论文部分内容阅读
提出一种新型的递归中值滤波器，抛掉了统计参数的制约，将滤波算法转化为一种优化处理。该方法兼顾了滤波处理的光滑连续性及抑制噪声的累积特性，可有效地消除脉冲型干扰的影响，同时也从理论的角度上对该算法进行了分析。为消除加性高斯噪声，提出了一种基于图像边缘方向的小波线性滤波器，它仅仅处理边缘信息。该算法的极大优点是克服了边缘模糊效应，小波的去噪逆向重构的处理方法对边缘为阶跃型的层析图像非常实用。
Proposed a new type of recursive median filter, throw away the constraints of statistical parameters, the filtering algorithm into an optimization process. The method takes into account both the smoothness of the filtering process and the suppression of noise accumulation, which can effectively eliminate the influence of impulsive interference and at the same time analyzes the algorithm theoretically. In order to eliminate additive Gaussian noise, a wavelet linear filter based on the edge direction of the image is proposed, which only deals with the edge information. The great advantage of this algorithm is to overcome the edge blur effect. The wavelet de-noising and inverse reconstruction method is very practical for the edge-based tomographic image.