JPH0122666B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field) The present invention relates to a figure recognition device, and particularly to detection of center point coordinates of an isolated figure. The present invention can be applied to detecting the coordinates of the center point of a seal imprint in automatic seal stamp verification.

(Background Art) Conventionally, the coordinates of the center point of a figure are detected by raster scanning a binarized figure pattern with the X-axis direction as the main scanning direction and the Y-axis direction as the sub-scanning direction, as shown in FIG. 1A. Among the black points (logical "1" points) of the binarized figure pattern existing on each main scanning line, the X coordinate value of the black point with the minimum X coordinate value X _S1 ~ X _So ,
After detecting the maximum X coordinate value X _E1 ~ X _Eo of the black point and completing the sub-scanning of the entire figure pattern,
Let the minimum value of X coordinate value X _S1 ~ _{X So} be X _S , and
Let the maximum value of X _E1 to X _Eo be X _E. Then Figure 1B
As shown in Figure 2, by performing raster scanning with the Y-axis direction as the main scanning direction and the X-axis direction as the sub-scanning direction, the minimum Y coordinate value Y _S and the maximum Y coordinate value Y _E are determined by the same operation as above. obtain.

Then, X ₀ = (X _S + X _E )/2, Y ₀ = (Y _S +
By calculating Y _E )/2, the center point coordinate C
(X ₀ , Y ₀ ) was obtained. This method of detecting center point coordinates can detect accurate center point coordinates with simple processing if there are no blurs or noise points outside the contour line in the figure pattern, but 1
As shown in Figure D, if there is a blur or a noise point outside the contour line (so-called dirt) in the figure pattern, the disadvantage is that the influence is clearly reflected in the detection error of the center point coordinates. It is abbreviated as the frame-centered method].

Another conventional example is a method in which the center of gravity of all black points of a graphic pattern is set to the coordinates of the center point. That is, as shown in Figure 2A, the X coordinate is i and the Y coordinate is j.
(where i and j are integers) is defined as P _ij , and when P _ij is a black point, |P _ij |=1, and when P _ij is a white point (logical "0" point), |P _ij | = 0, the center point G (X ₀ , Y ₀ ) is It can be calculated by Note that m and n are the outermost shell coordinates of the region that allows the existence of graphic patterns on the X and Y axes, respectively. This center point coordinate detection method has higher detection accuracy than the frame center method because even if there is a noise point outside the contour line in the figure pattern, the influence of it is greatly weakened, as shown in Fig. 1D. As shown in Figure 2B, if the complexity of the internal characters, such as the number of strokes, is not equal, such as in the case of a seal impression pattern, the detected center point coordinates C (X ₀ , Y ₀ ) will move toward the more complex character than the true center point. The detection accuracy will be degraded due to displacement. For example, even in the case of a template-drawn figure, if the thickness of the line is uneven as shown in Figure 2 C, the detected center point coordinates C (X ₀ , Y ₀ ) will be closer to the thick line than the true center point. The disadvantage was that the detection accuracy decreased due to displacement (hereinafter, this method will be abbreviated as the centroid method).

By the way, to accurately detect the coordinates of the center point of an isolated figure pattern, for example, after performing position matching between the registered seal imprint pattern and the to-be-verified seal imprint pattern,
This is an essential item in many figure recognition devices, such as a figure recognition device for seal stamp matching that checks whether two seal imprint patterns match.

(Problem to be solved by the invention) The purpose of the present invention is to eliminate these drawbacks.
The center of gravity method is used to reduce the negative effects of external noise, but in order to avoid being affected by internal patterns such as internal characters, the center of gravity of a graphic pattern that has been filled with weights is set at the center point. Its features include a reading section that reads the figure pattern using electrical signals, a fill pattern creation section that fills in the inside of the figure read by the reading section, and a weighted point creation section that and a center point coordinate detection unit that determines the barycenter coordinates as the center point coordinates of the graphic pattern, and the filling pattern creation unit has: (Condition 1); the Y coordinate is Y _j equal to the point of interest, and the X coordinate is There is a black point of a graphic pattern with a value less than or equal to _i , and the Y coordinate is equal to the point of interest Y _j
, and there is a black point of the graphic pattern whose X coordinate is greater than or equal to _{X i} ; (Condition 2);
There is a black point of a graphic pattern whose Y coordinate has a value of Y _j or less, and a black point of a graphic pattern whose X coordinate is X _i, which is equal to the point of interest, and whose Y coordinate is a value of Y _j or more; According to the following two conditions, each point of interest (X _i , Y _j ) in the area where the existence of a graphic pattern is allowed is determined according to the surrounding graphic pattern situation; (1) Condition 1 is satisfied and Condition 2 is satisfied; (2) Points of interest that satisfy either Condition 1 or Condition 2; (3) Points of interest that do not satisfy Condition 1 and Condition 2; This figure recognition apparatus is such that weighting in a point coordinate detection section is performed for each type of classified point of interest.

