JP2718322B2 - Cylindrical character reader - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for reading characters drawn on the surface of a cylindrical body.

[0002]

2. Description of the Related Art Thin steel plates manufactured at steelworks are shipped as coils wound in a cylindrical shape. In this case, each coil is marked with a character such as an identification number.

[0003]

In such a character marking on a steel material, it may be difficult to recognize the marking correctly due to the following reasons. If there is uneven temperature on the coil surface, the adhesion of the marking ink changes, and characters are cut or jumped. During marking, if the distance between the marking nozzle and the coil surface fluctuates, characters will be blurred or blurred. If the printing surface has irregularities due to the cutting of the steel sheet, enlargement, reduction, or bending of the character occurs. Poor lighting on the printing surface makes it difficult to see clearly due to reflection and contrast. If the coil diameter changes, the reading area changes and reading is performed erroneously.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems, and has a cylindrical character reading method in which a plurality of TV cameras are used to photograph an object in an overlapping manner, and the data is interpreted using the most recognized data. It is intended to provide a device.

[0005]

In order to achieve the above object, a TV camera A installed at a position substantially at the center of the cylinder opposite to the cylindrical body and a reading target portion arranged on the left and right sides of the TV camera A overlap. The TV camera B and TV camera C that are configured to capture images and the characters of the image data input from the TV camera A are separated and the number of characters is the same as the number of marked characters.
If the number recognizes this separation character, wherein the time different from the number of characters that characters has said marking TV camera B, also to separate the text for the image data inputted from the TV camera C, the TV camera A, A character recognizing unit is provided for recognizing characters separated by input from B and C, and for recognizing the character of the input data by adopting the character at the same position that is best recognized.

An illumination unit is provided on the left and right sides of the TV camera A to irradiate the light from a light source to a reflecting plate to irradiate the surface of the cylindrical body.

[0007]

The data characters captured by the TV camera A arranged at the center are separated, and if the number of characters is equal to the number obtained by the data instructing the marking, all the marked characters are imaged. Therefore, recognition work of the separated data is performed. If the number of separated characters does not match the number obtained from the data instructing the marking, the characters are separated from the image data of the TV camera A and the image data of the TV cameras B and C. 3
If at least two of the TV cameras A, B, and C are configured to capture an image of an object in an overlapping manner, two or more image data can be obtained for one character at the same position. The character that can be recognized best is recognized as the character marked at that position. As described above, since the most recognized character is selected from two or more data, the probability of correctly recognizing the marking character is improved. In addition, TV
If no character can be separated from the image data of camera A, it means that an image of a position without marking is taken, or the lighting is too dark or inappropriate such as causing halation. Correct and perform the recognition process again.

[0008] Illumination units are provided on the left and right sides of the TV camera A so that the light applied to the reflection plate is irradiated on the target surface.
As a result, the arc surface can be illuminated fairly uniformly by the scattered light from different directions depending on the surface light source, so that accurate images can be input from the TV cameras A, B, and C, and correct recognition can be performed. Is more likely to be performed.

[0009]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a configuration of an embodiment of the present invention. 1
Is a hot coil, which is marked and flows on the belt conveyor. 2 is a TV camera with 3 units, A in the center,
Let B be the right side and C be the left side. An illumination device 3 is provided so as to irradiate the cylindrical surface from an oblique direction different from that of the TV camera. Reference numerals 4 and 5 denote sensors for detecting the passage of the hot coil 1, and reference numeral 6 denotes a recognition device that analyzes the image data of the TV camera 2 and recognizes the marking. 7 transports the hot coil 1 by a belt conveyor.

FIG. 2 is a sectional view taken along line XX of FIG. The illuminating device 3 includes a light source 8, a reflecting plate 9, and an enclosure surrounding the light source 8, and a transparent window for dust prevention is provided at a portion where light is emitted.
By using a sodium lamp as the light source 8 and using a diffuse reflection plate as the reflection plate 9, the hot coil 1 can be uniformly illuminated by the scattered light from the surface light source. Also, white ink characters printed by a sodium lamp can be clearly projected. By irradiating a coil surface with a character from an oblique direction that is greatly inclined at 10 to 30 degrees, regular reflection light enters the TV camera 2 installed at an angle nearly perpendicular to the surface, causing halation. Preventing.

FIG. 3 is a block diagram showing the configuration of the recognition device 3. An input selection unit 11 for selecting inputs of the TV cameras A, B, and C, an A / D converter 12 for converting the input data from analog to digital, and converting the converted data to blue, green,
An input buffer 13 to be stored in each of the red buffers, a bus 14, a CPU 15 for separating constituent characters of input image data, identifying characters and controlling the whole, an image processor 16 for performing image processing of image data, a CPU 15 A memory 17 for storing a program for determining the operation of the image processing, a binarized image memory 18 for storing data subjected to binarization processing and gradation image processing by the image processor 16, a gradation image memory 19, and an output for temporarily storing output data It comprises a buffer 20, a mixer 21 for synthesizing the output data, a D / A converter for converting digital data into analog data, and a CRT 23 for displaying output data.

