一种基于三维小波变换的视频图像消噪新方法

根据视频图像帧与帧之间在多数情况下存在较强相关性的特点，提出了一种基于三维小波变换的视频图像消噪方法。将被加性平稳高斯噪声污染的连续几帧视频图像进行帧间一维小波变换，根据变换后各帧的特点．采用不同的方法分别进行消噪；对变换后的低频帧进行帧内二维小波变换，在小波域中采用自适应阈值进行消噪；对变换后的其余各高频帧直接采用固定阈值进行消噪。实验结果表明该方法消噪效果较好。

New method of video image denoising based on 3_D wavelet transform

According to the great relativity between frames of video image in most cases, this paper proposes a method for video image denoising on the basis of 3_D wavelet transform by performing multiscale decomposition of the several continuing frames corrupted by additive stationary guassian noise using interframe 1_D wavelet transform. Then, different methods are used for denoising according to the characteristics of the transformed frames. For the low frequency frame, an adaptive threshold is used for denoising in the wavelet domain based on the multiscale decomposition of frame using intraframe 2_D wavelet transform; for the high frequency ones, an universal threshold is used for denoising directly. The results demonstrate that we better effect by using the proposed method can be achieved.