JPS63145579A - Method for extracting hole - Google Patents

Abstract

PURPOSE:To efficiently extract a hole of a picture with a simple method, in a short processing time by obtaining X-coordinate of black and white picture elements in one line of a raster scan. CONSTITUTION:When X-coordinate of the black picture element at the left end in applying raster scan to a binary image is Bs, the X-coordinate of the black picture element at the right end is Be, the X-coordinate of the white picture element between he coordinates Bs and Be in the transition from the black picture element to the white picture element is Ws, and the X-coordinate of the white picture element in moving from the white picture element to the black picture element is We, when the state is transited from a 1st state where no white picture element exists between the coordinates Bs and Be in one line of the raster scan into a 2nd state where a white picture element exists between the coordinates Bs and Be through the consecutive raster scanning, a flag 1 is set. In a picture shown in figure, the flag 1 is set by a raster 3 and the coordinates Ws, We and the coordinates Bs, Be at one preceding raster 2 are checked. When the conditions of Bs2<Ws3 and We3<Be2 are satisfied, it is considered that there exists a hole.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field) The present invention relates to a method for extracting holes in an image in an OCR (optical character recognition device).

(Prior art) As a conventional method for hole extraction, an image is labeled, and a final check is made to determine whether or not a white pixel area surrounded by a black pixel area is connected to the outside of the black pixel area. There is a way to extract it as a hole when it is confirmed that it is not.

However, this method has the problem that processing is complicated and takes a lot of processing time.

(Object of the Invention) The present invention provides a method for extracting holes in an image through simple processing.

(Structure of the Invention) When a binary image such as a character is raster scanned, the X coordinate of the leftmost black pixel is Bs, the X coordinate of the rightmost black pixel is Be,
The X coordinate of the white pixel when moving from a black pixel to a white pixel between Bs and Be is Ws, the X coordinate of the white pixel when moving from a white pixel to a black pixel is We, and Bs in one line of raster scan. Bs in the raster when changing from the first state where there are no white pixels between Be to the second state where there are white pixels
, Be.

Ws, We, or Bs, Be in the raster when changing from the second state to the first state.

Holes are extracted using Ws and We.

(Example) Hereinafter, embodiments will be described in detail based on the drawings.

1 to 4 each show the operational flow of an embodiment of the present invention, and FIG. 8 shows the operation flow of the embodiment of the present invention.

This figure shows the hardware configuration of the present invention.

Dog class 1 First, FIG. 5(a) shows a binary image.

When this image is raster scanned, the X coordinate of the leftmost black pixel is Bs, the X coordinate of the rightmost black pixel is Be, and Bs and B
Let Ws be the X coordinate of a white pixel when moving from a black pixel to a white pixel between e, and We be the X coordinate of a white pixel when moving from a white pixel to a black pixel, then each raster ■, ■, ■, ・In the scan of . . . , B s HB e +W s , respectively.
A pixel corresponding to W e exists as shown in FIG. 5(b). Therefore, one of the raster scans
When the first state in which there are no white pixels between Bs and Be in the line changes to the second state in which there are white pixels between Bs and Be by further raster scanning, a flag 1 is set. In the image in Figure 5, flag 1 is shown in raster ■.
At this time, Ws, We, and the previous raster ■
Bs in

Check Be. and.

Bs■<Ws■, We■<Be■ When the condition (1) is satisfied, it is considered that there is a hole there. When the raster scan continues and the state changes to a first state in which there are no white pixels between Bs and Be in one line of the raster scan, this is extracted as a hole. A similar check is performed on the raster parts ■ and ■, and holes are extracted.

The operation flow of FIG. 1 shows the procedure of the hole extraction method described above. Here, Wsk, Wek and k mean Ws and Ws in raster k when changing from state 1 to state 2, and -1 in B5k-1°Beh-0 means the change in raster. Bs in previous raster
, represents Be.

Even if condition (1) in Example 1 is satisfied, the image shown in FIG. 6 is not a hole. Therefore, the following conditions are used.

1IBe■-Bs■1-IBe■-Bs■II<TH-
(2) where TH is a predetermined threshold;

In short, by performing a raster scan using the same procedure as in Example 1, we captured changes in the state, and
Compare the distance between s and Be with the distance between Bs and Be in the previous raster ■, and if there is a large difference between the same distances as shown in Figure 6, it must be a hole. Conceivable. Therefore, if the difference is small, that is, if it is smaller than a predetermined threshold TH, it is extracted as a hole.

