JPS6128157B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention, in the field of image processing, relates to a matching device that reads a pre-registered pattern and a target pattern and determines whether the target pattern is the same as the registered pattern. A conventional device of this type is shown in FIG. In the figure, 1 is an image input device, 2 is a binarization circuit that binarizes the output of the image input device 1 and takes it out as a digital quantity, and 3 is a two-dimensional addressable image that stores the output of the binarization circuit 2. memory,
4 is a registered pattern memory that stores information on registered patterns; 5 is a calculation memory that stores intermediate results of calculations; 6 is a projection histogram calculation unit that calculates the horizontal and vertical projection histograms of the image; 7 is registration. A cross-correlation calculating section 8 calculates the cross-correlation coefficient of the projection histograms of the pattern and the target pattern, and 8 is a matching section that determines whether the registered pattern and the target pattern match based on the value of the cross-correlation coefficient. Next, the operation will be explained. First, an image of a subject is input to the image input device 1. The image input device is
For example, like a television camera, it is a device that consists of an optical system, a scanning mechanism, and an optical conversion mechanism, and performs a raster scan of a spot on the subject, and converts the gradation information on the spot into an electrical signal using a photoelectric conversion mechanism, and outputs the signal. . Next, in synchronization with the scanning mechanism of the image input device 1, a binarization circuit compares a preset threshold voltage with the output of the image input device 1 using a sample pulse for digitally reading the subject. Pixels fixed to the object, that is, pixels on the object, are output as "1" and backgrounds are output as "0", and are sequentially written into the image memory 3. Reference numeral 4 denotes a registered pattern memory in which a plurality of patterns are registered in advance and has a function of specifying a specific pattern using, for example, a keyboard. The following explanation will be given assuming that the size of the digital image to be handled is M×M. The projection histogram calculation section 6 reads out the target pattern and the registered pattern from the image memory 3 and the registered pattern memory 4, obtains horizontal and vertical projection histograms, and stores them in the calculation memory 5. For example, a horizontal projection histogram is a count of the number of pixels fixed on an object for each horizontal line starting from the topmost line of a digital image. Similarly, the projection histogram in the vertical direction is obtained by arranging the counts for each line in the vertical direction starting from the leftmost end. The matching unit 8 calculates the similarity between the target pattern and the registered pattern from the horizontal and vertical projection histograms of the target pattern and the registered pattern, and determines whether the target pattern is the same as the registered pattern based on the similarity of both. Make a judgment. The cross-correlation calculation unit 7 obtains a horizontal or vertical histogram (i), j=1 to M; target pattern g(j), j=1 to M; registered pattern (g(j)=0) from the calculation memory 5. j<1orj>M) and the cross-correlation coefficient Λ(τ) seek. The matching unit 8 calculates the similarity S from the output of the cross-correlation calculation unit 7.

