CN104574448A - Method for identifying connected pixel blocks - Google Patents

Abstract

The present invention is applicable to the field of image processing, and provides a method for identifying connected pixel blocks. The method includes: scanning an image to be identified line by line, and obtaining a run segment of each line of the image to be identified; Include the run nodes of the image in this line, divide the run nodes into target run nodes and background run nodes, and obtain the run information of the run nodes; used for identification based on the run information of the run nodes and the preset connectivity The relationship identifies whether the target run-length nodes of two adjacent rows are connected; update the pointer of the ThisRow target run-length node until the last row of image processing is completed. In the embodiment of the present invention, the image to be recognized is scanned line by line, and the run segment of each line of the image to be recognized is obtained, and the run segment is included in the run node of each line, and the run information of the run node is obtained, and identified according to the preset connectivity The relationship judges whether two adjacent lines of images are connected, and solves the phenomenon of label redundancy.

Description

A Recognition Method of Connected Pixel Blocks

technical field

The invention belongs to the field of image processing, in particular to a method for identifying connected pixel blocks.

Background technique

Blob refers to a block composed of pixels that have similar image features (such as color, texture, etc.) and are spatially connected. Blob analysis is a very important part of computer vision and image understanding. It can easily locate objects and calculate various geometric features of objects without any prior knowledge of the position, shape and size of objects in the image. , so that objects can be selected and classified based on these features.

The core of Blob analysis is the connected region detection algorithm. The current algorithms can be roughly divided into two categories according to the processing method: the method based on pixel scanning and the method based on run-length coding. Representative pixel-based marking methods are the recursive method and the two-scan method. The recursive method scans the image, finds the X point that has no mark, assigns it a new mark L, and recursively assigns the mark L to the neighbors of the X point until all points have marks. This method needs to scan the image repeatedly, and the space complexity and time complexity are high. The two-scanning method improves the above method: in the first scan, all connected areas are temporarily marked, and at the same time, the equivalence table is used to record all label redundancy; in the second scan, the equivalence table is used to use a specific search algorithm , for equivalence marker merging. The method based on run-length coding first performs run-length coding on the binary image, and then marks the connected regions of the coded image, and also uses the equivalence table to record the label redundancy. This method can effectively compress the image and make the search space smaller. , thereby reducing the space complexity of the algorithm search. Use the equivalence table to record mark redundancy to avoid repeated scanning of connected regions. However, when the number of connected regions in the image is large or the shape of the connected region is complex, the phenomenon of mark redundancy is serious, and the structure of the equivalence table is complex. Searching the equivalence table The algorithm to do equivalence tag merging is very time consuming. The images collected in the industrial environment have a lot of noise after binarization processing due to noise, lighting and other reasons. The algorithm based on the equivalence table is very time-consuming, and when the noise is very serious, the algorithm is very unstable and difficult. Applied to industrial visual inspection.

Contents of the invention

In view of this, the present invention provides a method for identifying connected pixel blocks to solve the technical problem of label redundancy in the prior art.

The embodiment of the present invention is achieved in this way, a method for identifying connected pixel blocks, the method includes the following steps:

Scan the image to be recognized line by line, and obtain the run segment of each row of the image to be recognized;

Count the run segment into the run node of the row image, divide the run node into a target run node and a background run node, and obtain the run information of the run node. The run information includes: tag information known, index Information index, start coordinate start, end coordinate end and gray value value;

It is used to identify whether the target run nodes in two adjacent rows are connected according to the run information of the run nodes and the preset connected identification relationship, wherein, ThisRow represents the image of the current row, and LastRow represents the image of the previous row;

Update the pointer of the ThisRow target run node until the last row image processing is completed.

