摘要
传统的基于噪声检测的中值滤波算法,对于检测出的噪声点往往采用相同的滤波处理方法,而忽视了噪声点自身特性的差异,限制了其滤波性能的进一步提高。针对这一问题,提出了一种区域自适应中值滤波器,算法在噪声检测的基础上,应用改进的区域判别方法实现对噪声点所属区域的判别,进而根据其所属区域的不同有针对性的选取滤波处理方法。仿真研究表明,该滤波器算法在有效抑制噪声的同时,获得了更强的边缘细节保护能力,较之传统方法具有更为出色的滤波性能。
In the result of adopting unitary filtering algorithm no matter the discrepancies between the regions that the noise pixels locate in, the conventional decision-based median filters run short of flexibility and their performances are grievously constrained. A region-adaptive median filter was therefore proposed instead of the conventional one, which determines the regions in which the pixels locate firstly and then switches the filtering algorithm for each pixel according to the peculiarity of the region it belongs to, so as to protect efficiently the details of edges with the noise suppressed simultaneously. Simulation tests verifiy the effectiveness of the improved region-adaptive median filter for edge detection.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2007年第23期5411-5414,共4页
Journal of System Simulation
基金
国家自然科学基金资助项目(60374003)
国家重点基础研究发展规划项目(2002CB312201)
流程工业综合自动化教育部重点实验室开放课题基金资助项目(PAL200511)
关键词
区域自适应
中值滤波
噪声检测
边缘保护
region adaptive
median filter
noise detecting
detail preserving
