Research On The Edge Detection Method For Image Segmentation

Image segmentation is a key step in image processing. This paper starts from the image edge detection in image segmentation research, mainly focus on two theoretical methods:phase information theory and ant colony algorithm, and applies these two theories to image edge extraction. Phase information theory is a new algorithm which is rising lately in recent years. This algorithm is not affected by image brightness and contrast, so it has strong advantage in image edge detection. Ant colony algorithm is a new optimization algorithm, it is rarely used in the image edge detection. In this paper, ant colony algorithm is applied to image edge detection.This paper firstly studies several classical edge detection operator of the basic concepts and principles of segmentation. With simulation experiment, these types of edge detection operator's respective characteristics in a noisy environment and the noise-free environment are discussed.Then a new image edge detection algorithm is proposed, whose core is phase information theory, combined with the theory of mathematical morphology. In color image processing, to realize better-performing color image edge detection, color space multi-channel integration of technology is introduced.Finally, the paper systematic study of the basic principles of ant colony algorithm, and ant colony algorithm for image edge detection is improved. The Otsu algorithm and pheromone update strategy are introduced into the algorithm.Experimental results show that the proposed algorithm provides good performance.