JP2792697B2 - Vehicle type identification device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to a vehicle type discriminating apparatus applied to discriminating a vehicle type of a toll machine product.

[Prior Art] FIG. 6 is a diagram showing a configuration of a conventional vehicle type discriminating apparatus.
The vehicle separator 11 detects the entry of the vehicle, the image pickup device 12 and the image processing device 13 recognize the license plate of the vehicle, the tread plate measures the tread of the vehicle, the vehicle height detector 14 measures the vehicle height, and the overhang detector 15 corresponds to the vehicle. The distance from the tip of the tire to the first axis tire is measured, and the recognition result and the measurement result are comprehensively determined to determine the vehicle type.

[Problems to be Solved by the Invention] As shown in FIG. 6, the conventional vehicle type discrimination requires a large number of sensors, and treads and the like need to be embedded in the road. Vehicles had to be closed, and it was considered difficult to introduce a vehicle type discrimination device on an existing road with heavy traffic.

An object of the present invention is to configure a device configuration for vehicle type discrimination by a vehicle separator, an image pickup device, and an image processing device, and to measure a vehicle width, a vehicle height, and the like which are measured by a conventional tread plate and a vehicle height detector. It is an object of the present invention to provide a vehicle type discriminating apparatus capable of shortening a vehicle closing time required for installation of devices and reducing equipment costs by performing image processing.

[Means for Solving the Problems] A vehicle type discriminating apparatus according to the present invention includes a vehicle separator for detecting a vehicle passing through a toll gate, and an imager for photographing a front surface of the vehicle with a vehicle detection signal from the vehicle separator. A license plate extracting means for extracting a license plate portion from an image photographed by the imager; a plate size determining means for determining a size of the license plate extracted by the means; and a license plate extracting means for extracting the license plate. A classification number recognizing means for extracting a vehicle type classification number portion from the obtained license plate and recognizing the vehicle type classification number; a compression means for compressing an image taken by the image pickup device; and a vehicle from the image compressed by the compression means. A vehicle cutting means for cutting out a portion, and a vehicle width and a vehicle height are measured from an image cut out by the vehicle cutting means. Vehicle shape feature measuring means for measuring the vehicle shape feature, and the license plate size determined by the plate size determining means, the classification number recognized by the classification number recognition means, and the vehicle shape feature measuring means. And comprehensive determination means for comprehensively determining the vehicle type from the vehicle shape characteristics.

[Operation] According to the present invention, since license plate recognition and vehicle shape feature measurement are performed by image processing, a license plate existence range (about 1.5 m horizontally × 1 m vertically) required for conventional license plate recognition is included. Use a high-resolution electronic shutter camera capable of resolving the small 6-mm-thick license plate in the above-mentioned imaging field of view using a wide imaging field of view that covers the full width of the lane (horizontal 3m). Where lm is lm, the required horizontal resolution Ah is Ah>
It will be 1000 l / 6. In addition, the feature measurement of the vehicle shape is performed by compressing the entire image photographed in the imaging visual field according to the measurement accuracy, measuring the vehicle width and the vehicle height, and recognizing the license plate in the imaging visual field. By extracting the plate, determining the plate size, extracting the vehicle type classification number, and recognizing the vehicle type classification number for the license plate existence range in the image, the vehicle type determination is based on the license plate recognition result and the vehicle shape. The feature measurement results can be comprehensively determined and performed.

[Embodiment] Fig. 1 is a diagram showing a configuration of a vehicle type discriminating apparatus according to an embodiment of the present invention. A vehicle separator 1 is an optical sensor for detecting a vehicle passing through a lane, and is arranged in a lane width direction. A plurality of light beams are passed through the vehicle, and the vehicle is detected when the tip of the vehicle blocks any light beam. The imaging device 2 uses the vehicle detection signal to photograph the front of the vehicle obliquely at a shutter speed of about 1/1000 with a horizontal field of view of about the lane width (about 3 m). Since the camera needs to resolve the lowercase 6mm width of the license plate, use a high-resolution electronic shutter camera with a horizontal resolution of 3m ÷ 6mm = 500 or more lines. The image processor 3 receives the still image video signal captured by the imager 2, performs AD conversion, performs license plate recognition processing and vehicle shape feature measurement processing, and comprehensively determines the recognition result and the measurement result. Determine the vehicle type.

FIG. 2 is a diagram showing a vehicle type discrimination conceptual diagram according to an embodiment of the present invention, wherein a1 shows an image field of view, an image of the entire vehicle taken obliquely, and b1 shows a range in which a license plate exists. ing. When performing license plate recognition processing, processing is performed for the range of b2 obtained by cutting out b1. This is to speed up the processing,
As a result of the processing, the determination of the b3 plate size (large plate / normal plate) and the recognition of two digits of the vehicle type classification number are obtained. a2 indicates that a1 is compressed to the measurement accuracy required for vehicle type determination and the vehicle is cut out. In this figure, the width between the lower left corner of the vehicle and the right tire is the vehicle width w, and the highest point from the bottom of the tire Up to the vehicle height h. Here, the reason why a1 is compressed to a2 is to increase the processing speed. As a processing result, a3 vehicle width w and vehicle height h are obtained.

For the comprehensive judgment of vehicle types, those that can be clearly judged by the plate size and vehicle type classification number perform the vehicle type judgment irrespective of the vehicle shape measurement result, and when the plate size and vehicle type classification number alone are not clear, consider the vehicle shape measurement result And make a decision.

