Edge Detection Algorithm Based On The Direction Of The Differences Between Multi-scale Analysis Of Digital Images

Edge detection is an important topic in the field of digital image processing. In this field the main method contains:Classical edge detection operator, Wavelet analysis method, surface fitting law, Mathematics morphology method, and so on. In this paper, firstly, I describe the history and current situation of digital image processing and image edge detection; secondly, I recall the classical edge detection operator; thirdly, I review the theory of Wavelet analysis and its associated method of edge detection; Finally, I raise better edge detection algorithm--the edge detection algorithm based on the different change in different direction between the real edge and the noise. Until now, it's still hard to distinguish between noise and real edge, and the result of edge detection often be interfered by noise. Considering the difference between the noise and the real edge:in the gradient direction and the non-gradient direction the modulus of noise can change largely, however, the real edge doesn't have this characteristic, in this paperⅠraised the edge detection algorithm based on the different change in different direction between the real edge and the noise. Considering the difference between noise and real edge:After being carried out wavelet transform, the local maximum of modulus caused by noise can be quickly enlarged along with the increase of scale, however, the local maximum modulus caused by the real edge can be quickly reduced along with the increase of scale, plus the edge result detected by wavelet transform in low scale has much more noise and rich detail, in high scale has much less noise and a rough outline, in this paperⅠraised the edge detection algorithm based on the harmonization with different scales.