An improved contourlet transform image denoising method: 7th international conference on modelling, identification and control

In order to remove the noise of the traction machine and the steel wire rope effectively and protect the details of the image better, the image denoising method based on an improved Contourlet are proposed, which considers the non-independence among Contourlet coefficients fully in denoising process. The traditional image denoising methods of Contourlet coefficients are obtained by Contourlet transform, the sub-band coefficients are convergence to eliminate the noise in the Sub band. In order to avoid the traditional neighborhood convergence method threshold, which is too high to erase many coefficients. This paper, the universal threshold and convergence factor is improved by the partial parameters in the neighborhood window. It makes more coefficients be corrected by using improved the universal threshold and new convergence factor. More coefficients is built instead of missing. The Experiments show that the algorithm is improved in the denoising performance and the visual effects. Meamwhile, the image details are protected better.