为消除彩色图像中的脉冲噪声,提出一种新的基于等级阈值的矢量中值滤波器。该滤波器设计了一组随滤波窗口内当前像素排序位置而等级变化的阈值,并根据当前像素与经典矢量滤波器输出的距离差值来判断噪声像素的存在。仿真实验证明,该方法较其他非开关型和开关型矢量中值滤波器能更好保存原图像细节和消除脉冲噪声。
A new switching vector median filter using ranked thresholds is proposed to remove the impulsive noise from color images. The new method utilizes thresholds related rank position of the central pixel in filtering window and the distance between the output of the classic vector median filter and the original vector to determine the noise probability. Compared with other non-switching and switching vector median filters, simulation experiments show that the new filter outperforms them in terms of both noise suppression and detail preservation.