JPH04153900A - Vehicle detection device - Google Patents

Abstract

PURPOSE:To make optimum adjustments of the state of a detection place and an operation mode by photographing an image of a road surface where a vehicle travels by a camera and storing the image as a still image, and collating it with images of the road surface in a vehicle detection area and a specific pattern drawn thereupon and detecting the presence of the vehicle from the difference image between the both. CONSTITUTION:The TV camera 2 is installed, an area 16 is set initially by a setting part 13 for a speed indication '40', and an image when there is no vehicle present is stored in a reference image memory 6. When the vehicle passes in a setting state, an object 16 of recognition in the visual field of the TV camera 2 is shielded by the vehicle body and then the processing result of collation of the image in a still image memory part 5 with the reference image in the reference image memory part 6 is stored in a difference area memory part 8 to obtain the part of the vehicle which overlaps the object area. A vehicle detection part 9 decides that there is a difference and a detection signal output part 10 outputs a vehicle detection signal. Consequently, the optimum adjustments of the width of lanes in the detection place, the travel state of the vehicle, and the operation mode are made.

Description

[Detailed Description of the Invention] Industrial Application Fields The present invention is applicable to measuring the number of passing vehicles on public roads or expressway toll gates, detecting traffic jams from passing intervals of vehicles, and triggering the passing of vehicles. The present invention relates to a vehicle detection device used for issuing tickets and opening/closing gates.

BACKGROUND OF THE INVENTION In recent years, as the number of vehicles has increased, road traffic management has become one of the important issues, and vehicle detection means are important.

Conventional devices of this type include: ■ A tape sensor method in which a tape sensor is pasted on the road and the resistance value changes when a vehicle steps on it.

■ Loop coil method that detects magnetic changes by burying a coil in the ground.

■ An ultrasonic method that uses ultrasonic sensors installed overhead or on the roadside to detect temporal changes in reflected waves from the vehicle body or road surface, as well as the Doppler effect.

■ Radar-only type, with a radar sensor installed in front or on the roadside to detect time changes in reflected waves from the vehicle body and road surface, as well as the Doppler effect.

■ A light method that uses a light emitter and light receiver installed overhead or on the roadside to detect changes in the amount of light received when a vehicle passes by.

■ Image processing methods have been devised that recognize the shape of a vehicle by capturing images with a TV camera.

Problems to be Solved by the Invention However, in such conventional detection devices, the above method (2) has a short lifespan, and only the number of axles of a vehicle can be determined, but the number of vehicles cannot be determined. Installation is difficult with method (2).

In method (2), the detection timing is delayed and inconsistent due to the firing cycle depending on the installation location, and adjustment is required and it cannot be easily relocated.

The detection timing is fast with the method ①, but adjustment is required.
Not suitable for low-speed vehicles such as traffic jams. Method (2) requires adjustment of the direction of the optical axis and is difficult to relocate.

In method (2), since vehicles have a variety of shapes, it is difficult to identify common features, making it difficult to obtain stable and high detection accuracy.

The present invention solves these conventional problems,
Existing signs can be used, and even if there are no signs, targets can be set by simply marking the road where vehicles will pass, and it is easy to adjust and can be relocated, so it can be used as a portable type. It is an object of the present invention to provide a vehicle detection device that can select from a wide variety of operating modes so as to be optimally adjusted according to the width of the lane depending on the vehicle detection location, the driving condition of the vehicle, and the mode of operation.

Means for Solving the Problems In order to achieve the above objects, the present invention provides: (1) a television camera for photographing a predetermined area of a road surface on which a vehicle runs, which has a predetermined pattern; means for storing a screen shot by the camera as a still image; and means for constantly updating and storing a screen when there is no vehicle in the predetermined area as a reference image;
means for comparing the still image and the reference image in a predetermined area range and storing a difference; means for detecting the presence of a vehicle from the difference storing means; and outputting a detection signal based on the detection result of the detecting means. (2) The vehicle detection device is provided with an operation mode setting means, and the operation mode setting means adjusts the operation mode of the television camera according to the conditions for detecting a vehicle. The operating conditions of the still image storage means, the reference image storage means, the difference storage means, and the vehicle detection means are set, and updating of the reference image is prohibited to prevent changes in road surface brightness due to vehicle illumination lights. a reference image update prohibition mode that does not correspond to the update; and an order check mode that detects the passage of a vehicle based on three changes: before passing, during passing, and after passing;
Detailed mode performs detailed detection in a narrow detection area, and
A check mode detects the presence or absence of a boundary between the pattern and the road surface, a leading edge matching mode detects changes in the vehicle-side leading edge of the pattern, and detects changes in each of multiple blocks in which the detection area is divided. The vehicle detection device is configured to be set in combination with an average detection mode.

