JPH06162177A - License plate image photographing device - Google Patents

Abstract

(57) [Summary] [Purpose] A license plate image capturing device for capturing an image of a license plate mounted on a vehicle, and an object thereof is to enable the license plate image to be captured as a single image at high speed and in a large size. . [Structure] A laser light projecting device 13 irradiates a fan-shaped laser light obliquely and horizontally to the road surface near the hood or rear gate of a vehicle. The image of the projection locus 32 is captured by the television camera 10 via the convex mirror of the rotating mirror 15, and the center position of the vehicle body is detected by noting that the vehicle body shape is generally center symmetric. Based on the information, the rotation angle of the plane mirror portion of the rotary mirror 15 is controlled, and the zoomed-in license plate image is captured by the TV camera 10 as one image.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a license plate image capturing apparatus for capturing an image of a license plate (registration number mark) mounted on a vehicle.

[0002]

2. Description of the Related Art When a license plate is photographed and its characters are recognized, it is necessary to magnify and photograph it in order to distinguish not only the serial number but also the name of the Land Transport Office from the image obtained from a TV camera. is there. This is because if the image signal obtained from the camera is converted into a digital image by A / D conversion for computer processing, the resolution of a normal camera is about 720 x 480 pixels, so if the license plate is within the field of view, This is because if the image is taken small, the character data will be destroyed and it will not be possible to identify it.

Generally, for vehicles, the license plate mounting position range is stipulated by law. Most passenger cars are mounted in the number stay position at the time of manufacture, so the vertical variation of the license plate is about ± 40 cm at most. On the other hand, in the vehicle width direction (direction parallel to the road surface), it is assumed that there is considerable variation (± 80 cm or more) on the general road and in the parking lot. Further, the license plate itself is horizontally long compared to the shape of the shooting range of the camera. Therefore, when images are taken without missing the surroundings, the margin in the horizontal direction is considerably smaller than in the vertical direction.

Therefore, in consideration of variations in the approaching vehicle position, it is difficult to zoom in and shoot a single image without missing the surroundings without obtaining the vehicle license plate position. In order to solve that,
As a device for capturing the image of the license plate mounted on the vehicle, the image captured by one wide-field camera is processed to detect the plate position and the serial number, and the high image captured by the multi-camera based on the position information is processed. There is known one that processes only one of the resolution images that is determined necessary (for example, Japanese Patent Laid-Open No. 62-196686).

[0005]

However, in the conventional method of processing the image to detect the position of the license plate plate or the character position, and thereby to position the license plate of the vehicle, the positioning is not sufficient. However, there is a problem in that it requires a long calculation time and causes a decrease in processing speed.

SUMMARY OF THE INVENTION It is an object of the present invention to solve the above-mentioned problems and to provide an apparatus capable of photographing a license plate portion at a high speed at a high speed.

[0007]

According to the apparatus of the present invention, a television camera is installed in front of or behind a vehicle, two mirrors, a convex surface and a flat surface, are arranged in front of the camera, and the mirror is rotated to rotate the camera. As a field of view of the vehicle, there is a mechanism for switching and setting two types of fields: a large field of view including the entire front or rear of the vehicle, and an enlarged field of view that vertically covers the license plate mounting position specified by law. In addition, it has a mechanism that captures images by irradiating a fan-shaped laser beam diagonally and horizontally to the road surface near the hood or rear gate in synchronization with the time when the camera observes the entire area including the front or rear of the vehicle with the convex mirror. In addition, a circuit that processes this and calculates the vehicle center position, and controls the rotation angle of the plane mirror from the obtained position to cover the vertical position of the license plate mounting position in the center of the vehicle when the camera field of view is reached. And a control circuit for synchronously capturing an image. As a result, the image including the license plate is zoomed up and photographed, and the characters on the license plate including the name of the Land Transport Office can be processed with high accuracy.

[0008]

The present invention focuses on the fact that the shape of the vehicle body is generally centrally symmetric, detects the vehicle body center position by high-speed processing of the waveform extracted by the slit projection method, etc. The license plate image can be zoomed up and taken as a single image without missing the surrounding area.

The slit projection method is generally used for the purpose of acquiring the shape of an object, but in the present invention, it is used as a waveform extraction method for positioning the center of the vehicle body. Also,
To control the viewing angle and viewing position of the camera, two types of mirrors, a convex surface and a flat surface, are switched and used.

Although the slit projection method and the method of processing images of two viewing angles to improve efficiency are known techniques, a combination of these two methods as a configuration proposed by the present invention results in one unit. The greatest feature is that it is possible to obtain a remarkable effect that the license plate image can be magnified and photographed with the minimum number of cameras and without missing the surroundings.

