JP2783739B2 - Drop-out color adjustment method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an OCR apparatus for reading a character printed on a form, in which the color of a frame printed on the form is erased (dropped out) to read only the printed character. It relates to an out-color adjustment method.

[0002]

2. Description of the Related Art In an OCR apparatus for reading characters printed on a form on which a frame is printed, it is necessary to erase the color of the frame printed on the form at the time of reading and to read only the characters printed on the form. Conventionally, in the above-mentioned OCR apparatus, only a form in which a frame is printed in a recommended drop-out color is read.

That is, a frame of a form is printed in a recommended drop-out color that is relatively difficult to read, and a slice value for binarizing a read signal is set to a value that does not read the frame color. When reading a form,
The color of the frame was erased and only the characters printed on the form were read. However, even if the form is printed in the recommended drop-out color, the result of the actual reading may be that the frame is read by the ink etc. between the printing manufacturers, or even if the frame is not read, There has been a problem that a large number of notches are generated (the frame portion is partially read, and noise is included in the read result), and the recognition rate is reduced.

[0004]

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above-mentioned conventional problems, and has been made in advance to judge the quality of a drop-out color printed on a form, or By making it possible to set the optimum slice value according to the drop-out color, it is possible to drop out the frame color of the form printed in various colors, and to improve the character recognition rate・ To provide an out-color adjustment method.

[0005]

FIG. 1 is a block diagram showing the principle of the present invention. In FIG. 1, reference numeral 1 denotes a form, reference numeral 2 denotes a filter, reference numeral 3 denotes an optical sensor, and reference numeral 4 denotes an output of an optical sensor 3 which is compared with a slice value. Binarizing means 5 for binarizing the image data, 5 is an identification means for identifying a character based on the output of the binarizing means 4, and 6 is a control means for switching the filter 2, adjusting a slice value, and the like.

According to the first aspect of the present invention, as shown in FIG. 1, a character of a form 1 on which a frame is printed is read by an optical sensor 3 and the read data is read by a predetermined slice value. Binarizing means 4 for binarizing by comparing with
Identification means 5 for identifying a character based on binarized data
In the drop-out color adjustment method in the OCR device having the following, the slice value is adjusted to obtain a first slice value at which the frame color cannot be read,
Control means 6 is provided for obtaining a second slice value at which characters cannot be read by adjusting the slice value, and obtaining an intermediate value between the first and second slice values from the first and second slice values. The data read by the optical sensor 3 is binarized by using the obtained intermediate value slice value.

According to a second aspect of the present invention, in the first aspect, the optical sensor 3 for reading the form 1 is provided with a filter 2.
The control unit 6 switches the color of the filter 2, selects the optimal filter color, and reads the form 1. According to a third aspect of the present invention, in the first or second aspect, when a difference between the first slice value and the second slice value is smaller than a predetermined value,
A control means 6 for determining that the drop-out color printed on the form is unusable is provided.

[0008]

The data of the form 1 read by the optical sensor 3 via the filter 2 is binarized by the binarizing means 4 by comparing it with the slice value. The identification unit 5 extracts a characteristic pattern from the data binarized by the binarization unit 4 and compares the extracted characteristic pattern with the characteristic pattern stored in the dictionary, thereby identifying the character printed on the form.

The control means 6 switches the color of the filter 2 to select the optimum filter color and adjusts the slice value of the binarization means 4 so that the frame color cannot be read and the first slice value and the character A second slice value at which data cannot be read, an intermediate value between the first and second slice values is obtained from the first slice value and the second slice value, and the obtained intermediate value slice value is converted into a binarizing means. Set as the slice value of

When the difference between the first slice value and the second slice value is within a predetermined value, the control means 6 determines that the drop-out color printed on the form is unusable. According to the first aspect of the present invention, as described above, the slice value of the binarizing means 4 is adjusted so that the first slice value at which the frame color cannot be read and the second slice value at which the character cannot be read. Then, the slice value of the binarizing means 4 is set based on the first slice value and the second slice value, so that the optimum slice value can be set according to the drop-out color, and the character can be set. The recognition rate can be improved.