According to the above-described structure, the television camera captures an image of the road surface on which the vehicle is running, and the still image storage means stores the image as a still image.

A reference image storage means compares an image of the road surface of the vehicle detection area and a predetermined pattern drawn thereon with the reference image, and stores a difference image between the two. The detection means detects the presence of a vehicle from the different images, and the detection signal output means outputs a detection signal in response to the detection of the vehicle. Still, the operation mode setting means sets a detection mode suitable for vehicle detection conditions.

Embodiment Hereinafter, a vehicle detection device according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a vehicle detection device according to a first means of the present invention.

In the figure, reference numeral 1 denotes a lens unit provided in front of the TV left camera. 3 is a lens control unit that controls the iris of the camera lens; 4 is a mu/D conversion unit that converts the video signal from the camera into digital data; 5 is a still image memory unit that stores the image digital data as a still image; Reference numeral 6 is a reference image memory section in which a still image taken when no vehicle is present is stored in the initial setting, and is subsequently updated in response to a change in the noise level or dirt on a sign; 7 is a still image memory section 5; a specified area mask processing and image matching processing unit that compares only the portion specified in the initial settings of the image in the reference image memory unit 6;
Reference numeral 8 denotes a difference area image memory unit that stores only the different portions of the processing results of the processing unit A; 9 a vehicle detection determination unit that detects the presence of a vehicle from the differences in the difference area memory unit 8;
10 is a detection signal output section that outputs the result.

11 is a D/MU conversion unit that converts the digital images stored in each memory 5, 6, and 8 into an analog video signal; 12 is a monitor display unit that displays the digital images; 13 is a display unit that displays the image in the reference image memory unit 6 in the initial setting; This is a target area designation setting section that sets a target area such as a sign using a mouse or other means, and is used at the initial setting and adjustment stage.

Reference numeral 14 is an overall control unit that controls the above as a whole, and 15 is a continuous illumination device that is necessary only when the amount of light at night is insufficient.

FIG. 2 is a schematic diagram showing a general road on which the vehicle detection device of the present invention is installed and vehicles on the road.

FIG. 3 is a pattern diagram showing areas where target areas are specified in the initial setting of the vehicle detection device according to an embodiment of the first means of the present invention, and 16 is the specified target area.

The operation of the above configuration will be explained.

As shown in Figure 2, install the TV left camera, adjust the angle of view, and use the setting section 13 in Figure 1 to set the 16 areas shown in the pattern diagram in Figure 3 for the speed indicator "40". This is assumed to be the initial setting. In this case, an image when no vehicle is present is stored in the reference image memory section 6. When a vehicle passes under such a setting state, the recognition target 16 within the field of view of the TV left camera is blocked by the vehicle body. Therefore, the part of the vehicle that overlaps the target area is stored in 8, which is the processing result of the process 7 that compares the image captured in the still image memory section 5 with the reference image of the reference image memory section 60. The vehicle detection and determination unit 9 determines that there is a difference, and outputs a vehicle detection signal from the detection signal output unit 1o indicating that the vehicle is being detected.

Since the sign 16 is imaged after the vehicle passes, the output from the difference 1o is that there is no vehicle, and the reference image is updated during this time. If there is no appropriate marking, you can draw a white line on the road with tape or paint. Furthermore, in order to separate and detect the vehicles in front and behind even if the inter-vehicle distance is short, cameras installed on the roadside should be installed as perpendicular to the direction of vehicle travel as possible. On the other hand, in order to prevent the towing vehicle from being separated into two vehicles, the camera is installed as parallel to the direction of travel of the vehicle as possible. In addition, by increasing the field of view of the camera and making the recognition target sign smaller, conversely, by reducing the field of view and increasing the recognition target, it also affects the above-mentioned ability to determine the distance between vehicles. Since the recognition universal display itself has different sizes, if it is large, a part of it may be set as the recognition target area during initial settings. However, when specifying a target area, it is necessary to include not only the sign but also some of the surrounding road surface. This is because the sign is a single color, V, and a black and white camera cannot distinguish it from the vehicle body, which has the same brightness. The condition of the vehicle is different. In addition, priority may be given to the detection target vehicle depending on the purpose. According to the present invention, adjustments can be easily made while visually confirming the conditions.

According to the vehicle detection device of one embodiment of the above first means,
a television camera for photographing a predetermined area of a road surface on which a vehicle travels and is provided with a predetermined pattern; a means for storing a screen shot by the camera as a still image; means for storing the screen as a reference image while constantly updating the screen when the image is not displayed; means for comparing the still image and the reference image in a predetermined area range and storing the difference; By providing a vehicle detection device comprising a means for detecting the presence of a vehicle and a detection signal output means for outputting a detection signal based on the detection result of the detection means, it is possible to detect existing markings on the road surface or a pattern drawn by simple painting. Vehicles can be detected using existing television cameras and with easily adjustable equipment.

