle position, the speed, the traveling lane position, and the vehicle width, and the vehicle data is transmitted to the integrated determination unit 14 via the communication I / F 19. Similarly, images (image data) captured by the infrared cameras 7, 8, and 9 are input to the vehicle detection unit 10,
After the image processing, each is sent to the integration determination unit 14. Then, the integration determination unit 14 performs a cross-correlation operation on the vehicle information, calculates the width, height, and length of the vehicle, and integrates the processing results to detect the vehicle. As will be described later, the vehicle data finally integrated and determined by the integration determination unit 14 is output as vehicle information.

Hereinafter, the details of the present invention will be described with reference to the determination conceptual diagrams of FIGS. 4 and 5 and the flowchart (integration determination processing) of FIG. Here, FIG. 4 shows an imaging screen when one large vehicle is traveling without a “concealed vehicle” and an image data processing process thereof, and FIG. 5 shows an imaging screen when a “concealed vehicle” is present. It shows the image data processing process, and here is a conceptual diagram of judgment when judging coincidence or non-coincidence of images by a cross-correlation operation. In this case, the ordinary vehicle is a “concealed vehicle” for the large vehicle. First, assuming that one large vehicle is running on the road as shown in FIG. 4, first, an image (front left) of the large vehicle is photographed by the infrared camera 6 (“screen A”), and the photograph is taken. Vehicle information extracted from the vehicle detection unit 10 based on the image is acquired (S
1) Then, similarly, an image (front right) of the large vehicle is photographed by the infrared camera 7 located in the opposite position of the infrared camera 6 (“screen B”), and the vehicle extracted from the vehicle detection unit 11 based on the photographed image Information is obtained (S2). That is, at this stage, the entire large vehicle is photographed from both sides, and the image data is detected as vehicle information.

At this time, an outline extraction process is performed on the image of the vehicle taken from each of the infrared cameras 6 and 7. In other words, the image data converted by the A / D converter 16 in the vehicle detector 10 is converted (from 0 to 255) from "black data (1)" to "white data (0)" for each pixel. Therefore, by performing the contour processing, only the contour portion of the image data having a high contour level value is set to “black (1)”, and the other portion is set to “white (0)”. In the process of this contour extraction processing, the position of the center of gravity on the image data in each extracted image is specified and set. For example, this contour extraction processing is performed by using a comparator or the like in which a slice level is set to a predetermined value. If the distances from the infrared cameras 6, 7 to the vehicle are different, a magnification conversion process is performed on one of the acquired image data (S
3). By this magnification conversion processing, the sizes of both image data become substantially the same. Further, after that, the left / right inversion processing is performed on the image data on the side enlarged by the magnification conversion processing (S4).

That is, when the vehicle is traveling on one side of the road (left side in FIG. 7) as shown in FIG. 7, there is a difference between the infrared cameras 6, 7 and the distances L ₁ and L ₂ to the vehicle. At the same time (L ₁ <L ₂ ), the taken image (front right) taken from the infrared camera 7 side is naturally small because the separation distance is long.
Further, since the infrared cameras 6 and 7 capture images from opposite directions with respect to the vehicle, the directions of the respective extracted images (see “Screens A and B” in FIG. 4) are different. For this reason, enlargement processing is performed on the extracted image in order to perform “cross-correlation calculation processing” described later, and similarly, left-right reversal processing is performed on the enlarged extracted image. This left-right reversal processing (mirror image processing) is performed by reading out image data for one scan on the line memory from the opposite direction.

Then, "cross-correlation calculation processing" is performed on the two images after the image processing (S5). Here, FIG.
The “cross-correlation calculation processing” will be briefly described with reference to (a) and (b). For example, when judging the identity between the “figure α” and the “figure β” having the same shape shown in FIG. 8A, the “figure α” and the “figure β” are respectively set in the “cross-correlation calculation processing”. , The product-sum processing is performed to shift the center of gravity slightly. In this case, when the figures are the same, the corresponding matrix (matching contour data) is the product of “1 × 1” and “1”, so the diagram at the center of gravity (correlation processing result diagram) Is "1" which is the vertex as shown in FIG. That is, the position of the center of gravity exceeds a preset “threshold”. Therefore, by performing the determination of the preset “threshold” correlation value (S6), in this case, it is determined that the extracted images captured from the two directions are the same vehicle (S6).
7). On the other hand, as shown in FIG. 8 (b), when the “figure α” and the “figure γ” are not the same shape, even when the sum-of-products processing is performed, as shown in FIG. The position also appears as a flat diagram that does not exceed the preset “threshold”. That is, as shown in FIG. 5, the image of the large vehicle (front left) captured by the infrared camera 6 is “screen A”, and the image of the large vehicle (front right) captured by the infrared camera 7 is “screen C”. , It is determined that there is no vehicle identity, and the presence of a “concealed vehicle” is detected.

Further, the procedure will be described with reference to the judgment flowchart of FIG. 6. If the sum of the correlation values is not equal to or greater than the "threshold", the vehicle detection positions of the vehicle detectors 10 and 11 are compared (S8). . Next, it is determined whether or not the existing positions match (S9). Specifically, it is determined whether or not the presence positions at the vehicle bottom detected by the infrared cameras 6 and 7 match, and if the result is determined to be “the presence positions match”, it is determined that the vehicle is the same vehicle. Is determined (S10). Then, when it is determined that the “existing positions are not the same”, “determination of the concealed vehicle” is performed (S11). That is, at any position, the integrated correlation value does not exceed the “threshold” and it can be determined that the vehicles are not the same vehicle.

