基于粒子群优化的Shearlet自适应图像去噪

Particle Swarm Optimization Based Adaptive Image Denoising in Shearlet Domain

研究Shearlet变换域图像去噪阈值选取的问题,提出Shearlet变换域图像去噪自适应阈值选取方法.该方法根据Shear-let变换域不同尺度和方向系数的分布特性,采用粒子群优化算法自适应地确定各尺度和方向的最优阈值,实现基于图像内容的自适应去噪.仿真实验表明,该方法能有效滤除图像的噪声,较好地保留图像的边缘信息.同时,去噪后图像具有更高的峰值信噪比(PSNR).

The threshold selection of image denoising in shearlet domain was firstly researched.And then,an adaptive threshold selection method was proposed.According to the distribution characteristics of shearlet coefficients in different scale and direction,the optimal threshold was adaptively determined by particle swarm optimization.Using the proposed optimal threshold,the noisy image can be adaptively denoised based on image content.Simulation results indicate that the method can filter noise effectively and preserve edge information of image preferably.Meanwhile,the denoised images achieve higher PSNR.