According to a second aspect of the present invention, in the first aspect, the optical sensor 3 for reading the form 1 is provided with the filter 2, and the control means 6 switches the color of the filter 2 to determine the optimum filter. Since the form 1 is read by selecting a color, the frame color of the form printed in various colors can be dropped out. According to a third aspect of the present invention, in the first or second aspect of the present invention, when the difference between the first slice value and the second slice value is smaller than a predetermined value, the drop size printed on the form may be reduced. Since the out-color is determined to be unusable, it is possible to determine the quality of the drop-out color printed on the form.

[0012]

FIG. 2 is a diagram showing the overall configuration of a system according to an embodiment of the present invention. In FIG. 2, reference numeral 11 denotes a filter for deleting a drop-out color of a form, and 1
Reference numeral 2 denotes an optical sensor such as a CCD for reading a form, 13 denotes a binarizing means for binarizing the output of the optical sensor 12, 14 denotes a scanner control unit for controlling a scanner, 15 denotes a memory, and 16 denotes a filter. A filter driving means 17 for switching is an identification means for identifying the read image data.

In the identification means 17, 17a is 2
A pre-processing unit for performing a pre-processing such as a filtering process or an emphasizing process on the output of the value conversion unit 13, 17b is a feature extracting unit for extracting a feature from the read data, and 17c is a feature extracting unit 17b. An identification unit 17d for identifying a character from the characteristic pattern of the read data, and 17d is a dictionary storing characters corresponding to the characteristic pattern.

FIG. 3 is a diagram showing an example of the configuration of the filter 11 shown in FIG. 2. FIG. 3A shows an example in which the filter 11 is formed in a rectangular shape and is switched between a rack and a pinion. () Shows an example in which the filter 11 is formed in a disk shape and the filter is switched by rotating the disk. In FIG. 3, reference numerals 111 and 112 denote filters, 111
a and 112a are filters for blue, for example, 111b and 1
12b is a green filter, for example, 111c, 112c
Represents a red filter, for example, 111d and 112d represent infrared filters, 111e represents a pinion, and 111f represents a rack. The filters shown in FIGS. 2A and 2B are filter driving means shown in FIG. Rotated by 16,
The filter is switched.

FIG. 4 is a diagram showing the arrangement of a form, a filter, and an optical sensor. In FIG. 4, reference numeral 21 denotes a form on which a frame color is printed; 22, a halogen or white light source; 23, a mirror; Is an optical sensor. As shown in the figure, the halogen or the light of the white light source 22 applied to the form 21 is reflected by the mirror 23 and passes through the filter 11 which is switched by the scanner control unit 17 (see FIG. 2). Input to the optical sensor.

Next, the operation of the embodiment of the present invention will be described with reference to FIG. In FIG. 2, the image data of the form read by the optical sensor 12 via the filter 11 is 2
In the value converting means 13, the value is compared with the slice value and converted into a binary signal of 1 or 0. The converted binary signal is provided to a pre-processing unit 17a of the identification unit 17 to perform a filtering process or a fattening process (a process such as thickening a read black portion when a part of a character is blurred). , A preprocessing such as recognition as a continuous line) is performed, and a characteristic pattern is extracted from the read image data in the characteristic extracting unit 17b.

The identification unit 17c identifies a character by comparing the characteristic pattern extracted by the characteristic extraction unit 17b with the characteristic pattern stored in the dictionary 17d, and outputs an identification result. In FIG. 2, in order to drop out the color of the frame printed on the form, a test form in which a frame is printed in the color to be dropped and characters are printed is read in advance.

