An Image Edge Detection Method Based on Wavelet Transform Modulus Maximum.

Wavelet transform is a multi-scale signal analysis method. It has the characteristics of low entropy, good timefrequency characteristic, multi-scale characteristic and de-correlation, which makes the wavelet transform widely used in image de-noising and edge detection. The multi-scale detection method mainly uses multiple edge detection operators of different scales to detect the edge of image correctly by effective combination. Wavelet multiscale edge detection method not only has good noise suppression capability, but also has complete edge preservation characteristics, which can take advantage of size and scale to obtain precise single pixel wide edge in edge extraction. The proposed algorithm in this paper uses the wavelet transform to detect the jump edge of the image in row direction and column direction, and then the image edge can be formed under certain rules, and useful edge information can be extracted by means of the multi-scale wavelet transform and the difference between signal and noise under the wavelet transform to multiply wavelet coefficient on the adjacent scale. In suppressing noise at the same time, such algorithm implies the the causal relationship of the edge on different scales, and holds that the edge point is the local modulus maximum point of the scale volume coefficient through the thresholding operation, and remove the noise interference and finally the detailed experiment is carried out to prove that this method can effectively solve the contradiction between extracting edge and restraining noise. [ABSTRACT FROM AUTHOR]