(Structure and operation of the invention) FIG. 3 shows a first embodiment of the invention, in which a graphic pattern 20 read and binarized by the reading unit 10 is shown.
After filling the inside of the outer frame in the filling pattern creation section 30, the center point coordinate detection section 4
The center point coordinates are detected by calculating the center of gravity of the filled pattern with 0.

Using Fig. 4, fill pattern creation section 3
Let's explain about 0.

As shown in FIG. 4A, the point whose X coordinate is X _i and Y coordinate is Y _j is defined as P (X _i , Y _j ), and the point P (X _i , Y _j )
The value of is defined as V=|P(X _i , Y _j )|. For convenience of explanation, Figure 4 shows graphic patterns that should originally be filled with black dots only as outlines, and also shows main scanning intervals and sub-scanning intervals that are much rougher than they actually are. There is.

The binarized figure pattern 20 in which the black point value is 1 and the white point value is 0 is raster scanned with the X-axis direction as the main scanning direction and the Y-axis direction as the sub-scanning direction, as shown in FIG. 4B. For each main scan, detect the X coordinate value X _S of the black point with the smallest X coordinate among the black points existing on the main scanning line, and the X coordinate value X _E of the black point with the largest X coordinate, and For example, 2 is added to the values of all points between X _S and X _E on the line. For example, in the case of main scanning line l, |P(X _S , Y _l ) |+2→|P(X _S ,
Since Y _l )|~P(X _E , Y _l )|+2→|P(X _E , Y _l )| is performed, the values at each point are shown in a diagram as shown in FIG. 4B. Such processing is performed over the entire area in the sub-scanning direction, that is, from l=1 to n, and the filling in the X-axis direction is completed.

Next, the graphic pattern that has been filled in the X-axis direction is raster scanned with the Y-axis direction as the main scanning direction and the X-axis direction as the sub-scanning direction, as shown in FIG. , a point on the main scanning line whose value is 3 (the value of a black point in a binarized figure pattern is always rewritten to 3, and even though it was a white point in the binarized figure pattern, A point that becomes a new black point by filling in the X-axis direction always has a value of 2), the Y coordinate value Y _S of the point with the minimum Y coordinate, and the Y coordinate value of the point with the maximum Y coordinate. Detect Y _E and Y _S ~ on the main scanning line
For example, add 4 to the values of all points between Y _{and E.} For example, in the case of main scanning line k, |P(X _k , Y _S )|
+4 → | P (X _k , Y _S ) | ~ | P (X _k , Y _E ) | +4 →
Since |P(X _k , Y _E )| is performed, the values at each point are shown in a diagram as shown in Figure 4C. This kind of processing
Filling is performed over the entire area in the sub-scanning direction, that is, from k=1 to m, to complete filling in both the X and Y axis directions.

An example of the fill pattern obtained in this way is illustrated in FIG. 4D. In addition, points drawn like "+" in Figure 4D (for example, Q and R)
is a point filled in both the X and Y axis directions,
Its value is 7 or 6, and the black dot in the binarized figure pattern is 7 (for example, Q), and 2
In the digitized figure pattern, the white dot is 6 (for example, R). Points drawn like "-" (there are no corresponding points in FIG. 4D) are points that are filled only in the X-axis direction, and their value is 2. A point drawn as "1" (for example, T) is a point filled only in the Y-axis direction, and its value is 4.

