Image denoising method based on primal sketch correction and matrix filling

作者：Mengjuan Lu, Binghui Yang · 年份：2022 · DOI：10.1109/icbar58199.2022.00010 · 研究领域：Image and Signal Denoising Methods、Advanced Image Fusion Techniques、Sparse and Compressive Sensing Techniques

This paper only aims to develop a SAR denoising algorithm based on Primal Sketch classification and SVD domain to improve MMSE estimation. The main contents of this paper are as follows: 1) In Primal Sketch algorithm, dual-domain contrast enhancement method is used to improve the energy image. 2) SAR images are divided into edge class and non-edge class by Primal Sketch algorithm. Then the NLSVD decomposition of the two kinds of pixels is carried out, and the singular value matrix is estimated by the minimum mean square error criterion containing the contracted silver. The estimated value of the edge class and the non-edge class are obtained by the inverse transformation. 3) The edge coefficient is calculated, and Butterworth fusion method is used to fuse the edge class and non-edge class boundary to get the denoising result. Our results show that this method can effectively remove speckled noise in SAR images, and keep the edge and point target information well.