JPS63122000A - Number plate reader - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a license plate reading device suitable for use in a toll collection system on a toll road or a toll collection system in a toll parking lot. Technology] In recent years, as a toll collection system for toll roads, a system has been considered that reads the license plate of a vehicle, determines the vehicle type from this read information, and calculates the toll based on this vehicle type information. Here, the reading of the license plate of the vehicle is performed using software processing by a computer as follows. In other words, the front shape of the vehicle is imaged using an imaging means such as a television camera, this image information is A/D converted, then imported into a frame memory, and image processing is performed according to a predetermined program procedure to determine the size of the license plate or the like. The plate number is encoded to provide vehicle type identification information.

[Problems to be solved by the invention] However, as mentioned above, in the conventional license plate reading device, image processing is performed on all images in the frame memory according to a program procedure pre-installed in the computer. was. Therefore, it took a long time to complete the entire process. As an example of a toll collection system for a toll road, if this license plate reader is applied to an unmanned system at entrance f-), the time required for the vehicle to proceed from the installation location of the license plate reader to the installation location of the automatic ticket issuing device. The vehicle type must be determined. In the past, the long processing time of the number rating reader was a major problem in making entrance darts unmanned. In addition, since the front image of the vehicle stored in the frame memory includes the headlights and other parts in addition to the license plate, the accuracy of image processing for the license plate area may deteriorate. There was a risk of misinterpretation.

The present invention was made based on these circumstances,
The purpose and focus is to provide a license plate reading device that can shorten processing time and improve reading accuracy.

[Means for solving the problems] In order to solve the above problems and achieve the purpose, the present invention has the following features:
The wheel distance and passing position of the vehicle are measured using the output from a wheel distance measuring device that is sensitive to the pressure on the vehicle's wheels, the vehicle's license plate installation position is estimated based on the measured value information, and this estimated position is calculated. The license plate information is read by cutting out a part of the image information of the vehicle taken by the imaging means based on the information and performing predetermined image processing on the cut out image.

[Effects] By taking such measures, the number plate is attached to the center position of the vehicle's wheels, so the number plate attachment position with respect to the vehicle can be quickly estimated, and the number plate can be attached to the front of the vehicle according to this estimated position. Since only the license plate portion is extracted from the image and subjected to image processing, there is no need to perform image processing on unnecessary portions for reading the license plate.

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

FIG. 1 is a state diagram illustrating an example of the arrangement when the number/grave reading device of the same embodiment is installed at a toll gate dirt of a toll road. In the same figure, the number plate (N) should be attached at least a little to the front.The vehicle 1 is to proceed in the direction of the arrow F in the figure on the tollgate dirt entrance lane 2, with the above lane 2 in between. Vehicle separators 4, 4 and shutter devices 5, 6 of the vehicle type discrimination device 3 are arranged facing each other. Further, a wheel distance measuring device 6 of the vehicle type discrimination device t3 is buried in the lane 2. The vehicle separators 4, 4 are composed of a plurality of phototube emitters and receivers, and detect each passing vehicle and output a vehicle detection signal. death.

This measures the vehicle's wheel width and wheel pressure position. Further, the shutter devices 5, 5 are connected to the vehicle separator 4.
．． The vehicle detection signal serves as a timing signal for commanding the shutter-on of the television camera 8 attached to the roof 7 of the toll plaza above the car 2. The television camera 8 has a field of view indicated by a person in the figure, and is set so that at least the license plate (N) of the vehicle 1 falls within this field of view A.
Furthermore, the television camera 8 is equipped with an infrared flash 5.9 that is filtered so as to emit long wavelength light that cannot be seen by Iruma, and the television camera 8 is particularly sensitive in the infrared region. It uses a built-in image sensor. 10 is a license plate discrimination device, and the vehicle type discrimination device 3. The license plate N attached to the front of the passing vehicle 1 is determined based on output signals from the shutter device 5 and the television camera 8.