Calculate using [Formula]. When the two patterns are compared, it is determined that they match when the conditions SH and SVS _O are satisfied, for example, based on predetermined threshold values S _O of horizontal and vertical similarities SH and SV. Note that the cross-correlation coefficient Λ(τ) is −1Λ
(τ) has a value of 1, and the equality sign of Λ(τ) = 1 is
(i)=c・g(i−τ) The equality sign of Λ(τ)=−1 is when (i)=−c・g(i−τ)c0 holds true. Therefore, the more similar both patterns are, the closer the value will be to 1. Conventional verification devices are configured as described above, so the patterns used need to be low-noise patterns.For example, when printing a seal, it is necessary to avoid minute differences in fonts such as blurring and blurring. For objects with poor image quality where fonts are distributed over the entire area, it is difficult to distinguish between marks with truly different fonts, and the marks are often judged to be different, thus increasing the overall reject rate. There were flaws. This invention was made in order to eliminate the drawbacks of the conventional ones as described above. Projection histograms of a target pattern and a registered pattern are obtained, one of the projection histograms is first divided into several large regions, and then the projection histograms are divided into several large regions. Divide a large area into several small areas. Then, each small area is matched with the other projection histogram, and when the number of matched small areas is distributed such that there is a certain number or more in all large areas, the marks are determined to match. It is an object of the present invention to provide a seal imprint collation device that can accurately collate even a mark of poor image quality in which minute differences in font are distributed throughout the entire body. FIG. 2 is a schematic diagram showing an embodiment of this invention.
1 is an image input device, 2 is a binarization circuit that binarizes the output of the image input device 1 and takes it out as a digital quantity, and 3 is a two-dimensional addressable image memory 4 that stores the output of the binarization circuit 2. A registered pattern memory stores registered pattern information; 5 is an arithmetic memory that stores intermediate results of calculations; 6 is a projection histogram calculation unit that calculates a horizontal and vertical projection histogram of an image; 7 is a registered pattern and a cross-correlation calculation unit that calculates the cross-correlation coefficient of the projection histogram of the target pattern; 9 a small area matching unit that determines whether the small areas match;
Reference numeral 10 denotes a distribution measuring unit that examines the distribution of matched small areas in the large area, and reference numeral 8 denotes a matching unit that determines whether marks match or do not match based on the output of the distribution measuring unit. In FIG. 2, the process up to obtaining a projection histogram is the same as in the conventional method, in which a subject is inputted by an image input device 1, converted into a binary digital image by a binarization circuit 2, and stored in an image memory 3. As the registered pattern, one of the patterns stored in the registered pattern memory 4 is designated and read out. The projection histogram calculation unit 6 stores horizontal and vertical projection histograms of the registered pattern and the target pattern in the calculation memory 5. The cross-correlation calculating unit 7 calculates a projection histogram (i), i=1 to M; target pattern g(j), j=1 to M; registered pattern (however, g(j)=0 j1 or jM) from the calculation memory 5. ) and calculate the cross-correlation coefficient Λ _R (τ) in the specified small region R (k, l). seek. The small area matching unit 9 matches the specified small area R.
At (k, l), it is determined whether the patterns match. This includes, for example, a given deviation τ
Near _O , change the deviation τ appropriately (|τ−τ _O |
ε; ε a constant determined as appropriate) Maximum value of the cross-correlation coefficient Λ _R (τ) of the output of the cross-correlation calculation unit 7 Find this value and set it to a certain predetermined threshold value
It is determined whether the condition X _R X _O is satisfied by comparing with _O , and if the condition is satisfied, it is determined that the small area is matched. The distribution measurement unit 10 first divides the projection histogram of either the target pattern or the registered pattern (here, the target pattern) into several equally spaced large areas.
R ₁ (k ₁ , l ₁ ), R ₂ (k ₂ , l ₂ ), ... Rα (k α, l
α), ...Rm (km, lm). Here kα=M/m(α-1)+1 lα=M/m・α M is a multiple of m Then, this large area Rα (kα, lα) is further divided into several small areas Rα ₁ (kα ₁ , lα ₁ ), Rα
₂ (kα ₂ , lα ₂ ),...Rαβ (kαβ, lα
β), ... Rαn (kαn, lαn). Here (M is a multiple of m・n kαβ=kα+M/m・n(β−1) lαβ=kα+M/m・nβ−1, β=1, 2,...n) Then, first, the deviation τ in the whole area is changed, and the cross-correlation calculation unit determines the point τ _O at which the cross-correlation coefficient is maximum. Then, using this deviation τ _O , the small area matching unit 9 calculates all the small areas Rαβ(kαβ, l
αβ), α=1~m, β=1~m, it is determined whether or not the small area matches, and the large area Rα,α=
Count the number of matching small areas every 1 to m. Let Cα be the count value in the large area Rα. Next, the matching section 8 determines whether the marks match or do not match based on the distribution of the number of matching small areas Cα, α=1 to m within the large area. For example, if a predetermined threshold value T is used, and the condition "Cα/n>T, or alld" is satisfied, that is, a certain number or more of matching small regions are included in all large regions, then It is determined that the marks match. According to this invention, when performing pattern matching using projection histograms of binary patterns, one projection histogram is divided into large regions, which are further divided into small regions, and all of the large regions are When the distribution is such that the number of small areas that match a certain number or more is determined to be a match, it is determined that the marks match, so minute differences in the pattern, such as blurring or blurring when stamping, are distributed throughout the mark. Even marks with poor image quality can be verified.

[Brief explanation of the drawing]

FIG. 1 is a schematic configuration diagram showing a conventional seal imprint verification device, and FIG. 2 is a schematic configuration diagram showing a seal imprint verification device according to an embodiment of the present invention. Note that the same reference numerals in the figures indicate the same or corresponding parts. 1: Image input device, 2: Binarization circuit, 3: Image memory, 4: Registered pattern memory, 5: Calculation memory, 6: Projection histogram calculation unit, 7: Cross-correlation calculation unit, 8: Collation unit, 9: Small area matching section,
10: Distribution measurement section.

Claims (1)

[Claims] 1. Equipped with means for inputting seal impressions affixed on documents such as securities as binary digital images, memory means for storing a large number of digital images of seal impressions, and means for selecting registered patterns to be compared from this memory means, In a seal imprint matching device that uses projection histograms that count the number of pixels with the value "1" in the horizontal and vertical directions to determine whether both marks are the same, one projection histogram is Divide into large regions, further divide each into several small regions, match each small region with the projection histogram of the other, and find that a certain number or more of matched small regions exist in all large regions. A seal imprint matching device characterized by determining that the marks match when performing distribution.