The scanner control unit 14 switches the colors of the filter 11, reads the entire area of the test form for each color of the filter, and selects an optimal filter. And if there is an optimal filter, it is used. Also,
If there is no optimal filter, the filter with the lowest output of the frame color is selected, and the slice value of the binarizing means 13 is varied. If there is a slice value that allows the frame color to be dropped out and the character to be read with a predetermined margin, the slice value is set to that value. If the above slice value does not exist, it indicates that the form printed with the frame color cannot be used.

FIG. 5 is a diagram showing the operation of adjusting the slice value of the binarizing means 13 in the embodiment of the present invention. In FIG. 5, the output of the optical sensor 12 when "white" is read is the slice value. The initial value is the optimal slice value,
Is the output of the optical sensor 12 when “black” is read,
A and B are the outputs of the optical sensor 12 when the frame color of the form is read, and C is the optical sensor 12 when the character area is read.
3 shows an example of the output.

In FIG. 5, the scanner control unit 14 reads a test form, changes slice values one by one, monitors the color of each frame, and determines that the color of the frame is “white”. The slice value is switched to the position, and the value at that time is stored (the value in FIG. 5). Next, the output at a certain position of the character is monitored, and similarly to the case of the frame color, the slice value is sequentially changed to store the position where the character is no longer output (FIG. 5).

Next, the slice value (FIG. 5) having no frame color is compared with the slice value (FIG. 5) having no character output, and the following processing is performed. (A) Difference between position of frame color and slice value of character (Δ in FIG. 5)
X) If the value is smaller than a certain value (there is almost no difference), it is determined that the drop-out color cannot be used. (B) Difference between position of frame color and position of slice value of character (Δ in FIG. 5)
X) If the value falls within a certain value range, an intermediate value of each slice value (FIG. 5) is used as a slice value, and when reading a character, the character portion is emphasized by filtering or fattening processing. Perform processing to improve recognition accuracy. (C) Difference between position of frame color and slice value of character (Δ in FIG. 5)
X) If the value is equal to or more than a certain value, the intermediate point is set as the optimum slice value.

Note that, for a plurality of forms on which frames are printed with different drop-out colors, the filter and slice value setting processing is performed in advance, and the filter color and slice value for each form are changed. If the data is stored in the memory 15, the color of the filter and the value of the slice value can also be immediately set by simply designating the type of the form.

FIGS. 6 and 7 are flowcharts showing the processing procedure of the slice value adjustment processing in the scanner control unit 14. In step S1 of FIG. 6, the positions of the character area and the frame area of the test form from the console panel etc. Is set, and an initial slice value is set in step S2.
Next, in step S3, reading of a test form printed in the color of the frame to be dropped out is started. In step S4, it is determined whether or not the frame cannot be read. If the frame color has been read, 1 is added to the slice value in step S5, and the process returns to step S4.

When the above processing is repeated and the slice value at which the frame is not read is determined at step S4, the process goes to step S6, and the slice value at which the frame becomes unreadable is stored as "A". Then, the process goes to step S7 in FIG. 7 to determine whether or not characters can be read.
If it is determined in step 7 that the characters cannot be read (both the frame and the characters cannot be read), the process proceeds to step S15, and the drop-out color of the form is determined to be inappropriate.

If the character can be read, the process goes to step S8, where 1 is added to the slice value, and in step S9, it is determined whether the character can be read. If the character can be read, the process returns to step S8, adds 1 to the slice value, and returns to step S9. When the above processing is repeated and the slice value at which the character cannot be read is determined, the process proceeds to step S10, and the slice value at which the character cannot be read is stored as the slice value "B".

In step S11, the slice value "A" and the slice value "B" are added and divided by 2 to calculate an optimum slice value "X". In step S12,
It is determined whether [X−A] is not smaller than the allowable amount Y,
If [XA] is not smaller than the allowable amount Y, it is determined that the slice value is normal and the slice value of the binarizing means 13 is set to the value "X", and the process ends normally.

