JPH0679329B2 - Image sensor read gradation inspection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Outline] A reading gradation inspection method for an image sensor, which can judge the quality by obtaining a difference between an average value of measured data and registered reference data, and performs a reading gradation inspection with high accuracy. An image sensor reading gradation inspection method for reading an image written on a medium by an image sensor, the method being provided with steps. A test pattern in which gradations of a plurality of levels are arranged in a predetermined width in the order of gradation levels, a storage unit for storing gradation-specific reference data for measuring each gradation level of the test pattern, and the test pattern by an image sensor The average value of the data for each gradation level read over a predetermined area in the main scanning direction and the sub-scanning direction is calculated and stored in the average value and storage means. And a calculation unit that calculates the difference between the gradation-based reference data corresponding to each gradation level,
The configuration is such that pass / fail is determined based on the calculation result by the calculation means.

[Industrial application field]

The present invention relates to a pixel reading gradation inspection method for an image sensor of an optical reading device, and in particular, a reading gradation inspection for an image sensor capable of determining pass / fail by obtaining a difference between an average value of measured data and registered reference data. It is about the method.

Recently, a method of reading characters, symbols, codes, etc. written on a medium is widely used, and a bar code in a POS system, a manuscript reading of a facsimile, a code showing a page of a bankbook such as a bank, or an optical character reading. Device (Op
tical character reader: hereinafter referred to as OCR). There is a demand for a method capable of accurately inspecting the reading performance of an image sensor when manufacturing and maintaining such a device.

[Conventional technology]

The OCR 1 shown in FIG. 5 includes, for example, a reading unit 3 including an image sensor 2a composed of a charge-coupled device (CCD), a binarizing circuit 4 for binarizing the read data, and a binarized data. Image buffer 5 to be stored, character recognition unit 6 to recognize characters,
It is composed of a display (hereinafter referred to as CRT) 7a, a printer 7b, and a control unit 8.

Therefore, the image data read by the image sensor 2a from a document or the like on which characters or the like are written is binarized and stored in the image buffer 5. The character recognition unit 6 cuts out data for each character from the stored image data, recognizes the characters, displays them on the CRT 7a, prints them out to the printer 7b as necessary, and sequentially recognizes the characters.

In this way, the image data read by the image sensor 2a is recognized and printed out as a character pattern by the printer 7b, or sent to an external data processing device and processed as numerical data.

Due to the demand for improvement of OCR1 character recognition accuracy, the structure of the image sensor element is becoming more and more miniaturized, the number of bits is increasing, and the read data is multi-valued. There is a growing need to inspect the key accurately.

Not only in the case of OCR of this example, but also in the case of barcode reading, facsimile, etc., the quality of recognition and document reproduction depends on the read gradation of the image sensor. The following method is performed as an inspection of the pixel reading gradation by the image sensor at the time of manufacturing or maintenance.

That is, the image sensor 2a reads a test pattern printed in stages with a predetermined gradation level, and the gradation of the analog value output to the synchroscope or the like is measured. In this case, the output data is multi-valued, and the gradation average value is visually measured.

[Problems to be solved by the invention]

According to the above-mentioned conventional method, as a method of determining the quality of the read gradation by the image sensor, since the average value of multi-valued data is visually measured from the analog output of the synchroscope, not only the number of steps is required, The accuracy is limited, and it is difficult to accurately inspect the read gradation. Moreover, it is necessary to accurately print the density of the gradation level of the test pattern, but in reality the density and variations of the printing tend to be inaccurate, so even if an average value is given, it cannot be said to be a reliable value. There is a problem.

It is an object of the present invention to provide a read gradation inspection method for an image sensor that can perform a read gradation inspection with high accuracy.

[Means for solving problems]

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

In the figure, 2 is an image sensor, 9 is a test pattern in which gradations of a plurality of graded levels are arranged in a predetermined width in the order of gradation levels, and 10 is a gradation in which each gradation level of the test pattern 9 is measured. Storage means for storing different reference data, 13 calculates the average value of the gradation level-specific data obtained by reading the test pattern 9 over a predetermined area in the main scanning direction and the sub scanning direction by the image sensor, and stores the average value and the storage. means
The calculation means calculates the difference between the gradation-based reference data corresponding to each gradation level stored in 10.

Therefore, the pass / fail judgment is made on the basis of the calculation result by the calculation means 13.

[Action]

The image sensor 2 reads the test pattern 9 on which gradations of a plurality of levels are printed over a predetermined area in the main scanning direction and the sub-scanning direction, and the arithmetic means 13 averages the data for each gradation level. Is calculated, and the difference between the average value and the gradation-based reference data corresponding to each gradation level stored in the storage means 10 is calculated, and the pass / fail is determined based on the calculation result.

In this way, by comparing the measured value of each gradation level of the test pattern 9 and the average value of each read gradation of the image sensor 2, the printing accuracy of the test pattern 9 can be relaxed and the read gradation is automatically adjusted. , And can be accurately inspected.

