Abstract: A Nonsubsampled Contourlet Transform(NSCT) image denoising method based on Linear decreasing Weight Particle Swarm Optimization(LinWPSO) is proposed in this paper. This method acquires the optimal threshold of Generalized Cross Validation(GCV) risk function by using LinWPSO in the NSCT domain, and removes the noise through soft threshold function, which does not need the prior information of noise variance. Experimental results show that the proposed method can more effectively reduce Gauss noise in remote sensing image and improve the Peak Signal to Noise Ratio(PSNR) of the image.

Key words: image denoising, soft threshold, Nonsubsmapled Contourlet Transform(NSCT), Particle Swarm Optimization(PSO), shift invariance, Generalized Cross Validation(GCV)

摘要： 提出一种基于线性递减权重粒子群优化(LinWPSO)阈值的非下采样Contourlet变换(NSCT)图像去噪方法。在NSCT域通过LinWPSO对广义交叉验证风险函数寻优以确定最佳阈值，通过软阈值函数去噪，利用NSCT的平移不变性抑制伪Gibbs失真效应，从而完整保留图像的纹理和边缘等细节信息。实验结果表明，该方法能有效去除遥感图像的高斯噪声，提高图像的峰值信噪比。

关键词: 图像去噪, 软阈值, 非下采样Contourlet变换, 粒子群优化, 平移不变性, 广义交叉验证

CLC Number: 

• TN911.73