If it is determined in step S12 that [XA] is smaller than the permissible amount Y, the process proceeds to step S13 as a margin shortage, and the preprocessing unit 17a adds the above-described emphasis to the read character portion. In step S14, it is determined whether the frame color has been dropped out and the character portion can be read with a margin. If the character portion can be read, the slice value of the binarizing means 13 is set to the value " X ”and ends normally.

If it is determined in step S14 that the document cannot be read, the process proceeds to step S15, where it is determined that the drop-out color of the form is inappropriate, and the process ends with an error. The scanner control unit 14 displays this to the outside. FIG. 8 is a flowchart showing a processing procedure of the filtering process in the scanner control unit 14. The filtering process will be described with reference to FIG.

At step S21, it is determined whether or not there is an optimum filter by switching the filter 11, and if there is an optimum filter, the process proceeds to step S2.
Go to 6 and perform normal reading. If it is determined in step S21 that there is no optimum filter, the process proceeds to step S22, where the output is adjusted to the filter closest to "white", and in step S23, the slice value is changed to change the color of the frame to " Find a slice value that can be judged as "white".

In step S24, it is determined whether the frame color is dropped out and there is a slice value from which the character can be read. If the slice value does not exist and the character cannot be read, the process proceeds to step S25. And the form printed in that frame color is displayed externally as unusable. Also,
If it is determined in step S24 that the characters can be read, the process proceeds to step S26, where normal reading is performed.

[0031]

As is apparent from the above description,
In the present invention, the color of the filter is switched to select the optimum filter color, and the slice value of the binarizing means is adjusted so that the first slice value at which the frame color cannot be read and the second slice value at which the character cannot be read. Find the slice value of
Based on the first slice value and the second slice value, 2
Since the slice value of the value conversion means is set, it is possible to determine in advance whether the drop-out color printed on the form is good or not, or to set the optimum slice value according to the drop-out color. It is possible to drop out a frame color of a form printed in various colors. In addition, this can improve the character recognition rate.

[Brief description of the drawings]

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

FIG. 2 is a diagram illustrating an overall configuration of a system according to an embodiment of the present invention.

FIG. 3 is a diagram illustrating an example of a configuration of a filter.

FIG. 4 is a diagram showing an arrangement of a form, a filter, and an optical sensor.

FIG. 5 is a diagram illustrating an operation of adjusting a slice value according to the embodiment of the present invention.

FIG. 6 is a flowchart illustrating a processing procedure of slice value adjustment processing.

FIG. 7 is a flowchart (continued) showing a processing procedure of slice value adjustment processing.

FIG. 8 is a flowchart illustrating a processing procedure of a filtering process.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 form 2, 11 filter 3, 12 optical sensor 4, 13 binarization unit 5, 17 identification unit 6 control unit 14 scanner control unit 15 memory 16 filter driving unit 17 a preprocessing unit 17 b feature extraction unit 17 c identification unit 17 d dictionary

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) G06K 9/20 G06K 9/38

Claims (3)

(57) [Claims] 1. A binarizing means (4) for reading a character of a form (1) on which a frame is printed by an optical sensor (3), and comparing the read data with a predetermined slice value to binarize the data. In a drop-out color adjustment method in an OCR apparatus having an identification means (5) for identifying a character based on binarized data, the first slice value is adjusted so that the frame color cannot be read.
And a second slice value at which a character cannot be read by adjusting the slice value. An intermediate value between the first and second slice values from the first slice value and the second slice value. A drop-out color adjustment method characterized by providing control means (6) for obtaining the data, and binarizing the data read by the optical sensor (3) using the obtained intermediate slice value. 2. A filter (2) is provided in an optical sensor (3) for reading a form (1), and a color of the filter (2) is switched by a control means (6) to select an optimal filter color. Report
2. The method according to claim 1, wherein (1) is read.
Out color adjustment method. 3. A control means for judging that a drop-out color printed on a form is unusable when a difference between the first slice value and the second slice value is smaller than a predetermined value. 3. A drop-out color adjustment method according to claim 1, wherein the color adjustment method is provided.