Here, the positional relationship between the attachment position of the number grade N of the vehicle 1 to this device and the passing position is 5VC as shown in FIG. That is, the vehicle 10 wheel distance is expressed as the distance from both ends of the W9 wheel distance detector 6, that is, the distance from the road shoulder to both sides of the wheel, respectively, tl and t2°.
If the distance to the edge of is tS and the distance from the edge of the field of view of the television camera 8 to the center of the vehicle 1 is Xn, then
3 + Xn = t l + (W / 2)
``' The relationship shown in (1) is. Therefore, if we assume that the license plate N is attached to the center of the vehicle 1, and if one end P of the person in the field of view of the television camera 8 is taken as the reference point, then the license plate N is position, that is, Xn
is expressed as Xn=(11+W/2')-13-(2).

Further, the positional relationship of the license plate N with respect to the vehicle 1 is as shown in FIG. That is, if the vertical coordinate axis with respect to the vehicle body is Y and the horizontal coordinate axis is X, and if we toss the reference point (0,0) on the lane 2 at one end P of the field of view A of the television camera 8 in FIG. The position of N is (Xn.

Yll).

Generally, the mounting position of the number plate N differs depending on the vehicle type, as shown in FIG. That is, the license plate of a regular passenger car exists within the range α in FIG. 4, and the license plate of a large vehicle exists within the range β in the same figure. Expressing these in the coordinate relationship described above, the license plate presence area of a regular passenger car is (XI, Yl) = (Xnl±x 1, Yll±yJ) ++・
(3) becomes. Also, the area where the license plate of a large vehicle exists is (X2.Y,?) = (Xn2±x 2 , Yn2shiy 2 )...
・(4) becomes.

On the other hand, the wheelbase distance measuring device 5 is composed of left and right contact bodies 50 and 50', as shown in FIG.
, s o' are two upper contacts 51 and 51' formed of flat metal such as stainless steel, and these upper contacts 5.
A resistance element electrically connects a plurality of metallic lower contacts El, 52' disposed directly below 51' at predetermined intervals, and the contacts that are matched with each other at each lower contact 52, 52'. 53, 53'.

Therefore, when the left and right wheels T and T' of the vehicle 1 are not pressing each contact body 50 and 50', the contact body 50
, 50', and the resistance value between a/ and b/ is R, the wheels T and T' are connected to each contact body 50,
The resistance values when the upper contact 51.51' is deformed by pressing 50' and short-circuited with the predetermined lower contact 52.52' are represented by (R-rl) and (R-rl), respectively.

However, rl and rl are the total values of the resistance elements 53 and 53', respectively, where no current flows due to a short circuit.

Therefore, the distance t1. from both sides of the wheels T1, T' to the road shoulder. t2 and wheel distance W are expressed by the following equations (5) to (7).

tx=GcR-rl)-(5)L
2 = G (R-r 2 )...-(
6) It is a control block diagram. In the figure, an A/D converter 61 converts an analog image signal from a television camera 8 into a digital image signal of, for example, 512 x 512 x 8-bit gradation, and the digital image signal converted here is divided into frames by a distributor 62. It is distributed to a memory 63 and a differentiation circuit 64. The frame memory 63 stores an original image of 8-pit gradation. On the other hand, the differentiation circuit 64 calculates the difference in density between horizontally adjacent pixels and outputs its absolute value, and this absolute value output is binarized by a binarization circuit 65 button. That is, the value of 8 bits per pixel is compared with a predetermined threshold value, and if the value is larger than the threshold value, it is set as "1", and if the value is smaller than the threshold value, it is set as "0";
Compress to binary image data of 1 pixel. Further, the output of the binarization circuit 65 is distributed by a distributor 66 to a differential binary image memory 67 and a feature extraction circuit 68. The differential binary image memory 67 stores a binarized differential binary image. On the other hand, the feature extraction circuit 68 extracts the rt of the differential binary image.
10J is used to extract areas where there are many changes, that is, areas where there are many density changes in the original image. Specifically, the position information of the license plate is extracted, and this extraction signal is stored in the feature memory 69. Note that the frame memory 63. The differential binary image memory 67 and the feature memory 69 are composed of the same storage element.