In the embodiment of the present invention, the image to be recognized is scanned line by line, and the run segment of each line of the image to be recognized is obtained, and the run segment is included in the run node of each line, and the run information of the run node is obtained, and identified according to the preset connectivity The method uses the run-length linked list and the dynamic array method to judge whether two adjacent rows of images are connected, which only needs to scan the image once, and does not need to establish an equivalence table and merge equivalent marks, which solves the mark redundancy in the existing algorithm Phenomenon.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings that need to be used in the descriptions of the embodiments or the prior art. Obviously, the accompanying drawings in the following descriptions are only of the present invention. For some embodiments, those of ordinary skill in the art can also obtain other drawings based on these drawings without paying creative efforts.

Fig. 1 is a flowchart of a method for identifying connected pixel blocks provided by an embodiment of the present invention;

As shown in Figure 2a, it is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by an embodiment of the present invention;

As shown in Figure 2b, it is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention;

As shown in Figure 2c, it is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention;

As shown in Figure 2d, it is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention;

As shown in FIG. 2e, it is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention;

As shown in Figure 2f, it is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention;

As shown in Figure 2g, it is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention;

FIG. 2h is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by yet another embodiment of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

In order to illustrate the technical solutions of the present invention, specific examples are used below to illustrate.

Embodiment one

As shown in Figure 1, it is a flowchart of a method for identifying connected pixel blocks provided by an embodiment of the present invention, and the method includes the following steps:

Step S101, scan the image to be recognized line by line, and acquire the run length of each line of the image to be recognized.

In the embodiment of the present invention, the recognition system scans the image to be recognized line by line, so as to obtain the run segment of each line of the image to be recognized.

Step S102, counting the run segments into the run nodes of the row image, dividing the run nodes into target run nodes and background run nodes, and obtaining the run information of the run nodes, the run information including: label information known, index information index, start coordinate start, end coordinate end and gray value value.

In the embodiment of the present invention, after the recognition system acquires the run segment of each line of image, the run segment is included in the run node of the row image and divided into the target run node and the background run node, and the run information of the run node is obtained at the same time, Among them, the run length information includes but not limited to: mark information known, index information index, start coordinate start, end coordinate end and gray value value, where the meaning expressed by the run length information is as follows:

Mark information: used to indicate whether the run node is marked.

Index information: used to indicate the index of the run node in the image of the row, and record the position information of the run node in the image of the row.

Start coordinates: It is used to indicate the start coordinates of the run node in the current line image.

Termination coordinates: It is used to indicate the termination coordinates of the run node in the current line image.

Gray value: used to represent the gray value of the run node.

Step S103, for identifying whether the target run nodes in two adjacent rows are connected according to the run information of the run nodes and the preset connection identification relationship, where ThisRow represents the image of the current row, and LastRow represents the image of the previous row.

In the embodiment of the present invention, after the identification system acquires the run information of the run nodes, it identifies whether the target run nodes in two adjacent rows are connected according to the run information, wherein the preset connected identification relationship is: LastRow.srart -1≤ThisRow.end, ThisRow.start-1≤LastRow.end and LastRow.value=ThisRow.value. The method for identifying whether the target run nodes in two adjacent rows are connected according to the run information of the run nodes is specifically:

(1) If Thisrow.start-1>LastRow.end, then the target run nodes of two adjacent rows are not connected.

In the embodiment of the present invention, when Thisrow.start-1>LastRow.end, the target run-length nodes of two adjacent rows do not meet the preset connected identification relationship, and the tag information of the target run-length node is: LastRow→index++, that is, points to The next run of the LastRow row continues to judge. Figure 2a is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by an embodiment of the present invention, wherein white is the target run node, and black is the background run node.

(2) If LastRow.end+1≥ThisRow.start≥LastRow.start and ThisRow.end>LastRow.end, then the target run nodes of two adjacent rows are connected.

In the embodiment of the present invention, when LastRow.end+1≥ThisRow.start≥LastRow.start and ThisRow.end>LastRow.end, the target run nodes of two adjacent rows satisfy the preset connected identification relationship, and the LastRow target run If the start coordinate and end coordinate of the node are smaller than the target run node of ThisRow, the following operations are performed: the target run node of ThisRow is classified into the target run node of LastRow, LastRow→index++. FIG. 2b is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention, wherein white is the target run nodes, and black is the background run nodes.