[0017]

The present invention is as described above. According to the first aspect of the present invention, there are provided a plurality of infrared cameras arranged on the left and right at predetermined intervals along a road side of a traveling road, and a plurality of these infrared cameras. A plurality of vehicle detection units to which image data of a plurality of traveling vehicles captured by the infrared camera are input, and an integrated determination unit to which a plurality of vehicle information processed by each of these vehicle detection units are input, Since the vehicle information processed by the integration determination unit is integrated and the overall determination of the vehicle is performed by performing cross-correlation calculation processing on each image data, a large vehicle and a normal vehicle run side by side next to each other. In this case, the concealed vehicle can be detected even when the vehicle is running, and the detection accuracy of a plurality of vehicles traveling on the road can be improved. In addition, since an infrared camera is used instead of a visible camera, erroneous detection due to the influence of the vehicle itself or a building can be prevented, and illumination is unnecessary even at night, so that an accurate detection function can be exhibited. .

According to a third aspect of the present invention, in the first aspect of the present invention, the integrated judgment processing unit separately processes two image data respectively photographed by the two infrared cameras, and among them, After performing a vehicle contour extraction process on one of the image data, a magnification conversion process and a left-right reversal process are performed, and a cross-correlation calculation process is performed on the two image data to determine the vehicle identity and determine the concealed vehicle Therefore, even when a large vehicle and a normal vehicle run side by side adjacently, a hidden vehicle can be detected, and the detection accuracy of a plurality of vehicles traveling on a road can be improved. In addition, since an infrared camera is used instead of a visible camera, erroneous detection caused by the shadow of the vehicle or building can be prevented, and no illumination is required even at night, and an accurate detection function can be demonstrated. This has the effect.

According to a fourth aspect of the present invention, in the third aspect of the invention, image data processed based on vehicle speed and vehicle width information as vehicle information obtained by the integrated determination processing section are compared with each other, When these pieces of information do not match, since the vehicle is not determined, there is an effect that the detection accuracy of a plurality of vehicles traveling on the road can be further improved.

[Brief description of the drawings]

FIG. 1 is a front view showing an embodiment of the present invention.

FIG. 2 is a plan view showing the arrangement.

FIG. 3 is an overall block diagram showing an example of the functional configuration.

FIG. 4 is a conceptual diagram of the determination when the vehicle is the same.

FIG. 5 is a conceptual diagram of determination when a concealed vehicle is present.

FIG. 6 is a flowchart illustrating an integration processing determination procedure.

FIG. 7 is an explanatory diagram showing a difference in a separation distance.

FIG. 8 is an explanatory diagram illustrating a cross-correlation calculation process.

FIG. 9 is a schematic configuration diagram of a conventional traffic flow detection device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Road 2 Visible camera 3 Large vehicle 4 Ordinary vehicle 6,7,8,9 Infrared camera 10,11,12,13 Vehicle detection unit 14 Integrated judgment unit 16 A / D conversion unit 17 Frame memory 18 CPU 19,20 Communication I / F 21 Vehicle position determination unit 22 Output I / F

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) H04N 5/225 H04N 7/18 C 7/18 G06F 15/62 380 15/70 410 460A F term (reference) 5B057 AA16 BA01 CA02 CA08 CA12 CA16 CB02 CB06 CB12 CB16 CC03 CD05 CD20 DA08 DB02 DB05 DB09 DC16 DC34 5C022 AA01 AA14 AB61 5C054 AA01 AA05 CA05 CC02 CE02 CH04 DA08 EA05 EB05 ED11 EH07 EA00 5H180 BB00 DD06 FA06 FA34 HA03 HA08 JA11

Claims (4)

[Claims] 1. A plurality of infrared cameras arranged on the left and right at predetermined intervals along a road side of a traveling road, and a plurality of image data of a plurality of traveling vehicles captured by the plurality of infrared cameras are input. A vehicle detection unit, and an integrated determination unit to which a plurality of vehicle information processed by each of the vehicle detection units is input. The vehicle information processed by the integration determination unit is integrated, and A traffic flow detection device for performing a comprehensive determination of a vehicle by performing a cross-correlation calculation process on the traffic flow. 2. The vehicle detection unit includes an A / D conversion unit, a frame memory, a CPU, and a communication I / F, and the integrated determination unit includes a communication I / F, a vehicle position determination unit, and an output I / F. The traffic flow detecting device according to claim 1, wherein: 3. An integrated judgment processing unit separately processes two image data respectively photographed by two infrared cameras, and executes a magnification conversion process after a vehicle contour extraction process on one of the image data. 2. The traffic flow detection method according to claim 1, wherein the vehicle identity is determined by performing a cross-correlation calculation process on these two image data, and a concealed vehicle is determined. apparatus. 4. Comparing the image data processed based on the vehicle speed and vehicle width information as the vehicle information obtained by the integrated determination processing unit, and when these information do not match, it is determined that the vehicle is not determined. The traffic flow detection device according to claim 3, wherein