基于三维滤波的视频去噪算法研究

On algorithm for video denoising based on 3D filtering

视频去噪的目的是将原始视频从观测到的含噪视频中还原出来。对基于三维滤波的视频去噪算法进行了研究。首先利用贝叶斯阈值对视频序列的各帧在小波域中滤波,之后对帧间连续三帧图像进行帧间滤波。仿真结果表明该算法的有效性。

The aim of video image denoising is to recover the original uncorrupted video images from noisy ones. Video denoising based on 3D filtering is studied, which consists of spatially filtering in the wavelet domain and temporal filtering in the signal domain. Spatial filtering is performed by adaptive Bayesian shrinkage for each video frame in the wavelet domain. Temporal filtering is performed on the three successive interframe frames subsequently, which is on the basis of motion detector and recursive time-averaging. Experimental results show that the proposed method has better denoising effect.