Image segmentation method based on similarity interaction mechanism

Abstract

The invention provides an image segmentation method based on a similarity interaction mechanism, which mainly solves the problems of low segmentation efficiency, poor region homogeneity and detail information loss of the prior segmentation algorithm. The implementation steps of the image segmentation method are as follows: (1) the characteristics of each pixel point of an image to be segmented are extracted; (2) the characteristics of region blocks are obtained; (3) the similarity among the region blocks is calculated; (4) phase values of the region blocks are obtained; (5) the region blocks are classified; and (6) an image segmentation result is output. The image segmentation method has the advantages of high segmentation efficiency, strong region homogeneity of the segmentation result, much detail information and good edge effect, can be effectively used for segmenting texture images and synthetic aperture radar (SAR) images and can be used for recognizing image targets.

Description

technical field [0001] The invention belongs to the technical field of computers, and further relates to an image segmentation method based on a similarity interaction mechanism in the technical field of image processing. The invention divides the image into regional blocks, classifies the regional blocks, realizes the segmentation of texture images and SAR images, and can be applied to target recognition. Background technique [0002] Image segmentation is to separate different regions with special meaning in the image according to certain characteristics of the image, and each region has the consistency of a specific region, while the attribute characteristics between adjacent regions have obvious differences. Image segmentation is an important issue in image processing. It plays a role in connecting the past and the future in the research of image analysis. It is not only a test for the effect of all image preprocessing, but also the basis for subsequent image analysis an...

AI Technical Summary

Problems solved by technology

But so far, there is no general method, and there is no objective standard for judging whether the segmentation is successful.

Although the threshold method is simple, it only considers the gray information of the image, does not consider the neighborhood information of the image pixels, and ignores the spatial information of the image, so it is difficult to obtain accurate results.

At present, many clustering algorithms have been used in image segmentation. K-means clustering is one of the simplest and most commonly used methods. It uses iterative optimization to find the optimal solution and performs on a compact hyperspherical distribution data set. However, when the data structure is non-convex, or the data points overlap each other seriously, the algorithm often fails, and the algorithm cannot guarantee to converge to the global optimal solution

The FCM algorithm also does not consider the spatial information of the image, and only clusters all samples as scattered sample points, resulting in poor regional consistency in the final segmentation results, and there are noise points inside the region. Relatively sensitive, easy to fall into local optimum, resulting in poor segmentation effect

The image segmentation technology based on graph theory usually maps the image into a weighted undirected graph, which essentially transforms the image segmentation problem into a graph optimization problem, and the optimal partition problem of a graph is an NP-hard problem, which makes the graph At the same time, the segmentation method based on graph theory only uses the information of adjacent pixels or regions, but ignores the global information of the image.

The disadvantage of this method is that although the influence of noise is suppressed in terms of the characteristics of the image, the subsequent fuzzy C-means clustering method does not consider the spatial information of the image. At the same time, it is sensitive to the initial value and is easy to fall into local Optimal, resulting in poor regional consistency in segmentation results

Images