To Describe the Complexity of Image Edge Based on the Information Entropy Method

Abstract

A method of image edge complexity description based on information entropy is proposed. Firstly, the long axis of the object is rotated to the horizontal direction, then the target image is united in size, and then the connected region is computed to obtain the coordinates of the target edge points. The first order difference and the second order difference of the adjacent edge point coordinates are obtained in turn. A number of statistical difference and second order difference take different values, calculating the probability of different value. According to the entropy formula, the first order differential information entropy and the second order differential information entropy of the edge are calculated. This method can avoid the effect of target rotation on calculation results, and can describe the complexity of the target edge very well, and it can be used as one of the identification features of the target.