Repair Algorithm Of Image Edge Detection

Image inpainting is the technology which is used to populate the pixel of the region of loss by using the image around the damaged region. It is widely used in inpainting the old movies and old photographs as well as many other aspects. Color image inpainting often requires determining the scale of regions to restore. Cluster analysis is a useful tool to extract the image edges.Three kinds of cluster analysis algorithms for extracting image edges have been studied in this paper from different levels. The first algorithm is the gray value based simple Euclidean distance cluster. This method firstly requires to covert the color images into gray images, then several cluster centers are determined and all the pixels are clustered to the cluster center that gives the smallest gray difference. The second algorithm is called the Euclidean distance-based dynamic clustering. It first assigns all the pixels a cluster center from the initial cluster centers on the principle of the nearest Euclidean distance and then updated all the centers by the average distance within the cluster until the new cluster centers converged. The third algorithm is color-image-based K-means clustering. This algorithm first assigns eight kinds of colors randomly as the origin cluster centers and then assigns each pixel to a nearest origin cluster. After this the origin cluster centers are updated iteratively by the average of all the pixels in the cluster until the new cluster centers unchanged. One of the advancement in the paper has adopted a kind method of cluster analysis. As a method of image edge extraction, cluster analysis has become one of the most common algorithms because of its simplification, fast and high efficiency. Consequently, the studies on cluster algorithms are of great practical significance.One of the methods to inpaint the image is convolution-based accounting method. This paper makes a deep research on the process of inpainting the pixel of the region of loss, and has achieved very good results. Convolution is a linear operator and an infinitesimal element of the cumulative form of multiplication. The common mask computing in image processing belongs to convolution operation. This technology is widely used in image filtering. Another of the advancement in this paper has used the image processing software imageboard. when inpainting the damaged regions by using convolution algorithm, the first step is to get the relevant coordinates of damaged region by dealing with the damaged region by the image processing software imageboard. The second step is to inpaint the image by populating the pixel of the region of loss through relevant coordinates and the non-damaged edge of the damaged region, according to the convolution multiplication of micro-element of the cumulative form.