JPH0816713A - Character recognizing system - Google Patents

Abstract

PURPOSE:To provide the character recognizing system with which characters or symbols described on a mobile object can be recognized. CONSTITUTION:A position detector 11 detects the mobile object and outputs a position detecting signal. When the position detecting signal is received, a TV camera 12 with electronic shutter function operates an electronic shutter and makes an auxiliary optical device 13 emit light at the time. An image input device 14 writes an image, which is photographed by the TV camera 12 with electronic shutter function, in a frame memory and when writing is completed, a write inhibited state is provided. A character recognizing device 15 reads the image from the write inhibited frame memory and performs character recognition.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention recognizes characters and symbols written on a moving object by combining a position detection device, a light emitting section for emitting auxiliary light and a TV camera with an electronic shutter function. This is a character recognition system for operating the TV, and is characterized by calculating the installation position of the TV camera and the shutter speed from the speed of the moving object and the size of the target area.

A means for fetching character information into a computer is an indispensable data input means in the case of converting an office work process into an electronic computer process for efficient operation. In recent years, as a part of factory automation, there are application examples such as using character recognition for quality inspection of characters engraved on a product and reading an automobile number plate. In the future, if a technology for recognizing characters in the moving state of the target moving object will be developed, quality inspection can be performed more efficiently than before in factory automation and the like.

[0003]

2. Description of the Related Art Most conventional character recognition apparatuses use a scanner as an input medium and perform character recognition based on a binary image obtained as a result. In this case, the target is a printed matter or a handwritten character, and it is premised that the target is a medium such as paper and the target is still.

On the other hand, there is also a method of character recognition using a camera as an input medium. In this case, it is possible to recognize characters from various objects other than printed matter. However, even when the camera is used as the input medium, it is possible to recognize only an object that can be regarded as stationary within the frame of 30 images per second, which is the processing speed of the camera.

For example, FIG. 4 is a diagram showing an example of a character recognition device by camera input. In the figure, 1 is a camera,
Reference numeral 2 is an A / D converter, 3 is a frame memory, 4 is a recognition processing unit, and 5 is an auxiliary storage device. The camera 1 is, for example, a TV camera, and has a two-dimensional image sensor. The camera 1 constitutes a video input unit. The A / D converter 2 converts an analog signal output from the camera 1 into a digital signal. The digital signal output from the A / D converter 2 is stored in the frame memory 3. The A / D converter 2 and the frame memory 3 form an image data conversion unit. The recognition processing unit 4 reads the image in the frame memory 3 and performs character recognition. The recognition result is stored in the auxiliary storage device 5. Recognition processing unit 4
The auxiliary storage device 5 constitutes a character recognition unit.

The camera 1 is a general camera having an analog output. Using the camera 1 as an input medium, the image data converter digitizes the analog signal and stores it in the frame memory 3 having a capacity of one screen. The character recognition unit performs character recognition based on the multivalued image stored in the frame memory 3, and stores the recognition result in the auxiliary storage device 5. The character recognition unit may be a dedicated device or a general-purpose computer.

[0007]

DISCLOSURE OF THE INVENTION Problems to be Solved by the Invention As described above,
In the conventional character recognition, it is possible to recognize only a stationary object or a character of a moving speed object that does not hinder the recognition. If the blur caused by moving does not affect the character recognition, character recognition is performed in the moving state, but in general, characters written on a moving object are input at high speed. Requires a process of stopping the object, inputting an image, and then restarting the movement.

For example, Table 1 and Table 2 show at what movement speed the movement stop and movement restart processes are not necessary.
And Table 3 demonstrates.

[0009]

[Table 1]

[0010]

[Table 2]

[0011]

[Table 3]

Table 1 shows shooting conditions as the performance of the camera. Under the shooting conditions of Table 1, the shooting range and the actually measured size of one pixel are as shown in Table 2.

In Table 2, the condition of the widest photographing range is that the focal length is 12.5 mm and the target distance is 6 m. Under this condition, the actual measurement range for one pixel is 1.9.
It is cm × 1.9 cm. On the other hand, the narrowest imageable range is when the focal length of the lens is 50.0 mm and the distance to the target is 3 m. The actually measured size of one pixel at this time is 0.24 cm × 0.24 cm.

On the other hand, Table 3 shows the movement amount per hour, second, 1/3.
It is shown at 0 seconds. As can be seen from Tables 2 and 3, when the lens focal length is 12.5 mm and the distance to the target is 6 m, the measured range of one pixel is 1.9 cm × 1.9 c.
Therefore, if the amount of movement for 1/30 seconds is within 1.9 cm, blurring can be contained within one pixel. When the speed at this time is obtained from Table 3, it is found that the speed is 1 km / h. When the lens focal length is 50 mm and the distance to the target is 3 m, the movement amount for 1/30 seconds is 0.
If it is within 24 cm, it can be read without blurring. Judging from Table 3, the speed must be 1 km / h or less.

