K-Means Sobel Algorithm in Edge Extracting of Mural Diseases

Currently, it is an effective means of collecting diseases information from mural collections by using edge detection and manual interpretation. But it is hard to extract better edges of mural diseases with existing edge detection algorithms. Therefore, a new K-Means Sobel algorithm is proposed. Firstly, we process a gray orthophoto map with K-means clustering algorithm to obtain a gray value matrix and then convolve this matrix with Sobel formula. Converting the processed matrix to a graph, we can finally get what information we want- a black-white binary edge image that contains edges of mural diseases. Considering the practical significance of research on mural collections, we give two evaluation factors to judge the effect of edge detection. Experimental results demonstrate that we can get a better effect by using new method than traditional algorithms.