The center point coordinate detection unit 40 calculates the center of gravity of the fill pattern obtained in this way, and uses it as the center point coordinate. The center of gravity is calculated by simply calculating the center of gravity by uniformly assuming that the value of all black points (points whose value after filling is 1 or more) in the filled pattern is "1", as in the conventional center of gravity method. However, in order to reduce the effect as much as possible even if there are blurs or noise points outside the contour line in the figure pattern,
It has been experimentally confirmed that it is preferable to assign the following weights. In other words, point P (X _i , Y _j ) is
Points filled in both Y-axis directions, that is, |P
The point where (X _i , Y _j )|=7 or 6 is |P(X _i ,
Y _j ) |=2, and the point P (X _i , Y _j ) is filled only in the X-axis direction or in the Y-axis direction, that is, the point where |P (X _i , Y _j )|=4 or 2 Ha｜
P(X _i , Y _j ) |=1, and point P(X _i , Y _j ) is X・Y
Points that are not filled in both axis directions, that is, points where |P (X _i , Y _j ) | = 0, remain as |P
After weighting (X _i , Y _j ) | = 0, the center of gravity of the fill pattern, that is, the center point C _p (X ₀ , Y ₀ ) This is a method of calculating by

As explained above, in the first embodiment, the center point coordinates are detected based on the idea of calculating the center of gravity.
As shown in Figure 1C or Figure 1D, even if there is a blur or a noise point outside the contour line in the figure pattern, the effect will not suddenly be reflected in the center point detection error as in the frame center method (frame center method). In the case of the center method, if the error in the coordinate values is Δd, the detection error of the center point is generally Δd/2), and only a portion of the black spots among the many black spots are reflected. Furthermore, since the number of black points to be used for calculating the center of gravity is increased by filling in compared to the conventional center of gravity method, the degree of its influence is generally reduced.
Not only is this method superior to the frame center method, but it can also detect the center point with higher precision than the center of gravity method.

In addition, Fig. 2B, which was a drawback of the conventional center of gravity method,
Alternatively, even in the case of a figure pattern such as the one shown in Figure 2 C, where the complexity of the internal characters of the stamp pattern is uneven, or the line thickness of the template drawn figure is uneven, it is possible to Since the influence of the pattern situation on center point detection is extremely reduced, highly accurate center point detection can be performed with almost no influence.

The first embodiment has been described using a binary graphic pattern as the graphic pattern 20, but it is also possible to use a multivalued pattern as the graphic pattern 20 and to binarize it immediately before filling. It is also possible to perform filling and central point coordinate detection using the multivalued pattern as it is. In other words, when the multi-value pattern has N values, for example, the constant value added as a point value when filling in the X-axis direction was set to 2 in the first embodiment, but if it is set to a value of N+1 or more, the original multi-value pattern is Since it is possible to fill in while the value information is saved, we created a fill pattern in both the X and Y axes directions while taking these considerations into account, and calculated the weighted center of gravity using the multi-value information. The center of gravity can also be set as the center point coordinates.

(Effects of the Invention) The present invention detects the center point of the graphic pattern by determining the center of gravity after filling the graphic pattern, thereby eliminating detection errors caused by fading or staining of the seal imprint pattern or the graphic pattern, and unevenness of the internal pattern. Since the amount is extremely reduced, it is possible to detect the center point coordinates with high precision, and it can be used in figure recognition devices, particularly figure recognition devices for seal stamp comparison that require high precision in position alignment during comparison.

[Brief explanation of drawings]

1A to 1D are explanatory diagrams of the frame center method, which is a conventional figure recognition device; FIGS. 2A to C are illustrations of the center of gravity method, which is a conventional figure recognition device; and FIG. The example block diagram, FIGS. 4A to 4D, are explanatory diagrams of a fill pattern creating section in one embodiment of the present invention. 10...Reading section, 20...Graphic pattern, 30
. . . Filling pattern creation section, 40 . . . Center point coordinate detection section.

Claims (1)

[Scope of Claims] 1. A reading unit that reads a graphic pattern using an electrical signal, a filling pattern creation unit that fills in the inside of the graphic read by the reading unit, and a filling pattern creation unit that calculates the barycentric coordinates of each weighted point of the filled pattern. and a center point coordinate detection unit that determines the center point coordinates of the figure pattern, and the filling pattern creation unit has: (Condition 1); the Y coordinate is Yj, which is equal to the point of interest, and the X coordinate has a value less than or equal to Xi. A black point of the graphic pattern exists and the Y coordinate is equal to the point of interest
Yj, and there exists a graphic pattern black point whose X coordinate is greater than or equal to Xi; (Condition 2); The X coordinate is Xi, which is equal to the point of interest, and the Y coordinate is less than or equal to Yj. exists and the X coordinate is equal to the point of interest
There is a black point of a graphic pattern that is (1) A point of interest that satisfies condition 1 and condition 2; (2) A point of interest that satisfies either condition 1 or condition 2; (3) A point of interest that satisfies condition 1. A figure recognition device characterized in that points of interest are classified into three types: points of interest that do not satisfy condition 2;