Hereinafter, a vehicle detection device according to an embodiment of the second means of the present invention will be described with reference to the drawings. FIG. 4 is a block diagram showing the configuration of a vehicle detection device according to an embodiment of the second means of the present invention. The device of this second means is a device that is provided with an operation mode setting means in the device of the first means, and the configuration and operation of other parts are the same as those of the first means, so the explanation will be omitted. The operation mode setting means provided in the second means will be explained below. In FIG. 4, 18 is an operation mode setting section. This operation mode setting section determines the conditions when detecting a vehicle, such as when you want to avoid the influence of vehicle headlight illumination, when you want to reliably detect passing vehicles, when the pattern on the road is very dirty, and when the vehicle is detected. Set the mode in combination according to conditions such as when you want to detect the front side of the camera. For example, when using the vehicle at night, the road surface becomes extremely bright due to the influence of the vehicle's headlights. One countermeasure is to install the camera in front of the recognition target sign, as shown in Figure 5, to avoid direct reflection of light, and the other is to update the reference image. A mode for limiting and a mode for weighting image changes at the boundary between the road surface and the marking within the recognition target area are selected in combination. This mode includes an update prohibition mode that does not update the reference image if there is a large change even when there is no vehicle, as in the case of lights, and an update prohibition mode that does not update the reference image when the target area is brightened by headlights. This process identifies whether it is a light or a vehicle body by specifically checking whether the boundary of the symbol remains.

In addition, if the purpose is to quickly detect the leading edge of the vehicle, such as when using it for the shutter trigger timing of another camera, it is possible to detect only changes at the boundary between the road surface and the marking within the recognition target area. Select the mode to match. In this case, differential processing is required, but the amount of data is small and vehicle determination processing is faster.

Furthermore, in order to improve the accuracy of vehicle detection, the size of the sign area to be recognized is set to a large size, and a mode is selected in which the order of changes in the front and rear of the target area, such as the sign, is checked. In other words, since the direction of travel of the vehicle is fixed, it is possible to check the vehicle when there is no vehicle present, when the vehicle is entering the vehicle, when the vehicle is passing, and when the vehicle is about to exit. It can reduce false detections caused by the influence of sunlight.

Furthermore, if there are many trucks loaded with earth and sand passing by, and the target sign changes significantly due to dirt, a mode is selected in which verification is performed using the average of partial blocks. This reduces detection sensitivity, but reduces false detections due to dirt.

Furthermore, if there are many trucks loaded with earth and sand passing by, and the target sign changes significantly due to dirt, a mode is selected in which verification is performed using the average of partial blocks. Although this reduces υ detection sensitivity, it reduces false detections due to dirt. A combination of modes can be set in response to conditions such as when the road pattern is extremely dirty or when the front of the vehicle is detected. These functions correspond to the purpose of preventing false information indicating that a vehicle has passed, and the purpose of detecting a vehicle.

For example, when used at night, a vehicle's headlights provide false information about the passing of a vehicle by changing the brightness of the image before the vehicle passes. In order to deal with this, we select a mode that focuses on checking the existence of the boundary between the road surface and the marking pattern to confirm that there are no vehicles, and select a reference image prohibition mode to temporarily check the presence of the boundary between the road surface and the marking pattern. Updates are prohibited when the reference image changes.

Furthermore, shadows caused by sunlight appear on the pattern, giving false information about vehicle passage. In order to cope with this, by selecting a mode that looks at three changes before, during and after the vehicle passes, pattern changes due to shadows are canceled out and removed. In addition, trucks loaded with earth and sand partially stain the pattern of the road surface, giving false information about the passing of vehicles. To deal with this, the area is divided into blocks, and a mode is selected in which an average judgment is made based on the ratio of changed parts to unchanged parts, etc., to remove changes due to dirt. However, detection sensitivity deteriorates. In addition, when vehicles are detected without being separated during traffic jams, conversely, towing vehicles may be detected as being separated, giving incorrect information on the number of vehicles. To cope with this, the detection area is set wide to capture multiple vehicles, and a detailed mode in which detection is performed pixel by pixel is selected to confirm the connection. Also, when detecting a vehicle for the purpose of obtaining a shutter trigger for the pond camera that captures the vehicle registration number,
A mode is selected to match the leading edge of the pattern on the vehicle side, and the moment when the front of the vehicle passes is detected.