According to the present invention, (1) the center coordinates in the left-right direction of the position where the license plate is mounted can be extracted by the slit projection method which is a fast / simple and stable slit projection method for the axis of symmetry of the vehicle body, Since only one image is captured based on the obtained position information and the license plate can be projected in that area without missing the surrounding area, the license plate image can be acquired faster and more stably than the conventional method. Will be possible.

[0012]

[Example] An example of the implementation of the present invention will be described below by taking an example of application to character recognition of a license plate in entering / exiting management of a parking lot.

FIG. 1 is a block diagram of an embodiment of the present invention, showing a device configuration for enlarging and photographing a license plate image of a vehicle entering a parking lot. Further, FIG. 2 shows an arrangement example of devices when enlarging and photographing a license plate image of a vehicle entering a parking lot.

In the figure, reference numeral 15 denotes a rotating mirror, which has a convex surface and a flat mirror, and the back surfaces of the mirrors are bonded together to form a unit. The rotary mirror 15 is configured so that it can be controlled to rotate about its axis by the digital servomotor 14. Further, the laser light projecting device 13 is arranged near the hood or rear gate of the vehicle so as to radiate a fan-shaped laser light obliquely and horizontally to the road surface.

The TV camera (ITV camera) 10 is arranged in front of (or behind) the entering vehicle 20 in such a manner that its optical axis deviates from the irradiation direction of the laser light projecting device 13. Then, the fan-shaped laser light emitted from the laser light projecting device 13 projects the cut surface onto the vehicle hood, and at the same time, the convex mirror portion of the rotating mirror 15 gives the television camera 10 an image including the front portion of the vehicle as a visual field, and By controlling the rotation angle of the rotating mirror 15, the plane mirror portion is zoomed up so that the TV camera 10 can cover the vertical direction of the license plate mounting position of various vehicle models of various passenger cars, and the field of view positioned in the vehicle width direction. The TV camera 10 is arranged so that the

The frame memory 11 is a TV camera 1.
This is a device for storing an image input from 0. The license plate image obtained from the enlarged field of view is temporarily stored in the frame memory 11 and read out for character recognition and the like by a control means such as a computer (not shown). The central position detection device 12 is a device that extracts the central position of the entering vehicle 20 by the slit projection method when the vehicle enters the warehouse.

The operation of this embodiment is as follows.
It should be noted that the correspondence between the position in the image and the position in the real space is obtained in advance for each of the convex mirror and the plane mirror in the rotating mirror 15, before the actual number plate imaging.

Immediately after entering the vehicle, the rotating mirror 15 is switched to the convex mirror side by using a gate opening / closing control signal which is essential for warehousing management, and a large field of view is obtained. At the same time, the laser light projecting device 13 irradiates the fan-shaped infrared laser light obliquely and horizontally on the road surface, and projects a line by light cutting on the bonnet.

FIG. 3 and FIG. 4 show the acquisition status of the large field of view images seen from the top and side surfaces, respectively. In the figure, 30 is a laser beam, 31 is a field of view of the television camera 10, 32 is a projection locus, and 40 is a license plate.

The projection locus 32 of the image captured by the TV camera 10 has a high brightness due to the laser light.
Becomes a waveform 32 'as shown in FIG. For convenience, as shown in FIG. 5, the direction corresponding to the center axis of the vehicle body in the two-dimensional coordinates is the y-axis, and the other direction is the x-axis. Only the part of the laser locus is extracted from the data obtained by the above method, and i = 0, 1,
2, ..., (Point number in x direction-1) is set as a point sequence, and this is a one-dimensional waveform format, that is, y [i] = f (x
Convert to [i]) format.

The waveform thus obtained is a locus obtained by cutting the vehicle body obliquely horizontally to the road surface. Since the vehicle body of a general passenger car is bilaterally symmetrical, it has a shape that is folded back and symmetrical with respect to the center position of the vehicle. Therefore, if we apply the correlation function in analytic,

[0022]

[Equation 1]

I = 0, 1, ..., (Number of pixels in x direction−
The calculation is performed for 1), and the symmetry center, that is, the x coordinate of the vehicle center axis is obtained as the value of i at which r [i] is maximized.
Note that in this calculation, due to the so-called Wiener-Khintchine relationship, speedup by convolution using the fast Fourier transform can be applied. Also, since the correlation function value has a strong peak, it is possible to speed up by thinning out. When the thinning method is used, a 32-bit microprocessor (clock
The central position calculation time of the waveform data obtained from an image of 512 × 512 pixels by 20 MHz, with a floating point coprocessor built-in is actually less than 4 ms, which is sufficiently faster than the frame rate of a typical TV camera of 30 ms. Required to.

