Image Denoising Method for Unknown Noise Based on 2-D FWT with Optimal Fractional Order

The optimal fractional order should be determined for image denoising by 2-D fractional wavelet transform (FWT). However, the actual application environment is complex, and the input image has already been polluted by unknown noise frequently in the process of capture and transmission. It is impossible to get the optimal fractional order on the basis of the objective evaluation standard in existence. Therefore, in view of the unknown image noise, a method to get the estimated value of optimal fractional order is put forward. Firstly, new objective evaluation standards for image denoising in fractional wavelet domain are defined, and its optimal value is obtained based on noise estimation. Then the optimal estimated fractional order is gained. The experiment results show that, 2-D FWT with the optimal fractional order can be selected reasonably and the unknown image noise can be filtered effectively in the estimated optimal fractional wavelet domain.