By the way, the speed of the belt conveyor of the production line is several meters / second to several tens of meters / second, and if it is assumed that the moving object is taken into the TV camera under the conditions of Table 1 and Table 2, The image is blurred and it is impossible to recognize the characters. The present invention has been made in view of this point, and an object thereof is to provide a character recognition system capable of recognizing characters, symbols, and the like written on a moving object.

[0016]

FIG. 1 is a diagram illustrating the principle of the present invention. In the present invention, in order to solve the above problems, the position detecting device 11, the TV camera with an electronic shutter function 12, the auxiliary light device 13, the image input device 14, and the character recognition device 15 are combined as shown in FIG. . The character recognition device 15 is the same as a conventional character recognition device.

The position detection device 11 detects whether or not an arbitrary object is present and outputs a signal. The TV camera 12 with an electronic shutter function is for converting an image into an analog signal at a preset shutter speed. The auxiliary light device 13 emits light by controlling the timing for starting light emission by a signal from the outside.

The image input device 14 has a function of converting an image photographed by the TV camera 12 into a signal of a format which can be processed by the character recognition device 15 and storing image data in a frame memory by a signal from the outside. ing. Although not shown, the frame memory exists in the image input device.
The character recognition device 15 is a device for recognizing characters or symbols written on a moving object based on the data stored in the frame memory.

[0019]

The operation of the system of the present invention will be described with reference to FIG. When an object passes in front of the position detection device 11, a position detection signal is output. This position detection signal is sent to the TV camera 12 and the image input device 14. The position detection signal is a signal that controls the timing of frame memory control of the auxiliary light device 13 and the image input device 14. The change over time in this case will be described with reference to the timing chart of FIG.

FIG. 2A shows an internal synchronizing signal of the TV camera 12. The TV camera 12 takes in an image according to this synchronizing signal. An ordinary TV camera, as shown in Fig. 2 (b), has a certain fixed interval of 1/30 seconds,
It receives light and converts it into electrical signals.

FIG. 3C shows the case where the electronic shutter function is used. In this case, light is converted into an electric signal only for a predetermined fixed period. For example, 1/100 second, 1 /
1000 seconds or so. Since the image is not converted into an electric signal during a period other than this fixed period, it is possible to capture an image instantaneously. For example, assume that the position detection device 11 outputs the signal shown in FIG. The position detection device 11 has a mechanism for instantaneously outputting a pulse when an object can be detected.

When the object passes, the position detecting device 11 outputs a position detecting signal as shown in FIG. 2D, and this position detecting signal is sent to the TV camera 12. When the TV camera 12 with the electronic shutter function receives the position detection signal,
The electronic shutter is operated at the next image capturing timing and a light emission signal as shown in FIG. The auxiliary light device 13 emits light upon receiving the light emission signal.

The position detection signal from the position detection device 11 is also sent to the image input device 14. Upon receiving the position detection signal, the image input device 14 receives a frame / frame after 1/30 seconds from the next timing (at the next next video capture timing).
Prohibits writing to memory. When the character recognition is completed, the write protection is released. It should be noted that not only an analog image signal but also a video capture timing signal is sent from the TV camera 12 with an electronic shutter function to the image input device 14.

[0024]

[Embodiment] In this embodiment, a case of recognizing a running train or a vehicle number of a train will be described. If the train speed is 80 km / h and the size of one character is 10
It is assumed that four characters are written in cm × 10 cm, and the distance between the characters is 5 cm. The camera used is the same as the conditions in Table 1. Under the above conditions, the calculation of the camera installation position and the electronic shutter speed is as follows.

Step 1 80 km / h → 22.2 m / sec → 74 cm (1/30
Second) Moves 74 cm in 1/30 seconds. Step 2 Width of vehicle number → 10 cm × 4 + 5 cm × 3 = 55 cm Step 3 In order to read the vehicle number for any 1/30 seconds, the width of the screen is 74 cm + 55 cm = 1.
29 cm or more is required.

Step 4 Normally, the aspect ratio of the TV is 3: 4, so the height of the screen is 1.
It is 29 cm × 3/4 = 97 cm. Here, if the effective range of the screen is 97 cm × 129 cm, the image of the vehicle number portion can be captured. Step 5 When the relationship between the focal length of the TV camera and the TV camera and the train is summarized from step 4, the vehicle number can be read in the case of the combinations in Table 4.

[0027]

[Table 4]

Step 6 When the actually measured value of one pixel is obtained from the effective screen range, it is 129 cm / 320 = 0.4 cm Step 7 Even if the speed of 80 km / h is used, in order to keep the blurring of the image within one pixel, Step 1 From the second speed obtained in step 1 and the actually measured value of one pixel obtained in step 6, it is obtained as follows. 2220 / 0.4 = 5550 That is, the shutter speed may be set to 1/5550 seconds.