The second example shows the procedure of the hole extraction method in Example 2.
It is an operation flow of the figure.

Husband 1 Artisan 1 In the method of Examples 1 and 2, in the case of an image like that shown in Fig. 7,
The part above "6" is extracted as a hole. Therefore, in this embodiment, a more accurate hole extraction method will be described.

That is, from a first state in which there are no white pixels between Bs and Be in one line of raster scanning, a raster scan is performed to change to a second state in which there are white pixels, and then raster scanning is performed from the second state. When the state changes to the first state, Bs and Be in that raster (raster ■ in FIG. 7) are compared with Ws and We in the previous raster (raster ■ in FIG. 7). Then, when the following conditions are satisfied, it is extracted as a hole.

Bs■<Ws■, We■(Be■ - (3) In Figure 7 (b), even if raster ■, which is flag 1, is considered to be a hole, it is determined that it is not a hole by checking raster ■ and ■. The operation flow of FIG. 3 shows the procedure of the hole extraction method in Example 3. Here, B sn.

Those with Be1l and n mean Bs and Be in raster n when changing from state 2 to state 1, and W
8n-i, n-1 of Well-□ is raster n
represents Ws and We in the previous raster.

In the case of the large-scale isolated area shown in FIG. 7, another checking process may be performed. That is, a raster scan is performed from the first state to change to the second state, and then a raster scan is performed to change to the first state.
Bs in the raster and the previous raster when the state changes to .

From Be, 11Be■-Bs■l-1Be■-Bs■II<TH-
When the condition (4) is satisfied, it is extracted as a hole. In this case, as in Example 2, Bs in raster ■,
The distance between Be and Bs in the previous raster ■
The condition is that there is no large difference between the distances between Be and Be.

FIG. 4 shows the operational flow of the hole extraction method of this embodiment.

(Effects of the Invention) As explained above, according to the present invention, by a simple method,
Furthermore, holes in images can be extracted efficiently with a short processing time.

[Brief explanation of the drawing]

1 to 4 are diagrams each showing an operational flow of an embodiment of the present invention, FIGS. 5 to 7 are diagrams each showing a specific example of an embodiment of the present invention, and FIG. 8 is a diagram showing a hardware FIG. IIBe=-BSkl-1 day eh-I-ask-dl<
THFigure 5Figure 6Figure 7

Claims (2)

[Claims] (1) In raster scanning of binary images such as characters,
Bs is the X coordinate of the leftmost black pixel, Be is the X coordinate of the rightmost black pixel, Ws is the X coordinate of the white pixel when moving from a black pixel to a white pixel between Bs and Be, and is the X coordinate of the white pixel when moving from a white pixel to a black pixel. Let We be the X coordinate of the white pixel at the time, and when the first state in which there is no white pixel between Bs and Be in one line of raster scan changes to the second state in which there is a white pixel, the following condition is satisfied. A hole extraction method characterized in that when at least one condition is satisfied, the hole is extracted as a hole. Condition 1: When changing from the first state to the second state at raster k, Bs_k_-_1<Ws_k, We_k<Be_k_-
If _1 is satisfied. Condition 2: When changing from the first state to the second state at raster k, ||Be_k−Bs_k|−|Be_k_−_1−Bs
＿k＿−＿１｜｜<TH is satisfied. However, those with k are Bs and B in raster k.
Those with e, Ws, We, and k-1 are 1 in the raster.
Bs, Be, and TH in the previous raster are predetermined thresholds. (2) In raster scanning of binary images such as characters,
Bs is the X coordinate of the leftmost black pixel, Be is the X coordinate of the rightmost black pixel, Ws is the X coordinate of the white pixel when moving from a black pixel to a white pixel between Bs and Be, and is the X coordinate of the white pixel when moving from a white pixel to a black pixel. When the X coordinate of the white pixel is We, the first state in which there is no white pixel between Bs and Be in one line of the raster scan changes to the second state in which there is a white pixel, and then the second state 1. A hole extraction method, characterized in that when at least one of the following conditions is satisfied when the state changes from a state to a first state, the hole is extracted as a hole. Condition 1: When changing from the second state to the first state at raster n, Bs_n<Ws_n_-_1, We_n_-_1<Be
If __n is satisfied. Condition 2: When changing from the second state to the first state at raster n, ||Be_n_−_1−Bs_n_−_1|−|Be_
When n-Bs_n||<TH is satisfied. However, those with n are Bs and B in raster n.
Those with e and n-1 are predetermined threshold values for Bs, Be, Ws, We, and TH in the raster immediately before raster n.