Image edge detection based on the grey prediction model and discrete wavelet transform

A novel image edge detection algorithm is presented by combining grey prediction model and discrete wavelet transform. Firstly, data are preprocessed according to GM(1,1) modeling condition. In order to remove the noise in edge information image, which is gained by preprocessed image subtracting from prediction image. Template of 3 × 3 pixel blocks is selected as the first choice in ascending order to predict maximum pixel value with GM(1,1). According to the information of template pixels with no difference, information of neighborhood pixels is used for improving differences of 3 × 3 template pixels. Secondly, median filter is employed to eliminate the isolated point noise in edge information images. Finally, image edge information with two-dimensional discrete wavelet weighted coefficient is obtained. Simulation results show that the proposed algorithm has advantages such as precisely locating, abundant weak edge, and better anti-noise performance comparing the classical edge detection operator methods.