Since this operation is based on one-dimensional data processing,
The calculation time can be made extremely shorter than that of the conventional method (for example, Japanese Patent Laid-Open No. 62-196686) in which the number plate position is extracted or the two-dimensional data is processed by extracting a series of characters.

Since the calculation using the correlation function is generally resistant to noise on the waveform, the vehicle center position can be measured stably and at high speed from the trajectory of the laser light by using the shape symmetry of the vehicle. You can Therefore, based on the vehicle center position in the image obtained by the above method, the actual vehicle center position can be estimated by using the correspondence with the actual space position which is obtained in advance.

After the vehicle center position is obtained, the plane mirror of the rotary mirror 15 is rotationally controlled as shown in FIG. 6, and an image is taken when the center of the camera visual field 60 is closest to the vehicle center position. It is possible to capture 40 in a large size in the image without missing its periphery.

The license plate image thus obtained is zoomed up, and the character crushing when converted to a digital image is reduced. Therefore, if it is processed by a so-called license plate character recognition device, it is included in the name of the Land Transport Office. It is obvious that the recognition rate of kanji etc. can be improved.

If the device according to the present invention is not used, the deviation in the vehicle width direction when entering the parking lot is generally ± 30.
Since it is assumed that the size is about cm, it is necessary to take a large number of images and process them with the license plate character recognition device in order to take the image at the same magnification as the device according to the present invention without missing the surroundings. It is clear that the recognition throughput must be reduced because the image of is processed.
However, according to the present invention, the number of images processed by the license plate character recognition device is reduced, and the processing time in the license plate character recognition device can be shortened.

[0029]

As described above, in the present invention, focusing on the fact that the vehicle body shape is generally center-symmetrical, the waveform extracted by the slit projection method is processed at high speed to detect the vehicle body center position, and the information thereof is obtained. It is possible to configure a device that can take a license plate image as a single image with a large size of the license plate image without missing the surrounding area.

Although the technique itself for controlling the viewing angle and viewing position of the camera by the slit projection method and two kinds of mirrors is a known technique, by combining these two methods, a license plate image can be obtained with a smaller device configuration. There is a remarkable effect that it is possible to take a magnified image of.

As a conventional method, an image picked up by one wide-field camera is processed to detect a plate position and a serial number, and it is judged that it is necessary based on the position information. When compared with a method of processing only one sheet (see, for example, Japanese Patent Laid-Open No. 62-119686), the present invention (1) determines only the center position based on the symmetry of the vehicle body shape as positioning information, and Faster than image processing for position and sequence number detection. (2) A desired field of view is determined by controlling the rotation of the mirror based on the positioning information, and this is taken, so multiple cameras are not required. (3) It is different from the conventional method in that only one image is taken based on the positioning information, and when an auxiliary light source such as a strobe is used, the number of lights or the amount of light can be reduced.

[Brief description of drawings]

FIG. 1 is a diagram showing a device configuration according to an embodiment of the present invention in the case of application to license plate character recognition in entry / exit management of a parking lot.

FIG. 2 is a diagram showing an arrangement of main parts in the device shown in FIG. 1 when applied to license plate character recognition in entry / exit management of a parking lot.

FIG. 3 is an explanatory diagram illustrating acquisition of a large-field-of-view image viewed from the top according to an embodiment of the present invention.

FIG. 4 is an explanatory diagram illustrating acquisition of a large-field-of-view image viewed from the side according to an embodiment of the present invention.

FIG. 5 is a diagram showing a locus shape on a vehicle body obtained by projecting a fan-shaped laser beam in an example of the present invention.

FIG. 6 is an explanatory diagram of a license plate photographing with an enlarged field of view in the embodiment of the present invention.

[Explanation of symbols]

 10 TV Camera 11 Frame Memory 12 Central Position Detection Device 13 Laser Light Projection Device 14 Digital Servo Motor 15 Rotating Mirror 20 Cargo Vehicle 30 Laser Light 31 Field of View of TV Camera 32 Projection Trajectory

Claims (1)

[Claims] 1. A television camera installed in front of or behind a vehicle, two convex and flat mirrors arranged in front of the television camera, and a field of view of the television camera by rotating the mirror. A mechanism that switches and sets the large field of view including the entire front or rear of the vehicle and the area of the enlarged field of view that covers the license plate mounting position specified by law in the vertical direction, and is diagonal to the road near the hood or rear gate of the vehicle. A mechanism for horizontally radiating a fan-shaped laser beam, a means for observing a large field of view including the entire front or rear of the vehicle through the convex mirror by the television camera, and capturing an image including a projected locus of the laser beam, And means for calculating the vehicle center position,
A rotation angle of the plane mirror is controlled from the obtained position, and means for capturing an image in synchronism with the time when the area covering the license plate mounting position in the central portion of the vehicle in the vertical direction becomes the camera visual field, License plate image capturing device.