Study on the Basic Principle and Image Denoising Realization Method of Curvelet Transform

Abstract

This paper studied the basic principle and realization method of Curvelet transform on the basis of analysis Ridgelet transform. Wavelet transform is more applicable in processing point singular signal, however, it is obviously insufficient in the analysis of the image edge feature of the curve of two-dimensional image or a straight, while the Curvelet transform has more directional and has better approximation and sparisity description of the smooth region and edge parts. This paper first introduces the concept and the implementation algorithm of the curvelet transform, then applies Curvelet transform to denoiseing of the standard image and polymer materials image and compared with the denoising effect of wavelet transform. In the experiments, the Wrap (Wrapping-based transform) algorithm was used to realize the Curvelet transform, which adds a wrap step to the USFFT method to one to one map arbitrary region to the origin affine regions through the cycle technology. The results of the experiments show that Curvelet transform has a better denoising result and a certain increase in PSNR.