JP2924063B2 - Image processing type traffic flow measurement device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing type traffic flow measurement device that detects a vehicle by image processing using an industrial television (ITV) camera or the like, and particularly to a measurement object. The present invention relates to a traffic flow measurement device capable of detecting a shadow generated around a certain vehicle with high accuracy and removing the influence of the shadow.

[Prior Art] In a traffic flow measuring device of this type, a shadow generated around a vehicle may be erroneously detected by incorrectly measuring the size of the vehicle, or may be erroneously detected by another vehicle. This may lead to erroneous detection of the number of measurement vehicles. Therefore, it is necessary to accurately detect the shadow generated around the vehicle and remove the influence of the shadow.

As a method of detecting the shadow of a vehicle in this type of conventional traffic flow measuring device, the shadow of the vehicle is detected using the fact that the shadow of the vehicle is darker than the road surface and the temporal change in brightness is almost uniform. The method is generally known (see Hashimoto, Kumagaya et al., “Development of an Image Processing Traffic Flow Measurement System”, Sumitomo Electric 127, p. 59, September 1986).

[Problems to be Solved by the Invention] However, in the related art, as described above, the fact that the brightness of the shadow portion is darker than the road surface or the temporal change of the brightness value at the measurement point is uniform is used. Therefore, it was difficult to determine that the vehicle was a black vehicle whose luminance value was lower than the road surface reference luminance.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an image processing-type traffic flow measurement device that solves the above-described problems, can accurately remove the influence of shadows, and can improve the accuracy of discriminating a vehicle type and the accuracy of measuring the number of vehicles.

[Means for Solving the Problems] In order to achieve the above object, according to the present invention, a portion having a luminance different from that of a road surface portion is obtained from an input image obtained by photographing a road surface and a vehicle existing on the road surface. In a device for detecting the vehicle by extracting as a feature point, a calculating means for calculating a difference value between a luminance value of the input image having a luminance lower than a reference luminance of the road surface portion and a reference luminance value of the road surface portion, By comparing adjacent difference values using all of the difference values calculated by the calculation means, a two-dimensional distribution pattern of the difference values calculated by the calculation means is determined along the longitudinal direction or the transverse direction of the road surface. And a shadow detecting means for detecting a portion which continuously changes with a certain inclination as a shadow portion of the vehicle.

[Operation] In the present invention, the distribution density pattern of the difference value between the luminance of the input image captured from above the measurement target vehicle and the road surface reference luminance is focused on the characteristic that the shadow of the vehicle becomes thinner as the distance from the vehicle increases. A part that changes sequentially in a certain direction is detected as a shadow of the vehicle. As described above, in the present invention, the shadow of the vehicle is detected by utilizing the property that the shadow becomes thinner as the distance from the vehicle increases, so that the shadow can be detected stably irrespective of the vehicle color and shape of the traveling vehicle and the stationary vehicle. Even if the vehicle extends on the lane, the shadow is not erroneously detected as a black vehicle. Therefore, according to the present invention, the influence of the shadow can be effectively removed, and the accuracy of discriminating the vehicle type and the accuracy of measuring the number of vehicles can be improved.

[Example] Hereinafter, an example of the present invention will be described in detail with reference to the drawings.

FIG. 1 shows a circuit configuration of a main part of an image processing type traffic flow measuring device according to an embodiment of the present invention. In FIG. 1, reference numeral 1 denotes a camera unit (fixed camera) such as an ITV camera as an image pickup unit for picking up an image of a vehicle traveling on a road, and 2 generates an image around the vehicle by processing an output image signal of the camera unit 1. It is an image measuring unit that detects a shadow to be cast. The image measurement unit 2 includes an A / D (analog / digital) conversion unit 21, an image storage unit 22, and an image ternary processing unit.
23, a vehicle body detecting unit 24, a shadow determining unit 25, a vehicle measuring unit 26, and an output unit 27.

The A / D converter 21 digitizes the output image signal of the camera unit 1. The image storage unit 22 is composed of a RAM (random access memory) or the like, and stores a digitized image signal as a luminance signal for each imaging screen. The image ternarization process 23 performs ternarization (1, 0, -1) for each measurement sample point using the stored luminance signal. The vehicle body detection unit 24 detects the vehicle body from the shift in luminance at the measurement sample point. The shadow determination unit 25 determines the shadow of the vehicle based on the distribution pattern of the difference between the luminance value of the input image and the road surface reference luminance value, as described later. The vehicle measurement unit 26 identifies the vehicle using the vehicle shadow data obtained from the shadow determination unit 25, calculates various traffic flow measurement data without the influence of the shadow, and outputs the measurement data from the output unit 27 to a remote traffic flow. Send to the central control unit.

