Color Image Segmentation Method Based On Infrared Image Edge Information

Image segmentation refers to the image into regions each with characteristics and to extract the region of interest in which the methods and techniques. Image segmentation in medical, aerospace, military and other fields have played a huge role. With the rapid development of collection equipment, people demand more and more diverse and complex in image processing. Color image contains a rich diversity of color feature information, which is not a substitute for other image sources, therefore, the processing of color images of people being more and more attention.This paper presents a new idea of color image segmentation, it through infrared images and color images combine for complimentarily of information to complete and accurate extraction the goal of color image. Through a multi-source device mediation infrared image, using the histogram method to enhance the edge contrast of infrared images, the edges of the image more clearly, help to improve edge detection in the follow-up on the edge of the LOG operator recognition. Extraction of the contours of the longest edge line identified as the target edges, can be extracted the target area of interest from the infrared image. At the same time the profile information change into the chain code for information storage, the use of 8 Neighborhood code to record of coordinates and direction of the edge, and applied to the same point color image, until the close link to extracted from the region of interest from color images for image segmentationCompare the traditional image segmentation methods, this method solve color image segmentation for color space conversion requirements, and will avoid the extraction of single-channel color images and multi-channel images calculated to restore the quality of the image itself. Compared to color image segmentation operations, the use of low information content of infrared image, gray contrasting characteristics to distinguish object and background, the target can more easily extract the edge. In the natural environmental conditions, this method solution between the target and background color of a high degree of close, hard threshold method for effective segmentation problems. Infrared images provide the basis for completion of the conditions, thus avoiding the algorithm operation is conducive to maintaining the image pixels invariance.The main contents include the following:1. Using multiple sources of existing images for image processing methods were reviewed, the existing multi-source analysis of image acquisition equipment, study and master the characteristics of color image color space.2. On the infrared image analysis and processing, including the infrared image contrast enhancement, analysis and comparison of edge detection algorithms, achieve maximum extraction of the edge, as well as records of appropriate information. Infrared image edge information processing method: line transformation and comparative analysis of the grid sampling, and on the edge of the record and read information programming.3. Using MATLAB 7 to implement of color images segmentation. Compared with the traditional method of color image segmentation image segmentation efficiency, response time and subjective comparison segmentation.