Research on Thangka Image Edge Detection Based on Differential Operator

Edge detection is not only the early work of image segmentation, but also the basis of image understanding, differential operator is a simple technique for edge detection and is typically accomplished by calculating gradients of horizontal and vertical gray values. However, there are many curves that two-direction operator can detect that is not so clear in Thangka image. In this paper, the edge detection effects of several common differential operators on Thangka image are compared. On the basis, a new algorithm for Thangka image detection is proposed. Fist the image is grayed. The the grays image is contrast reduced. At last edge detection of image is carried out by using four-direction Sobel operator.