In the figure, a CPU circuit 70 is in charge of number grade discrimination [necessary arithmetic processing control, and connects the frame memory 63, differential binary image memory 67, and feature memory 69 via a pass line 71; Further, a vertical histogram memory 72 and a horizontal histogram memory 7
3. License plate processing section 74. A license plate position memory 75 and an input/output interface circuit 76 are connected. The vertical histogram memory 72 calculates coordinates
This is a memory that scans in both the direction and the Y direction and counts and records the number of times the black level is "1" in the vertical direction. The horizontal histogram memory 73 is a memory that scans the image including the license plate in the feature memory 69 in the coordinate X direction and the Y direction and counts and records the frequency of the black level "1" in the horizontal direction. The license plate discrimination processing unit 74 also has a storage area for storing license plate reference values, encoded license plate number reference values, etc. in advance, and performs predetermined processing under the control of the CPU circuit 70. According to the program, each reference value and each memo +)63,6
A comparison is made with the stored contents of 7.69° 72 and 73 to determine the size of the number grade N of passing vehicle l and to code the plate number.

The number grade position memory 75 is connected to the vehicle type identification device 3.
The number great position information obtained based on the output signal from the input/output interface circuit 76 is taken in and stored. The input/output interface 76 is
It transmits and receives signals between the CPU circuit 7o, the vehicle type discrimination device 3, and the shutter device 5), that is,
Vehicle detection information 1 wheel distance and vehicle traffic position information are provided from the vehicle type discrimination device 3, and information for determining the size of the number grade and coded information of the grade number are provided from the CPU circuit 7o.

Next, the operation of the license plate reading device of this embodiment will be explained with reference to the flowchart shown in FIG. Now, when the passing vehicle 1 enters the tollgate and approaches the installation position of the vehicle separators 4, 4, a vehicle detection signal is sent from the vehicle separators 4, 4 to the input/output interface circuit 76 of the number rate discrimination device 10. Then, in response to this vehicle detection signal, the wheelbase distance measuring device 6 is notified of the passing position of the vehicle and the start of measuring the wheelbase distance (STJ).
.

When the vehicle 1 approaches the shutter device 5.5, a vehicle detection signal is sent from the shutter device 5,5 to the input/output interface circuit 76 (ST, ?).

Then, in response to this vehicle detection signal, the TV camera BVC
On the other hand, the start of photographing the curved surface of the vehicle is notified.

(ST4). Then, a front image of the vehicle 1 is captured by the television camera 8, and this captured image is taken into the number grade discrimination device 10. (ST5).

Note that at this time, the infrared flood 9 illuminates the field of view A.

On the other hand, in response to vehicle detection signals from the shutter devices 5, 5, the wheelbase distance measuring device 6 is notified of the termination of the wheelbase distance measurement operation (ST5), and then the wheelbase distance measuring device 6 identifies the license plate. The wheelbase distance W and passing position information t of the vehicle 1 are sent to the input/output interface circuit 76 of the device 10 (ST7).

As a result, in the license plate discrimination device 1o, the position of the license plate N is calculated based on each measurement value during the arithmetic processing of the CPU 70 according to a predetermined program in the license plate discrimination processing section 74, and this calculated value and Based on the wheelbase distance W, the position of the license plate N of the vehicle 1 is estimated from the above formula (3) (for regular cars) or the above formula (4) (for large cars), and is stored in the license plate position memory 75. (ST8) 6 On the other hand, the analog video signal of the front part of the vehicle including the driving license plate N captured and captured by the television camera 8Vc is converted into a digital value by the A/D converter 6, and then differentiated with the frame memory 63. circuit 64. In the frame memory 63, for example, a 512 x 512 x 8 bit digital image showing the field of view A is stored on one screen.

