CAJ | 학술논문

为了有效地保留视频去噪后的纹理细节，利用Surfacelet变换特性和自适应阈值的方法，提出了一种新的视频去噪算法。 Surfacelet变换是超小波分析领域的最新发展，具有良好的多维信号处理能力，真正实现了对多维信号的多方向多尺度分解。本文介绍了Surfacelet变换的结构，根据Surfacelet变换的视频信号的系数特性和噪声分布特征，自适应估计每个方向子带的系数阈值。实验表明，本文算法能更好保留视频细节信息，对于视频的噪声也具有更强的抑制能力。
위료유효지보류시빈거조후적문리세절，이용Surfacelet변환특성화자괄응역치적방법，제출료일충신적시빈거조산법。 Surfacelet변환시초소파분석영역적최신발전，구유량호적다유신호처리능력，진정실현료대다유신호적다방향다척도분해。본문개소료Surfacelet변환적결구，근거Surfacelet변환적시빈신호적계수특성화조성분포특정，자괄응고계매개방향자대적계수역치。실험표명，본문산법능경호보류시빈세절신식，대우시빈적조성야구유경강적억제능력。
In order to effectively retain details and texture while de-noising videos,a news method for video denoising was developed which brings the strong points of surfacelet transform and adaptive threshold.Surfacelet transform is the new development of hyper wavelet domain,it is good at multidimensional signal processing and it provides a true directional and multiresolution decomposition of high dimensional signals.In this paper,the construction of surfacelet is introduced.According to the correlation ofneighborhood coefficients and the distribution characteristics of noise in surfacelet transform domain, the adaptive threshold is accurately calculated corresponding toeach directional subband coefficient. The experimental results show that the method can better retain video details and these noises in video can bemore effectively removed.