In the example of b3 in FIG. 2, the ordinary plate 52 is recognized,
The vehicle is determined to be a normal vehicle regardless of the vehicle shape measurement result, but if the license plate cannot be recognized, it is determined that the vehicle is a normal vehicle based on the vehicle width w and vehicle height h of the a3 vehicle shape measurement result.

FIG. 3 is a diagram showing an example of measurement of a large vehicle or the like exceeding the imaging field of view. When the vehicle type is determined from the measurement results of the vehicle width and the vehicle height, the vehicle size is classified according to the vehicle type classification intended. This indicates that clipping may be performed with the entire imaging field of view in order to make a determination.

Since the actual vehicle type determination differs depending on how the vehicle type is classified, the concept will be described here with reference to the example of the vehicle type determination table in FIG. In Table 1 of FIG. 4, the size M of the license plate recognition result indicates a normal plate, and L indicates a large plate. The number is the recognized vehicle type classification number, and X is
Indicates Don'tcare.

The figures given to the light vehicle to the oversized vehicle indicate the possibility of the vehicle type, 100 may be considered as 100%, 0 may be considered as 0%, and k1 ~
k9 is the probability determined from the corresponding vehicle type traffic ratio for the license plate recognition result. For example, in the case of the plate recognition result L1X, k1 is the probability of being a large vehicle, k2 is the probability of being an oversized vehicle, and k1 + k2 = 100. .

In this table example, it is assumed that the check is performed in order from M40.

Since the vehicle type is clearly determined from the M40 to M9X license plate recognition results in Table 1, the vehicle type is determined regardless of the shape measurement results in Table 2. L1X ~ Tab for others
The probability obtained by le1 and the probability obtained by the shape measurement result of Table 2 are added for each vehicle type, and the vehicle type having the highest probability is determined as the determination result. Table2 shows the vehicle width and the vehicle height as 3
Shown below is an example of classification, where vehicle width is WS <WM <WL and vehicle height is HS <HM
<HL, and the actual measurement values w and h give the probability of the corresponding section. k10 to k25 are the same as k1 to k9 in Table 1, and are the probabilities determined from the vehicle width and the vehicle ratio to the vehicle height, and taking the vehicle width WM as an example, the probability of an ordinary car is
k12, the probability of a medium-sized car is k13, the probability of a large car is k14, and k
12 + k13 + k14 = 100.

FIG. 5 is a diagram showing a process flow of the above-described vehicle type discrimination. In the vehicle shape characteristic measurement process flow, it is determined that the left tire can be recognized in the feature measurement in the case of a bus overhang (from the vehicle head to the first axis tire). Distance) is large, and depending on the installation height of the imaging device, the left tire is hidden behind the bonnet and cannot be seen, indicating that this can be used as a feature of the bus.

[Effects of the Invention] According to the present invention, the following effects can be obtained.

(1) Since a conventional tread plate and a vehicle height detector are not required, vehicle closing time during installation of the equipment can be shortened, so that a vehicle type discriminating apparatus can be introduced to an existing road with a large traffic volume. Also, equipment costs can be reduced.

(2) As a proposal of a vehicle type discrimination device for an existing road with a large traffic volume, a method of discriminating only by license plate recognition has been conventionally proposed. However, the recognition rate of the license plate is limited, and the number of license plates is limited. Even if you can recognize the plate,
Depending on the classification of the vehicle type, there may be cases where the vehicle type cannot be determined only by the license plate information.

The present invention is intended to perform vehicle shape feature measurement and to enable vehicle type discrimination from information such as vehicle width and vehicle height even if the vehicle type cannot be determined by license plate recognition, thereby improving the vehicle type discrimination rate. Is measured.

[Brief description of the drawings]

FIG. 1 is a block diagram of a vehicle type discriminating apparatus according to one embodiment of the present invention, FIG.
FIG. 4 is a diagram showing an example of measurement of a shape exceeding an imaging field of view in one embodiment of the present invention. FIG. 4 is a diagram showing an example of a vehicle type determination table in one embodiment of the present invention. FIG. FIG. 6 is a diagram showing a vehicle type determining process flow, and FIG. 6 is a diagram showing a conventional example. 1 ... Vehicle separator, 2 ... Imaging machine, 3 ... Image processing machine.

──────────────────────────────────────────────────続 き Continuation of front page (56) References JP-A-62-179099 (JP, A) JP-A-61-213998 (JP, A) JP-A-61-249199 (JP, A) (58) Field (Int.Cl. ⁶ , DB name) G08G 1/015 G08G 1/017 G08G 1/04

Claims (1)

(57) [Claims] 1. A vehicle separator for detecting a vehicle passing through a toll gate, an imager for photographing the front of the vehicle by a vehicle detection signal from the vehicle separator, and a license plate from an image photographed by the imager. License plate extracting means for extracting a part, plate size determining means for determining the size of the license plate extracted by this means, and car type classification number part from the license plate extracted by the license plate extracting means Classification number recognizing means for recognizing the vehicle type classification number, compression means for compressing an image taken by the image pickup device, vehicle extraction means for extracting a vehicle portion from the image compressed by the compression means, and vehicle extraction means Vehicle shape feature measuring device that measures the vehicle width and height from the image cut out by the means and measures the vehicle shape feature And a license plate size determined by the plate size determination means, a classification number recognized by the classification number recognition means, and a vehicle shape characteristic comprehensively determined from the vehicle shape characteristics measured by the vehicle shape characteristic measurement means. A vehicle type discriminating device comprising: a judging means.