〔Example〕

An embodiment of the present invention will be described below with reference to FIGS. The same reference numerals denote the same objects throughout the drawings. Second
In the figure, those corresponding to FIG. 1 are shown surrounded by a one-dot chain line.

In FIG. 2, the test pattern 9a is a gradation pattern having a plurality of levels and a predetermined length in the horizontal direction.
Each has a predetermined width in the vertical direction, for example 2 mm width (pixel width 0.125 mm x 1
It is printed in 6) in order of the gradation level.

The reference data memory 10a stores, as reference data, a value obtained by measuring each gradation level of the test pattern 9a with a densitometer or the like. The reference data is input from a keyboard (not shown).

The conversion unit 11 converts the gradation of each bit of the image data (for example, 16 bits in the main scanning direction and 16 lines in the sub scanning direction) of the gradation pattern read from the test pattern 9a into, for example, 16 steps, and then converts the data. Send to memory 12. Part of the data is illustrated in FIG. 3 (a), and the numerical values of the gradation of each bit of the image data are arranged. The gradation has 16 gradation levels and is represented by 0 to 9, A to F (A = 10, ..., F = 15).

The operation unit 13a reads the test pattern 9a by the image sensor 2a in the main scanning direction and the sub-scanning direction over a predetermined area by gradation pattern data, 16 × 16 = 256 in FIG.
The average value x of the pixel data is calculated, and the difference between the average value x and the reference data S for each gradation stored in the reference data memory 10a, the difference (deviation) between the highest gradation and the lowest gradation, and the highest rank Calculates the key%.

The determination unit 14 includes a setting unit 15 in which a determination reference value L is set, compares the difference δ calculated by the calculation unit 13a with the determination reference value L, and if δ ≦ L, it is determined to be acceptable, and δ> L. If it is judged to be unacceptable.

The data processing unit 16 includes, for each gradation pattern, the result determined by the determination unit 14, the reference data S for each gradation, and the average value.
The x, the difference δ, the deviation, the maximum gradation and its percentage, etc. are edited and sent to the printer 7b.

Since it has such a configuration and functions, the operation at the time of testing will be described with reference to the flowchart of FIG.

First, the data of each gradation pattern obtained by measuring the test pattern 9a with a densitometer in advance is stored in the reference data memory 10a as gradation-specific reference data S.

By the image sensor 2, one gradation pattern of the test pattern 9a is 16 bits in the main scanning direction and in the sub scanning direction.
16 lines are read and stored in the image data image buffer 5.

The image data output from the image buffer 5 is a conversion unit.
It is converted into 16 gradations in 11 and stored in the data memory 12.

The calculation unit 13a calculates the average value x of the data read from the data memory 12, reads the corresponding gradation-based reference data S from the reference data memory 10a, and calculates the difference δ (= S−
x) is calculated and sent to the determination unit 14.

The determination unit 14 reads the determination reference value L from the setting unit 15,
The difference is compared with δ, and if the absolute value of δ is> L, it is determined as a failure.

The data processing unit 16 displays the judgment result and the reference data for each gradation.
S, the average value x, the difference δ, the maximum gradation, the% of the maximum gradation, etc. are edited and sent to the printer 7b for printing output. FIG. 3B illustrates a part of the print format.

Thus, to are repeated for all the gradation levels of the test pattern 9a. The clerk makes a comprehensive judgment based on all the output data.

In this way, the reading gradation of the image sensor 2 is automatically
And it can be inspected accurately.

Therefore, it is possible to use an existing printed matter without requiring accurate printing of the density of the gradation level of the test pattern.

〔The invention's effect〕

As described above, according to the present invention, the inspection is performed based on the measured value of the gradation level of the test pattern, so that the read gradation is automatically and accurately inspected without requiring the accuracy of the print density. The effect is that you can.

[Brief description of drawings]

 FIG. 1 is a block diagram of the principle of the present invention, FIG. 2 is a block diagram showing an embodiment of the present invention, FIG. 3 is an explanatory diagram of inspection data of the embodiment, FIG. 4 is a flowchart of the embodiment, and FIG. FIG. 3 is a block diagram of an OCR to which the present invention is applied. In the figure, 2 and 2a are image sensors, 9 and 9a are test patterns, 10 is storage means, 10a is reference data memory, 13 is operation means, and 13a is operation section.

Claims (1)

[Claims] 1. A reading gradation inspection method for an image sensor of a reading device for reading an image written on a medium by means of an image sensor (2), wherein gradations of a plurality of levels with steps are respectively arranged within a predetermined width. A test pattern (9) printed in the order of the gradation level; a storage unit (10) for storing gradation-specific reference data for measuring each gradation level of the test pattern (9); and the test pattern (9). The image sensor (2) calculates an average value of the gradation level-specific data read over a predetermined area in the main scanning direction and the sub scanning direction, and the average value and each floor stored in the storage means (10). Computation means (13) for computing the difference between the gradation-based reference data corresponding to the tonal level
And a read gradation inspection method for an image sensor, comprising the step of judging pass / fail based on a calculation result by the calculation means (13).