秩排序非线性滤波器的自适应优化算法  

Optimization for rank-order nonlinear filters

作　　者：宋焕生[1] 赵祥模[1] 王养利[2] 

机构地区：[1]长安大学,陕西西安710064 [2]西安电子科技大学,陕西西安710071

出　　处：《电路与系统学报》2006年第4期31-34,38,共5页Journal of Circuits and Systems

基　　金：国家自然科学基金资助项目(60272050)

摘　　要：在图像处理中,秩排序非线性滤波器具有非常重要的作用。本文从加权顺序统计滤波器的自适应优化算法出发,导出了加权中值滤波器、秩选择滤波器和置换滤波器等几种重要的秩排序非线性滤波器的自适应优化设计算法,这些算法采用迭代计算实现,并可以在MAE、MSE和Lγ准则下得到最优解。模拟实验表明,这些算法具有非常好的收敛效果,并可以很好地用于类似图像的非平稳信号的处理。The rank-order nonlinear filters play very important roles in image processing. Based on the weighted order statistic filter optimal algorithm, the adaptive optimal algorithms for several important nonlinear rank-order filter categories have been proposed. The filter categories include weighted median, rank selection and permutation filters. The proposed algorithms can implement by simple iterative calculations, and can reach the optimal solutions under MAE, MSE and L^r criterions. The experiments show that proposed algorithms have good convergent performance, and can be used in the image processing applications.

关 键 词：非线性滤波 滤波器设计 自适应优化 优化算法 

分 类 号：TN919.8[电子电信—通信与信息系统]