Note that each of the above modes has a detection processing time, and the processing time limits the time interval of imaging. Therefore, when targeting high-speed vehicles, it is appropriate to consider the processing time and select a mode with a small amount of data, such as the average change detection mode. According to the vehicle detection device of the embodiment of the second means described above, by providing the operation mode setting means and setting the mode according to the surrounding conditions and purpose at the time of detection, it is possible to adjust the operation mode to suit each case. vehicle detection.

Effects of the Invention As is clear from the above embodiments, the present invention comprises: (1) a television camera for photographing a predetermined area of a road surface on which a vehicle runs, which has a predetermined pattern; means for storing a screen shot by a camera as a still image; means for constantly updating and storing a screen when there is no vehicle in the predetermined area as a reference image; and storing the still image and the reference image in a predetermined area. A vehicle detection device comprising means for comparing in range and storing a difference, means for detecting the presence of a vehicle from the difference storage means, and detection signal output means for outputting a detection signal based on the detection result of the detection means. (2) The vehicle detection device of the first means is provided with an operation mode setting means, and the operation mode setting means selects the television camera, the still image storage means, and the still image storage means according to the conditions for detecting a vehicle. Operating conditions of the reference image storage means, the difference storage means, and the vehicle detection means are set, and updating of the reference image is prohibited and the update is not made to correspond to changes in road surface brightness due to vehicle illumination lights. a prohibition mode, a sequence check mode in which vehicle passing is detected by three changes before, during and after passing; a detailed mode in which detection is performed in detail in a narrow detection area; A combination of a check mode that detects presence, a leading edge matching mode that performs detection by comparing changes in the leading edges on both early sides of the pattern, and a mode that detects based on the average of changes for each of multiple blocks obtained by dividing the detection area. By using vehicle detection equipment, existing markings on the road surface can be used, or simple patterns can be drawn on the road surface, existing cameras can also be used, and special kesasa can be installed on the road surface. Moreover, since the device is adjusted visually using images, it can be easily done without requiring any special tools, and the following effects can be obtained.

If there are no limitations on the detection point, existing markings can be used, and even if there are no markings, the target can be set by simply marking the road on which the vehicle will pass.

Furthermore, since the detection principle can be visually confirmed, adjustment is easy. It can be moved and used as a portable type. In addition, a wide variety of operating modes can be selected to allow optimal settings to be made depending on the lane width depending on the vehicle detection location, vehicle driving conditions, and operation mode.

ゝ－珂＼

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a vehicle detection device according to an embodiment of the first means of the present invention, and FIG. Fig. 3 is a pattern diagram showing the markings on the road surface used by the vehicle detection device of the first means of the present invention and the pattern drawn on the road surface;
FIG. 4 is a block diagram showing the configuration of a vehicle detection device according to the second means of the present invention, and FIG. 5 shows a state in which the camera of the vehicle detection device of the present invention is arranged to avoid the influence of headlight illumination of a vehicle. It is a diagram. 9...Vehicle detection determination unit (means for detecting the presence of a vehicle), 1o...Detection signal output unit (detection signal output unit), 16...Predetermined pattern on the road surface , 18
...operation mode setting section (operation mode setting means),
19...Television camera, 2Q...
Still image storage means, 21... Different image storage means,
22...Reference image storage means. - Name of University of Science and Technology Patent Attorney/”・Akira Kaji Haga 2 persons*
m Y1 length 11tf! (MuAshidsik r1 ratio i6+
49) Gu 2 - Kasuri Iris Statue Slaying Tiger ~ Asajizu Diagram

Claims (2)

[Claims] (1) A television camera for photographing a predetermined area of a road surface on which a vehicle is traveling and is provided with a predetermined pattern; a means for storing the screen shot by the camera as a still image; means for constantly updating and storing a screen when there is no vehicle in the area as a reference image;
means for comparing the still image and the reference image in a predetermined area range and storing a difference; means for detecting the presence of a vehicle from the difference storing means; and outputting a detection signal based on the detection result of the detecting means. A vehicle detection device comprising a detection signal output means. (2) The vehicle detection device according to claim (1) is provided with an operation mode setting means, and the operation mode setting means selects the television camera, the still image storage means, and the reference image in accordance with the conditions for detecting a vehicle. A reference image update prohibition mode in which operating conditions of a storage means, the difference storage means, and the vehicle detection means are set, and updating of the reference image is prohibited and the update is not made to correspond to changes in road surface brightness due to vehicle illumination lights. , an order check mode that detects vehicle passage through three changes before, during and after passing; a detailed mode that performs detailed detection in a narrow detection area; and a check mode that detects the presence or absence of a boundary between the pattern and the road surface. A check mode in which detection is performed, a leading edge matching mode in which detection is performed by matching changes in the vehicle side leading edge of the pattern, and a mode in which detection is performed based on the average of changes in each of multiple blocks obtained by dividing the detection area are combined. A vehicle detection device.