JPS61243569A - System for labeling to digital picture area - Google Patents

Abstract

PURPOSE:To reduce an arithmetic amount necessary for labeling and to speed up the operation by providing a means which detects all provisional labels belonging to each area in an input digital picture based on the adjusted connection relationship and updates said labels. CONSTITUTION:A provisionally labeling part 10 puts sequentially provisional labels on respective areas in each unit picture at every partial picture of the input digital picture. A connection relationship detecting part 11 detects the connection relationship between the arbitrary provisional label in the partial picture and that adjacent to said label, and a connection relationship detecting part 12 detects the connection relationship of each provisional label between partial pictures in the boundary of each partial picture. Then provisional label adjusting parts 13 and 14 adjust the duplicated connection relationship of the provisional labels at every each part picture and partial picture, and a direct label detecting part 15 produces directly a connection table, which is transmitted to a label tracing part 16. Corresponding to the traced result a provisional label updating part 17 updates the label. In such a way the arithmetic amount can be reduced and a high speed processing can be attained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention divides an image region and deletes redundant connection relationships within and between partial images, thereby labeling a large number of connected regions. This paper relates to a labeling method for digital image regions that achieves high speed.

In image processing, there is a method of attaching a label to each pixel and performing the processing. Since the number of pixels involved in attaching such labels is enormous, it is important for high-speed image processing to be able to perform labeling processing at high speed.

[Conventional technology]

The conventional method for labeling image regions is to first attach a provisional label to the entire target image, detect and organize the connection relationships of the provisional labels, perform clustering processing on the provisional labels, and perform this processing. There is a method that attempts to enforce the desired labeling by updating the temporary label that has been created.

[Problem that the invention seeks to solve]

In the above-described method, it is essential to organize the connection relationships of provisional labels, and the amount of calculation required for this is proportional to the square of the maximum value of the number of provisional labels included in the image area. Because such an amount of calculation is required, high-speed processing cannot be expected, especially when an image including a large number of connected regions must be processed.

[Means for solving problems]

The present invention aims to solve the above-mentioned problems as much as possible.
The present invention provides a method for labeling digital image regions, which includes means for sequentially adding a temporary label to each region within each unit image for each partial image of an input digital image; Partial intra-image flaw connection detection means for detecting the connection relationship between any temporary label in an image and temporary labels in its vicinity, and inter-partial image connection for detecting the connection relationship of temporary labels between adjacent partial images. relationship detecting means; provisional label organizing means for deleting overlapping connections of provisional labels for each partial image and between partial images; The present invention is configured to include means for detecting a temporary label and updating the temporary label.

[Effect]

According to the method of the present invention, provisional labels are sequentially added to each region within each unit image for each partial image, and the connection relationship between any provisional label in each partial image and its neighboring provisional labels is Connection relationships of provisional labels between adjacent partial images are detected, and among the detected connection relationships, connection relationships of duplicate provisional labels that exist for each partial image and between partial images are deleted and organized. All provisional labels belonging to each area in the input digital image are detected based on the connection relationship thus determined, and each processing area is labeled.

As a result, the amount of calculation required for labeling processing can be significantly reduced, and the processing speed can be greatly promoted.

〔Example〕

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 shows an embodiment of the invention. In this figure, 1
0 is a temporary labeling section, which scans a field-shaped window as shown in FIG. 2 over partial images that are input sequentially, and generates and extends labels as shown in FIG. 3. . Such label generation and extension processing itself is a known method. Then, the connection relationship detection unit (within the partial image) 11 operates according to the logic shown in FIG. 4 to detect the connection relationship between any temporary label within the partial image and the temporary label adjacent thereto, and The connection relationship detection unit (between partial images) 12 operates according to the logic shown in FIG. 5, and detects the connection relationship of each provisional label between partial images at the boundary of each partial image. 6th
The figure shows an example of an image and detected connection relationships after undergoing the provisional labeling process as described above.

