基于GCV理论的小波自适应多阈值图像去噪方法

小波图像去噪是小波应用较成功的一个领域,而其中最重要的一个环节就是最优阈值的确定问题。提出了一种基于小波变换的多阈值图像去噪的改进方法。这种方法是在各分解尺度上的各子带选择不同的最佳阈值,而最佳阈值的选取是基于GCV理论。利用此理论不仅能获得最小均方意义上的渐进最优解,而且不需要知道噪声的先验知识。通过实验证明,这种方法不仅能获得很好的去噪效果,而且由于不需要对噪声进行估计从而减小了计算量。

Adaptive Wavelet Multi - thresholding Based on GCV Theory for Image Denoising

Image denoising based on the wavelet is a successful field of wavelet application.The most important step of wavelet denoising is how to select an optimal threshold.In this paper,the GCV theory is used to achieve the optimal threshold on every direction of each level.It has been proved that the GCV theory can achieve an approximate optimal resolute.Besides,the prior knowledge about noise distribution is unnecessary with this theory.Experiments show that this method is effective to image de noising.