Note that this screen is not rewritten unless a new image is input. On the other hand, the differentiation circuit 64 calculates the difference in density between horizontally adjacent pixels of the digital image, and outputs its absolute value to the binarization circuit 65. This can be easily constructed using known technology,
This output data is 512x512x8 bits.

In the binarization circuit 65, the value of 1 pixel (8 pixels) (7) is compared with a predetermined threshold value, and if the threshold value is a large value, it is "1", and if it is a small value, it is "0". After being compressed into binary image data of 1 pixel and 1 pit, the data is distributed to the differential binary image memory 67 and the feature extraction circuit 68. In the feature extraction circuit 68, rL10J of the differential binary image
Areas with large changes, ie, areas with large density changes in the original image, are extracted and stored in the feature memory 69Vc.

Therefore, in the CPU 70, only the license plate portion is extracted from the differential binary image from the feature memory 69 based on the license plate position information stored in the license plate position memory 75 (ST9), and only this portion is extracted. A predetermined memory scan is performed on the memory contents (sTIO). As a result, histograms of binary images in the vertical and horizontal directions of the coordinates are stored in the vertical histogram memory 72 and the horizontal histogram memory 73, respectively (STJJ), and the license plate discrimination processing unit 74 uses the stored contents of these memories to Number play) N
The size is determined and the grade number is coded (ST J, ?).

In this way, the license plate N is recognized and notified to a higher-level device such as a data processing device as a vehicle type identification device (
ST13).

According to the present embodiment configured in this way, the entire license plate information is not read by performing image processing over the entire area of the viewer's field of view of the television camera 8;
Measuring the wheel distance and passing position of the passing vehicle 1 using the wheelbase distance measuring devices 6, 6, estimating the parking plate position from this information, and processing only limited images of the license plate portion. Since the license plate information is read using the image processing system, image processing can be performed at high speed.

Furthermore, since noise parts such as those caused by headlights are discarded, highly accurate reading becomes possible. Therefore, by applying this embodiment to a toll road toll collection system,
The speed and accuracy of toll collection operations will be increased, and unmanned toll gates may be realized.

Mugio, the present invention is not limited to the above embodiments.
For example, it is of course applicable to paid parking lots.

[Effects of the Invention] As described in detail above, according to the present invention, the number plate mounting position on the vehicle is quickly estimated, and only the number plate portion is extracted from the vehicle front image according to this estimated position.゛ Image processing is performed, reducing processing time.

Also, it is possible to provide a license plate reading device that can improve reading accuracy.

[Brief explanation of the drawing]

Figures 1 to 7 are diagrams showing one embodiment of the present invention, in which Figure 1 is a schematic diagram showing the installation situation of this device at a tollgate, and Figure 2 is a diagram showing the passing position of the license plate and the vehicle. Figure 3 is a diagram showing the positional relationship between the vehicle and the license plate, Figure 4 is a diagram showing the mounting position of the license plate by vehicle type, and Figure 5 is a diagram showing the specific configuration of the wheelbase distance measuring device. FIG. 6 is a control block diagram of the number rate discrimination device, and FIG. 7 is a flowchart for explaining the operation of the device. DESCRIPTION OF SYMBOLS 1...Vehicle, 3...Vehicle type identification device, 4...Vehicle separator, 5...Shutter device, 6...Wheel distance measurement device, 8...TV camera, 10...Number Plate discrimination device.

Claims (1)

[Claims] A device for reading license plate information by performing predetermined image processing on image information of a vehicle imaged by an imaging means includes a wheel distance measuring device that is sensitive to wheel tread pressure of the vehicle; A measuring means for measuring the wheel distance and passing position of a vehicle; a position estimating means for estimating the license plate mounting position of the vehicle based on the measurement value information by the measuring means; image cutting means for cutting out a part of the image information of the vehicle imaged by the imaging means based on the image capturing means;
A license plate reading device comprising an image processing means for reading license plate information by performing predetermined image processing on the image cut out by the cutting means.