FIG. 2 shows an example of an actual mounting state of the camera unit 1 and the image measuring unit 2 in FIG. The camera unit 1 is fixed to an upper arm of a pole (pole) 3 erected on a road side 5 of a multi-lane road 4, and the image measuring unit 2 is fixed to a predetermined position of a trunk of the pole 3. The camera unit 1 captures an image of a predetermined range of the multi-lane road 4 from a predetermined height as a measurement area 6 at a predetermined angle, and the image measurement unit 2 performs a vehicle sensing process from the image data captured by the camera unit 1. In this embodiment, a case of three lanes on one side is considered as an example. In addition, 7 shows a center line (central separation zone).

FIG. 3 shows the principle of the determination operation in the above-mentioned shadow determination unit 25. As shown in FIG. 3 (A), the shadow of the vehicle body is divided into a front shadow and a horizontal shadow depending on the location where the shadow is generated. It is difficult to separate shadows. However, as shown in FIGS. 3 (B) and 3 (C), the shadow of the vehicle body becomes thinner as the distance from the vehicle body increases, both foreground shadows and lateral shadows. Using this phenomenon, the shadow determination unit 25 detects the shadow of the vehicle body as follows. Since the detection procedure is the same for the foreground shadow and the horizontal shadow, the case of the foreground shadow will be described as an example.

First, a measurement line is created by connecting each measurement sample point in the vehicle traveling direction and the road crossing direction. Therefore, the measurement points are arranged in a grid in the measurement area 4. In the case of a foreground shadow, a measurement line (a) in the vehicle traveling direction is used. A portion where the input luminance is darker than the road surface reference luminance in a continuous section of the measurement line in the vehicle traveling direction is detected. For each measurement sample point included in the detected section, a difference value between the road surface reference luminance and the input luminance is obtained. For example, when the measurement line (a) is scanned from the downstream side of the road to the upstream side, a gradient having the above-mentioned difference value is obtained. If the value gradually increases with, it is extracted as a shadow candidate.

Next, the same processing is repeated by shifting the measurement lines one by one in the direction of the road cross section, and the same processing is repeated. If a shadow candidate exists over a plurality of measurement lines and is extracted as a plane area, this area is shaded (previous). Shadow).

Further, the operation of the embodiment of the present invention will be described in detail with reference to the flowchart of FIG.

An analog image signal from the camera unit 1 is transmitted to an A / D converter 21
Is converted into digital luminance data of, for example, 256 gradations, and is recorded in the image storage unit 22 in units of screens (step
S1). Then, after the luminance data for each pixel at the measurement point is extracted by the pixel extraction, a difference (difference value) from the road surface reference luminance is calculated for each pixel by the luminance calculation (step S2). The difference value data for each pixel obtained in this way is “+1” (brighter than the road surface), “0” (same as the road surface), and “−1” (darker than the road surface) in the pixel ternarization processing unit 23. (Step S)
3). That is, the road surface usually has a substantially constant short luminance, and this luminance (road surface reference luminance) does not change in a certain short time interval, and is fixed by the measurement area. Therefore, this reference luminance is obtained in advance or as needed, and this is compared with the luminance data of each measurement point. Then, for example, the pixel is ternarized as a pixel having a certain level of luminance higher than the road surface... +1 A pixel having substantially the same luminance as the road surface...

Next, it is detected whether the ternary data of each pixel is darker than the road surface. (Step S4). That is, in the case of the present embodiment, since the ternary data of the road surface portion is “0”, a portion where the ternary data is “−1” is detected. Then, it is determined whether or not the dark portion thus detected is continuous in the vertical and horizontal directions (step S5). Subsequently, the measurement line (a) is scanned from the downstream of the road to the upstream of the continuous dark portion,
A case where the difference value between the road surface reference luminance and the input luminance performed for each measurement sample point gradually increases with a certain slope from the downstream side to the upstream side is recognized as a foreground shadow of the vehicle (step S6).

Here, assuming that the vehicle is traveling on the road surface toward the camera unit 1 or is stopped,
The bonnet always appears behind the foreground shadow, and the roof always appears behind the hood. The brightness of the hood and roof is higher than the brightness of the foreground shadow.
It is often higher than the brightness of the road surface. However, the luminance is not constant as compared with the luminance of the road surface. Therefore, the hood and the roof of the vehicle can be detected by examining whether there is a pattern having the next highest luminance or a pattern having a disturbed luminance after the foreground shadow. These detection results are sent to the next vehicle measurement unit 26 (step S7). Then, it is determined whether the detected foreground, hood, and roof of the vehicle appear within a certain range in the lateral direction (step S).
8). That is, as shown in FIG. 3 (A), the foreground shadow, the bonnet portion, and the roof portion of each vehicle should appear within a certain range in the lateral direction, and thus can be recognized as one vehicle. (Step S1
1).

On the other hand, when the lateral width of the hood and the roof is not within a certain range, it is generally a case that the vehicle is laterally shaded. Therefore, it is determined whether or not a portion darker than the road surface on the lateral side of the vehicle body is continuous in the vertical and lateral directions. Is checked (step S9). Subsequently, the measurement line (b) is scanned from the road side 5 of the road toward the center line 7 for the continuous dark part, and the difference between the road surface reference luminance and the input luminance performed for each measurement sample point is calculated. A case where the value gradually decreases (or increases) with a certain inclination from the roadside 5 toward the center line 7 is recognized as a lateral shadow of the vehicle (step S10), and is recognized as one vehicle based on the above measurement results. (Step S11).

