Method using improved intuitionistic fuzzy clustering algorithm to segment infrared ship image

Abstract

The invention relates to a method using an improved intuitionistic fuzzy clustering algorithm to segment an infrared ship image. The method includes: using a threshold obtained by an image variance matrix and a Chebyshev inequality to determine the possible area of a ship, and using the center of the possible area as the preset ship center; using the improved intuitionistic fuzzy clustering algorithm to segment the infrared ship image. The method has the advantages that the ship central position is considered, and the ship central positon is used to build the distance measurement from pixels to different clustering centers; the regular terms built by area variance is used to guide neighborhood information into a target function; the improved algorithm sufficiently considers the influence of the ship central position on data clustering, reasonably utilizes the neighboring information, and accordingly the infrared ship image segmentation result is improved evidently, and the method is promising in market prospect and high in application value.

Description

【Technical field】 [0001] The invention relates to a method for segmenting an infrared ship image by using an improved intuitive fuzzy clustering algorithm. The fuzzy clustering technique and the image segmentation technique are widely used in the field of image application and belong to the field of digital image processing. 【Background technique】 [0002] Image segmentation is an image processing technology that divides the image into several regions with unique properties according to the characteristic information of the image (such as grayscale, shape and gradient, etc.). Image segmentation adds the same label to pixels with common visual characteristics, which simplifies the representation of the image to a certain extent, and is beneficial to the understanding and analysis of the image. Generally, image segmentation is a very important preprocessing technique in the field of computer vision and image recognition. Therefore, a fast, accurate and universal image segment...

AI Technical Summary

Problems solved by technology

However, in the infrared ship image, the coast or river bank is usually equivalent to the pixel value of the ship target, so the general fuzzy clustering algorithm will misclassify it as the target class

At the same time, the mast or radar antenna part of the ship is often relatively small and trivial, and it is not easy to be classified with the main body of the ship when considering the influence of image noise, that is, some targets are lost

Images