Image Edge Detection Based on Complex Ridgelet Transform

Abstract

For the defects of the existing edge detection methods those can’t well guarantee the robust anti-noise character and the continuity of edges detected simultaneously, a new kind of image edge detection method is proposed which is based on the complex ridgelet transform and Laplace operator. First, the scalar one-dimensional wavelet transform in the traditional ridgelet transform is replaced by dual-tree complex wavelet transform, so that the ridgelet transform has translation invariance, and then the image is denoised by the complex ridgelet transform. Second, the Laplace operator convolutes the coefficients of the complex ridgelet transform so as to tell whether the points are edge points. The experiments show that this method has a better robust anti-noise character, and could reserve edge information in the image better at the same time of image de-noising, which could extract the object profile efficiently.