(3) If ThisRow.start≥LastRow.star and ThisRow.end<LastRow.end, then the target run-length nodes of two adjacent rows are connected.

In the embodiment of the present invention, when ThisRow.start≥LastRow.star and ThisRow.end<LastRow.end, the target run-length nodes of two adjacent rows satisfy the preset connected identification relationship, and the LastRow target run-length node is more than the ThisRow target run-length node The starting coordinate of is smaller but the ending coordinate is larger, then the following operation is performed: the target run node of ThisRow is classified into the target run node of LastRow, ThisRow→index++. FIG. 2c is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention, wherein white is the target run nodes, and black is the background run nodes.

(4) If ThisRow.start≥LastRow.star and ThisRow.end=LastRow.end, then the target run nodes of two adjacent rows are connected.

In the embodiment of the present invention, when ThisRow.start≥LastRow.star and ThisRow.end=LastRow.end, the target run-length nodes of two adjacent rows satisfy the preset connected identification relationship, and the LastRow target run-length node is more than the ThisRow target run-length node If the start coordinates are smaller and the end coordinates are the same, the following operations are performed: classify the target run node of ThisRow into the target run node of LastRow, ThisRow→index++, LastRow→index++. FIG. 2d is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention, wherein white is the target run nodes, and black is the background run nodes.

(5) If ThisRow.start<LastRow.star and ThisRow.end=LastRow.end, then the target run nodes of two adjacent rows are connected.

In the embodiment of the present invention, when ThisRow.start<LastRow.star and ThisRow.end=LastRow.end, the target run-length nodes of two adjacent rows satisfy the preset connected identification relationship, and the LastRow target run-length node is more than the ThisRow target run-length node The starting coordinates should be large, so the target run node of ThisRow may be connected with the target run node of the previous line of the LastRow target run node at the same time, that is, the target run node of ThisRow may have been marked, so the following processing is required: connect the target run node of ThisRow with the LastRow The target run nodes of are merged and do the following:

if(T.known==0) //judging that T is not marked

{Classify the ThisRow target run node into the LastRow target run node;}

else //T has been marked

{Merge the target run node of ThisRow with the target run node of LastRow;}

LastRow→index++;

ThisRow→index++;

FIG. 2e is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention, wherein white is the target run nodes, and black is the background run nodes.

(6) If ThisRow.start<LastRow.star and ThisRow.end>LastRow.end, then the target run nodes of two adjacent rows are connected.

In the embodiment of the present invention, when ThisRow.start<LastRow.star and ThisRow.end>LastRow.end, the target run-length nodes of two adjacent rows satisfy the preset connected identification relationship, and the LastRow target run-length node is more than the ThisRow target run-length node The starting coordinates are large, and it is necessary to determine whether the ThisRow target run node is marked, then perform the following operations:

if(ThisRow.known==0) //Judge ThisRow is not marked

{Associate the ThisRow target run node into the LastRow target run node;}

else //Determine that ThisRow has been marked

{Merge the target run node of ThisRow with the target run node of LastRow;}

LastRow→index++.

FIG. 2f is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention, wherein white is the target run nodes, and black is the background run nodes.

(7) If LastRow.end>ThisRow.end>LasrRow.start-1 and ThisRow.start<LastRow.start, then the target run nodes of two adjacent rows are connected.

In the embodiment of the present invention, when LastRow.end>ThisRow.end>LasrRow.start-1 and ThisRow.start<LastRow.start, the target run nodes of two adjacent rows satisfy the preset connected identification relationship, and the LastRow target run The node is larger than the start coordinates of the ThisRow target run node, and the end target is also larger. To determine whether the ThisRow target run node is marked, perform the following operations:

if(ThisRow.known==0) //Judge ThisRow is not marked

{Associate the ThisRow target run node into the LastRow target run node;}

else //Determine that ThisRow has been marked

{Merge the target run node of ThisRow with the target run node of LastRow;}

LastRow→index++.