After the above processing is completed, the vehicle measurement unit 26 performs tracking of the vehicle, determination of the lane, measurement of the vehicle, and the like. And
The result of the measurement is sent from the output unit 27 to a remote traffic control central device and output and displayed.

That is, the vehicle measuring unit 26 recognizes only the vehicle body by removing the shadow portion from the data of the image storage unit 22 using the vehicle shadow detected by the shadow determining unit 25 as noise. This vehicle body recognition is performed, for example, by detecting the rise and fall of the in-vehicle data in the measurement area. That is, the measurement area is divided into an upstream area and a downstream area, the traffic volume is counted as a passing vehicle at the rise of the vehicle data in the downstream area, and the distance between the rise of the vehicle data and the fall of the vehicle data in the downstream area is counted. The vehicle length is measured, large and small vehicle types are determined, and the speed (vehicle speed) of the vehicle is calculated from the travel distance and travel time of the vehicle. These traffic flow measurement data are transmitted from the output unit 27 to a remote central device. Therefore, the image measuring unit 2 can obtain accurate vehicle type determination data and the number of passing vehicles measurement data that are not affected by the shadow.

Note that the present invention is not limited to the above-described embodiment, and various modifications are possible.

In the embodiment, the shadow under the bumper of the vehicle is detected, but the invention is not limited to this. For example, the running vehicle may be photographed from behind by a camera, and a shadow formed behind the vehicle may be detected. Further, the vehicle is not limited to a vehicle traveling on a road surface, but may be a vehicle parked on a road side or parked in a parking lot.

As the camera 1, for example, a CCTV camera can be used, but it is not limited to this, and it is not essential to use the image storage unit 22. That is, the camera and the image processing device can be directly connected to perform online and real-time processing.

The brightness data is not limited to 2 ⁸ = 256 gradations, but 2 ⁴ = 16
Any gradation such as gradation, 2 ⁶ = 64 gradations may be used. Also,
Luminance data is not limited to ternary data, but can be binarized, quaternary,
Anything such as value conversion may be used.

[Effects of the Invention] As described above, according to the present invention, an input imaged from above a measurement target vehicle is used by utilizing the property that the shadow of the vehicle becomes thinner as both the foreground shadow and the lateral shadow become farther from the vehicle body. Since the distribution density pattern of the difference value between the brightness of the image and the road surface reference brightness continuously changes with a certain slope with respect to the vehicle traveling direction or the road crossing direction is detected as the foreground shadow or the side shadow of the vehicle, The shadow can be detected stably and accurately regardless of the color and shape of the running vehicle, so the accuracy of vehicle type determination (classification of large / small vehicles) can be improved, and even if the shadow extends over an adjacent lane. Since the shadow is not erroneously detected as a black vehicle, the effect of improving the accuracy of traffic measurement can be obtained.

Further, in the present invention, the shadow portion of the vehicle is detected over a plurality of lanes by detecting the shadow of the vehicle by comparing the adjacent difference values using all of the difference values calculated by the calculation means, and thereby, each individual shadow is detected. By recognizing the vehicle in combination with the detection of the hood and the like while identifying the vehicle, even if there is a vehicle traveling over two lanes or changing lanes, two or more lanes may be These can be accurately recognized even when the vehicles are running in parallel.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a circuit configuration of an embodiment of the present invention, FIG. 2 is a perspective view showing an installation example of the apparatus of FIG. 1, and FIGS. 3 (A), (B) and (C) show a vehicle. FIG. 4 is an explanatory diagram for explaining the nature of the shadow and the operation of determining the shadow. FIG. 4 is a flowchart showing the operation of the embodiment of the present invention. DESCRIPTION OF SYMBOLS 1 ... Camera part, 2 ... Image measurement part, 3 ... Support, 21 ... A / D conversion part, 22 ... Image storage part, 23 ... Image ternary processing part, 24 ... Car body detection part , 25 ... Shadow determination unit, 26 ... Vehicle measurement unit, 27 ... Output unit.

Claims (1)

(57) [Claims] 1. An apparatus for detecting a vehicle by extracting, as a feature point, a portion different in luminance from a road surface portion from an input image obtained by photographing a road surface and a vehicle existing on the road surface, Calculating means for calculating a difference value between the luminance value of the input image having a luminance lower than the reference luminance value of the road surface portion and the reference luminance value of the road surface portion, and using all of the difference values calculated by the calculating device By comparing adjacent difference values, a part in which the two-dimensional distribution pattern of the difference values calculated by the calculating means continuously changes with a certain slope along the longitudinal direction or the transverse direction of the road surface is changed. An image processing type traffic flow measuring device, comprising: a shadow detecting means for detecting a shadow portion.