矢量中值滤波及其应用

Vector median filtering and its application

矢量中值滤波方法是在中值滤波的基础上发展而来的,它具有与中值滤波几乎相同的滤波特性,即①任何单调变化的信号都将完全得到保留;②任何窄于(L-M)/2的脉冲都将被压制(这里 L 为滤波器的滑动时窗长度,M 为输出时窗长度)。与中值滤波相比,矢量中值滤波的最大特点是:在每个时窗位置处滤波后的输出是 M 个数据元素,因而运算速度是中值滤波的 M 倍。另外,矢量中值滤波是一种多参数滤波,在实际应用中具有更大的滤波选择性。文中最后讨论了矢量中值滤波在零偏移距 VSP 资料处理中的应用和效果。

Vector median filtering(VMF)is based on median filtering(MF); therefore,VMF has nearly the same properties as MF:(1)full pre- servation of monotonously varying signals and(2)suppression of the pulses that are shorter than(L-M)/2 samples,where L is moving win- dow length and M the output window length.Compared with MF, VMF has a major advantage:VMF is M times as fast as MF because VMF outputs M elements in each window.Besides,VMF is multipara- meter filtering and has more filtering options in practice.The paper fin- ally describes both the application of VMF to zero-offset VSP data processing and the effect.