论文部分内容阅读
人们一直在利用有效信号和噪声的某种特性差异 ,进行着对污染信号的恢复 ,这一过程依赖于对信号和噪声特性的精确描述 ,子波理论为这种描述提供了新的工具。文献 [1 ]的子波域降噪算法通过将相邻尺度的相关量 Corr2 (m,n)作归一化处理后直接与 W(m,n)进行比较决定 W(m,n)的取舍 ,并以噪声在各尺度上的方差作为终止迭代的准则 ,但是这样做使得在相邻尺度之间进行相关时会丢失一部分信号所产生的系数 ,从而不能使信号得到更好的重构。文中从大尺度的模极大值出发 ,在小尺度上的相应位置邻域内搜索对应于此大尺度模极大值位置 ,以改变信号边缘点的抽取 ,仿真结果表明 ,新算法在保留信号边缘的同时 ,能很好的抑制噪声 ,滤波效果较原算法有一定的提高。
People have been making use of the difference between the effective signal and the noise to carry out the recovery of the pollution signal. The process relies on the accurate description of the signal and noise characteristics. The wavelet theory provides a new tool for this description. The wavelet domain denoising algorithm in [1] determines the choice of W (m, n) by normalizing the correlation of adjacent scales Corr2 (m, n) directly with W (m, n) , And the variance of noise at each scale is used as the criterion to terminate the iteration. However, in doing so, coefficients generated by a part of signals will be lost when correlation between adjacent scales, so that the signal can not be reconstructed better. In this paper, starting from large-scale modulus maxima, the searching in the neighborhood of the corresponding position on the small scale corresponds to the location of the large-scale modulus maxima to change the extraction of signal edge points. The simulation results show that the new algorithm preserves the signal edge At the same time, can well suppress the noise, the filtering effect than the original algorithm has a certain increase.