The above conditions are summarized in Table 5.

[0030]

[Table 5]

FIG. 3 is a diagram showing a structural example of an apparatus for carrying out the present invention. In the figure, 21a is an infrared light emitting unit, 21b is an infrared light receiving unit, 22 is a TV camera with an electronic shutter function, 23 is a light emitting unit, 24 is a character recognition device, and 25 is a character recognition device.
Is an A / D converter, 26 is a frame memory, 27 is a character recognition unit, 28 is a recognition result output unit, 29 is a template character storage unit, T is a train, and R is a rail.

Infrared ray emitting section 21a and infrared ray receiving section 21b
When the train T passes between the two, the infrared ray receiving section 21b outputs a position detection signal as shown in FIG. 2 (d). Upon receiving this position detection signal, the TV camera 22 with an electronic shutter function operates the electronic shutter and sends a light emission signal to the light emitting unit 23.

The TV camera signal from the TV camera 22 is input to the character recognition device 24. The position detection signal from the infrared light receiving unit 22 is also input to the character recognition device 24. The character recognition device 24 includes an A / D converter 25 and a frame memory 2
6, a character recognition unit 27, a recognition result output unit 28, and a template character storage unit 29. A / D converter 25
Converts an analog TV camera signal into a digital signal. The digital signal output from the A / D converter 25 is written in the frame memory 26. Character recognition unit 2
7 reads the image data of the frame memory 26,
The characters are cut out, the characteristics of the input characters and the characteristics of the standard characters are compared, and the character type of the input characters is determined. The determined character type is output from the recognition result output unit 28 to the outside. The template character storage unit 29 stores the characteristics of standard characters.

The operation of the system shown in FIG. 3 will be described.
Most of the trains T have a vehicle number on the side surface of the vehicle body. The system configuration for recognizing the vehicle number is the infrared emitting section 21a and the infrared receiving section 2 as a position detecting mechanism.
1b is used in combination. When the train T is not running, the infrared ray receiving section 21b always receives the infrared ray emitted by the infrared ray emitting section 21a. When the train T approaches and blocks the infrared rays, it is determined that the train is detected. Hereinafter, processing is performed based on this signal.

The TV camera 22 which receives the position detection signal
Emits auxiliary light in synchronization with the beginning of the next frame. The image thus captured is the character recognition device 2
4 and the vehicle number is analyzed. The character recognition device 24 which has received the position detection signal, as shown in the timing chart of FIG.
At the end of / 30 seconds, writing to the frame memory 26 is prohibited.

[0036]

As is apparent from the above description, according to the present invention, it is possible to recognize a character or a figure even in an object which is moving at high speed.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a timing chart of the present invention.

FIG. 3 is a diagram illustrating an embodiment of the present invention.

FIG. 4 is a diagram showing an example of a character recognition device using a camera as an input medium.

[Explanation of symbols]

 1 Camera 2 A / D Converter 3 Frame Memory 4 Recognition Processing Unit 5 Auxiliary Storage Device 11 Position Detection Device 12 TV Camera with Electronic Shutter Function 13 Auxiliary Light Device 14 Image Input Device 15 Character Recognition Device 21a Infrared Light Emitting Unit 21b Infrared Light Reception Part 22 TV camera with electronic shutter function 23 Light emitting part 24 Character recognition device 25 A / D converter 26 Frame memory 27 Character recognition part 28 Recognition result output part 29 Template character storage part T Train R Rail

Claims (3)

[Claims] 1. A position detecting device (11) for detecting that a moving object moving on a predetermined orbit has passed a predetermined position, and a TV camera (12) with an electronic shutter function capable of varying a shutter speed. , An auxiliary light device (13) used as an auxiliary light and an analog TV camera signal output from a TV camera (12) with an electronic shutter function are converted into a digital signal, and the digital signal obtained as a result of the conversion is written in a frame memory. A TV camera with an electronic shutter function, which comprises an image input device (14) and a character recognition device (15) for identifying characters and figures from a multi-valued image stored in a frame memory of the image input device (14). (12) is a position detection device (1
When the position detection signal is output from 1), the electronic shutter is operated at the next image capturing timing, and the auxiliary light device (13) emits light in synchronization with the operation of the electronic shutter. . 2. A TV camera (12) with an electronic shutter function, wherein the shutter speed of the TV camera (12) with an electronic shutter function is determined from the focal position of the camera, the moving speed of the target moving object and the size of the target area. The character recognition system according to claim 1, wherein the position is determined from the focal position of the camera, the moving speed of the target moving object, and the size of the target area. 3. A multi-valued image of a frame memory in which writing is prohibited after a predetermined time has elapsed from the image capture timing when the electronic shutter of the TV camera (12) with an electronic shutter function operates. On the other hand, the character recognition device (15) performs character recognition, and the character recognition system according to claim 1 or 2.