Partial image flaw connection relationship detection unit] The connection relationship data detected in step 1 (1) of fa in FIG.
Overlapping portions of connection relationships supplied to 1 to 134 and indicated by those data (underlined connection relationship portions in Figure 7)
is deleted as a redundant part, and an organized provisional label (shown in (bl 1) in FIG. 7) is output. In addition, the connection relationship data ((a) 2 in FIG. 7) detected by the inter-partial image connection relationship detection unit 12 is also processed by the provisional label organizing unit 141 in the same manner.
to 143, duplicate connection relationships are deleted (as shown in (b) 2) in FIG. 7), and a sorted provisional label is output.

These organizing sections 131 to 134;14. From the organized connection relationships output from 143 to 143, small areas (areas given the same provisional label) within each partial image are identified.
The direct connection label detection unit 15 detects all directly connected small areas among the small areas and creates a direct connection table as shown in FIG. It is created in steps 154 to 154. In FIG. 8, a partial image area identification code 1: is added to the provisional label in each partial image to distinguish the labels.

The direct connection table is supplied to the label tracking unit 16, which searches for labels based on the contents of the direct connection table and detects all provisional labels that are directly or indirectly connected (this is called clustering). . This allows all temporary labels contained within each area to be detected.

Temporary label updating unit 1 that responds to this label tracking result output
In step 7, the temporary labels added to the small areas belonging to the areas in each partial image that should be the same area in the image area after provisional labeling U are updated to the same new label (the Figure 9) In the label area, processed image area data is output.

An example of how the amount of calculation can be reduced when dividing the image area and organizing the temporary labels as described above is as follows. That is, the size of the image is β×4, the maximum value of its provisional label is Lm,
Assuming that the case where the above division is not performed is (A), and the case where the division is performed (however, 8 divisions) is (B),
The differences between these two are shown in the table below.

As you can see from the table above, if you divide the image and organize its temporary labels, the amount of calculation will be approximately 1 in the case of 8 divisions.
This is a significant decrease to /4. This is particularly effective in processing complex images with a large number of regions where the maximum value Lm of the temporary label is large.

Furthermore, since the temporary label sorting sections can be operated in parallel, high-speed processing can be achieved.

〔Effect of the invention〕

As explained above, according to the present invention, the following effects can be obtained: (1) It provides a means for significantly reducing the amount of calculation required for labeling processing, and (2) It contributes to speeding up the processing.

[Brief explanation of drawings]

Fig. 1 is a diagram showing an embodiment of the present invention, Fig. 2 is a diagram showing an image scanning mode, Fig. 3 is a diagram showing a list of the relationships between label generation and extension processing, and Fig. 4 is a diagram showing an example of the present invention. Figure 5 is a diagram illustrating the process logic for detecting connection relationships within partial images. Figure 5 is a diagram illustrating the process logic for detecting connection relationships between partial images. Figure 6 is the image and detection after provisional labeling processing. Figure 7 is a diagram illustrating the process of deleting duplicated parts of connection relationships. Figure 8 is a diagram showing part of an example of creating a direct connection table. Figure 9
The figure shows a state in which a temporary label is updated to the same new label. In FIG. 1, reference numeral 10 denotes a four-page provisional label section; 11 denotes a section for detecting connection relations within partial image scratches; 12 indicates a section for detecting connection relations between partial images; 131 to 134; 141 to 143&;l provisional label sorting section; 154 is a direct connection label detection section, 16 is a label tracking section, and 17 is a provisional label updating section. ! [Seki Iho, who will be joining Sanskrit at the end of his 17th year] Sound Giant Car Picture Book Yoshiuchi (1) Sound l15 power #image I (3) Sound 15
Rin mountain” 1t-3 (2) sound 15 fraction 1 good 2 (Z1222245,, part game h1
Fu Sekiji (1) ■=■ (2) ■=■ (3) ■=qsu

Claims (1)

[Claims] Temporary labeling means that sequentially adds a temporary label to each region within each unit image for each partial image of an input digital image, and a connection relationship between an arbitrary temporary label within each partial image and temporary labels in its vicinity. an inter-partial image connection relation detection means for detecting the connection relation of provisional labels between adjacent partial images; and an inter-partial image connection relation detection means for detecting the connection relation of temporary labels between adjacent partial images; The present invention is configured to include provisional label organizing means for deleting connection relationships, and means for detecting all provisional labels belonging to each area in an input digital image based on the organized connection relationships and updating the provisional labels. A labeling method for digital image areas characterized by: