基于粒子群优化的Contourlet域图像去噪

IMAGE DENOISING BASED ON PARTICLE SWARM OPTIMIZATION IN CONTOURLET DOMAIN

本文研究了Contourlet变换域图像阈值去噪问题,提出了一种在Contourlet变换域改进的阈值去噪方法。该方法结合Contourlet变换能更加突出图像的边缘信息和方向信息的优势,在Contourlet变换域利用粒子群优化算法(PSO)迭代寻优找到最优阈值,以峰值信噪比(PSNR)的函数模型作为PSO的适应度函数,并且利用硬阈值函数规则对图像处理。实验结果表明,本文优化阈值的选取不仅更有利于保留图像的细节信息,还使得图像有着明显的去噪效果,PSNR值得到提高。

This paper studies the threshold problems for image denoising in Contourlet domain, and put forward an improved threshold denoising method . Combined with the advantage that Contourlet transform can protect image edge information and direction informa-tionpreferably, this paper uses particle swarm optimization PSO for iteration optimization to find the optimal threshold in Contourlet do-main, takes the function model peak signal-to-noise ratio PSNR as fitness function of PSO, and uses hard threshold function rules for image processing. The experimental results show that the selection of optimal threshold not only is more advantageous to keep the details of image, also makes the image has obvious denoising effect, and peak signal-to-noise ratio PSNR is improved.