FIG. 2g is a schematic diagram of the connection relationship between two adjacent rows of target run nodes provided by another embodiment of the present invention, wherein white is the target run nodes, and black is the background run nodes.

(8) If ThisRow.end<LastRow.start-1, the target run nodes of two adjacent adjacent rows are not connected.

In the embodiment of the present invention, when ThisRow.end<LastRow.start-1, the target run-length nodes of two adjacent rows do not meet the preset connected identification relationship, then perform the following operation: ThisRow→index++, pointing to the next run-length of ThisRow judge. FIG. 2h is a schematic diagram of the connectivity relationship between two adjacent rows of target run nodes provided by yet another embodiment of the present invention, wherein white is the target run nodes, and black is the background run nodes.

Step S104, updating the pointer of the ThisRow target run node until the last row of image processing is completed.

In the embodiment of the present invention, the image to be recognized is scanned line by line, and the run segment of each line of the image to be recognized is obtained, and the run segment is included in the run node of each line, and the run information of the run node is obtained, and identified according to the preset connectivity The method uses the run-length linked list and the dynamic array method to judge whether two adjacent rows of images are connected, which only needs to scan the image once, and does not need to establish an equivalence table and merge equivalent marks, which solves the mark redundancy in the existing algorithm Phenomenon.

As an optional embodiment of the present invention, before the step of identifying whether two adjacent rows of target run nodes are connected according to the run information of the run nodes and the preset connected identification relationship, the method further includes The following steps:

The connection identification relationship is preset.

Those of ordinary skill in the art can also understand that all or part of the steps in the method of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, so The storage medium mentioned above includes ROM/RAM, magnetic disk, optical disk, etc.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention should be included in the protection of the present invention. within range.

Claims (4)

1. A method for identifying connected pixel blocks, characterized in that the method comprises the following steps: Scan the image to be recognized line by line, and obtain the run segment of each row of the image to be recognized; Count the run segment into the run node of the row image, divide the run node into a target run node and a background run node, and obtain the run information of the run node. The run information includes: tag information known, index Information index, start coordinate start, end coordinate end and gray value value; It is used to identify whether the target run nodes in two adjacent rows are connected according to the run information of the run nodes and the preset connected identification relationship, wherein, ThisRow represents the image of the current row, and LastRow represents the image of the previous row; Update the pointer of the ThisRow target run node until the last row image processing is completed. 2. The method according to claim 1, wherein, before the step of identifying whether the target run nodes in two adjacent rows are connected according to the run information of the run nodes and the preset connected identification relationship, The method also includes the steps of: The connection identification relationship is preset. 3. The method according to claim 1, wherein the preset connected identification relationship is: LastRow.srart-1≤ThisRow.end, ThisRow.start-1≤LastRow.end and LastRow.value=ThisRow .value. 4. The method according to claim 3, wherein the step of identifying whether the target run nodes of two adjacent rows are connected according to the run information of the run nodes is specifically: If Thisrow.start-1>LastRow.end, the target run nodes of two adjacent rows are not connected; or, If LastRow.end+1≥ThisRow.start≥LastRow.start and ThisRow.end＞LastRow.end, the target run nodes of two adjacent rows are connected; or, If ThisRow.start≥LastRow.star and ThisRow.end<LastRow.end, then the target run nodes of two adjacent rows are connected; or, If ThisRow.start≥LastRow.star and ThisRow.end=LastRow.end, then the target run nodes of two adjacent rows are connected; or, If ThisRow.start≥LastRow.star and ThisRow.end=LastRow.end, then the target run nodes of two adjacent rows are connected; or, If ThisRow.start<LastRow.star and ThisRow.end>LastRow.end, then the target run nodes of two adjacent rows are connected; or, If ThisRow.start<LastRow.star and ThisRow.end>LastRow.end, then the target run nodes of two adjacent rows are connected; or, If ThisRow.end<LastRow.start-1, the target run nodes of two adjacent adjacent rows are not connected.