Image partitioning method based on mixed bipartite graph clustering integration

Abstract

The invention discloses an image partitioning method based on mixed bipartite graph clustering integration, and mainly solves the problems that type marks are required to be registered during integrated, and information utilization is insufficient in the prior art. The method comprises the following steps of: (1) inputting an image to be partitioned; (2) judging whether the image is a color image; (3) extracting image characteristics; (4) generating a characteristic data clustering grade; (5) generating a solution set; (6) generating a cascade block matrix; (7) constructing a mixed bipartite graph; (8) generating an embedding matrix; (9) clustering K mean values; (10) marking the image; and (11) partitioning the image. By the method, data and type information in a primary partitioning result of the image are effectively used, so that more details can be found; during integration, the type marks are not required to be registered, so that calculation resources are saved; all the image pixel characteristic data are integrated; and the problem that a sub test sample set has high influence on the partitioning result is solved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to an image segmentation method based on mixed bipartite graph clustering and integration in the field of image segmentation. The invention can be used for region segmentation of color RGB images or synthetic aperture radar images to achieve the purpose of identifying targets. Background technique [0002] Image segmentation is one of the basic problems in image processing, and it is the basis for realizing target recognition on images. The task of image segmentation is to divide the image into disjoint regions, each region satisfies a specific regional consistency, and different regions have significant differences. Image segmentation methods can be divided into two categories: region-based and edge-based. Among the region-based methods, threshold segmentation and spatial clustering methods are the most commonly used. The spatial clustering method to segment an im...

AI Technical Summary

Problems solved by technology

The present invention extracts the feature of each pixel of the image, and combines the multi-group segmentation results of all data feature samples of the image by clustering based on the mixed bipartite graph, so as to solve the disadvantage that the existing image segmentation technology is greatly affected by sub-test samples, and can Effectively merge multiple sets of initial segmentation results of images to achieve image region segmentation

Images