Image segmentation method based on iterative self-organization and multi-agent genetic clustering algorithm

Abstract

The invention discloses an image segmentation method based on iterative self-organization and multi-agent genetic clustering algorithm, which mainly solves the problem in the prior art that the segmentation results are overly dependent on initial parameters and easily fall into local optimum. The segmentation steps are: 1) extract the gray information of the image to be segmented; 2) apply the idea of ​​the iterative self-organizing algorithm ISODATA algorithm to the image to be segmented to obtain the optimal number of clusters; 3) according to the optimal number of clusters, use The multi-agent algorithm framework clusters the image to be segmented to obtain the optimal clustering label; 4) according to the optimal clustering label, classifies the pixels of the image to be segmented to achieve image segmentation. The invention does not need to explicitly determine the number of clusters, has good convergence effect, can easily obtain the global optimal value, can improve the quality of image segmentation, enhances the stability of segmentation results, and can be used for image target extraction and recognition and other subsequent processing.

Description

technical field [0001] The invention belongs to the technical field of digital image processing, relates to an image segmentation method, and can be used in the field of pattern recognition and computer vision. Background technique [0002] Image segmentation is a key technology in image processing, which is widely used in image processing research. For example, target recognition and target measurement are all based on image segmentation, and the results of image segmentation directly affect the follow-up tasks. Therefore, the research on image segmentation is of great significance. Image segmentation is a special image processing technology, which is essentially a process of classifying according to the image pixel attributes, namely grayscale, texture, and color. The commonly used methods in the existing image segmentation methods include image segmentation methods based on clustering and image segmentation methods based on edge extraction. Among them, the application o...

AI Technical Summary

Problems solved by technology

The advantage of this algorithm is that it is simple and easy to operate, but at the same time it also brings a lot of inconvenience, such as: depending on the number of clustering types initially set, easy to fall into local optimum, unsatisfactory clustering results, etc.

In order to solve this kind of problem, researchers have made many attempts. Some people use the combination of genetic algorithm GA and clustering algorithm, and get satisfactory results. However, due to the limitations of the global evolution mechanism of traditional genetic algorithm, the combined clustering method It still has the disadvantages of relying on the initial set category of clustering and easily falling into local optimal values, which leads to a decrease in the quality of image segmentation results and a decrease in the stability of segmentation results, which is not conducive to subsequent image analysis and processing

Images