FIG. 4 is an operation flowchart of this embodiment. When the sensors 4 and 5 shown in FIG. 1 detect the passage of the hot coil 1, the coil diameter is obtained from the speed of the belt conveyor 7, and the time required for the coil center to reach the center of the three TV cameras 2 is calculated. . At the time when the coil 1 is located at the center of the TV camera, image capture is started (step 3).
0). At this time, since the speed of the belt conveyor 7 is slower than the image capturing time, the image can be captured while the operation is continued, but the belt conveyor 7 may be stopped for a short time. Characters are cut out from the image captured by the TV camera A, and it is checked whether the number of characters matches the number of characters set for drawing the marking (step 31).

The character can be cut out one by one by examining the density projection distribution (horizontal and vertical directions) of the marking area. Further, since the number of characters and the character size are known in advance on the marking device side, the character spacing may be estimated and the character separation position may be examined in the vicinity thereof.

In step 31, the number N of characters cut out
i is compared with the number of characters Nd known in advance for the marking as described above. If there is no Ni, the marking surface is not facing the TV camera side, the coil 1 stop position is defective, or the illumination is defective. Since halation or stains on characters are conceivable, character reading is disabled (step 32). When Nd = Ni, since all the markings are included in the image data captured by the TV camera A, The cut character is read (step 33). Here, in reading the characters, first, the cut-out characters are normalized, superimposed on a registered reference pattern, and the character having the highest overlap is selected as a character represented by the selected reference pattern.

If Nd ≠ Ni in step 31, only a part of all the characters of the image marking captured by the TV camera A has been captured, or dirt or the like has been read as characters. The characters separated by the camera A are read (step 34), the characters are separated from the data captured by the TV cameras B and C (step 35), and the separated characters are read (step 35).
36).

After separating the image data input from the TV cameras A, B and C into characters, a reading process is performed. TV
Since the cameras A, B, and C image the markings in an overlapping manner, there may be a case where two or three characters are read at the same position. In this case, the result of reading a character having the highest overlap with the reference pattern is adopted. If there is only one read result, use that one (step 3
7,38). The characters selected in this way are combined with a character string in accordance with the imaged arrangement (step 39), and the result is displayed or printed out (step 39).
40).

FIG. 5 shows alphabets and numbers used in the present invention. Characters that are apt to be confused with other characters when the shape is lost have a unique shape. For example, B and D extend the upper and lower ends long, I greatly extends the upper and lower horizontal parts, O indicates a bar above,
Q has a bar below, S is the inverse of Z, and V is an inverted triangle. In addition, 0 indicates a circle with an oblique line, and 1 indicates that the bar at the lower end is elongated, and is inverted T. With such a shape, erroneous recognition of characters can be reduced.

In the above description, an example of reading characters on the convex surface of the outer surface of a cylinder called a coil surface is shown. However, the present invention is also applicable to the case of reading characters on a concave surface of a cylindrical inner surface such as the inner surface of a pipe by the same method. it can.

[0019]

As is apparent from the above description, according to the present invention, the object is imaged so as to be overlapped by three TV cameras, and the marking is decoded by the most recognized characters. Can be improved.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an embodiment of the present invention.

FIG. 2 is a sectional view taken along line XX of FIG.

FIG. 3 is a block diagram illustrating a configuration of a recognition device.

FIG. 4 is an operation flowchart of the present embodiment.

FIG. 5 is a diagram showing a character pattern that is effective when used in practicing the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Hot coil 2 TV camera 3 Lighting device 4,5 Sensor 6 Recognition device 7 Belt conveyor

Claims (2)

(57) [Claims] 1. A TV camera A installed at a position substantially at the center of a cylinder opposite to a cylindrical body, and a TV camera B and a TV camera arranged on the left and right sides of the TV camera A so as to capture an image of an object to be read in an overlapping manner. C and a sentence in which the characters of the image data input from the TV camera A are separated and the number of characters is marked.
If the same number as the number of characters recognized this separation character, when different from the number of characters that characters has said marking separates the TV camera B, the TV camera C characters on the image data input from, each TV A character recognizing unit for recognizing separated characters input from cameras A, B, and C, and recognizing input data by recognizing characters at the same position by adopting the best recognized character; A character reading device for a cylindrical surface. 2. The character reading device according to claim 2, wherein an illuminating unit is provided on the left and right sides of the TV camera A so as to irradiate a light from a light source to a reflector and irradiate the surface of the cylindrical body.