摘要
提出了一种新的在高密度噪声环境下自适应性更好的滤波方法——基于双极值判决的高密度噪声滤除算法。该方法可以有效去除图像中的脉冲噪声,能够可靠地检测出受强噪声污染的点目标,对面目标滤波时具有更好的细节保护能力。针对该算法中的阈值问题,提出了一种极值均方差计算与统计假设检验相结合的阈值选取法,克服了主观设定阈值的缺陷,提高了整个算法的客观适应性。仿真结果表明,该算法去噪性能优良,满足实时性和鲁棒性要求。
In order to keep the useful signals as much as possible when the noises are removed, a new better adaptive filtering algorithm an infrared image high-density noise filtering algorithm based on two-pass crest value judgment is presented. This method can filter ＂salt and pepper＂ noises efficiently and pick out the point spatial targets degraded by high density noise reliably. Aiming at the threshold selection, an efficient method composed with crest value mean square deviation and statistical hypothesis testing is given, it can overcome the limitation of subjective threshold and enhance arithmetic adaptability. The results show that this algorithm has perfect performance in eliminating noise and better real-time ability and robust operability.
出处
《红外技术》
CSCD
北大核心
2008年第3期168-172,共5页
Infrared Technology
基金
"十一五"预研资助项目
关键词
双极值判决
自适应滤波
阈值选取
Double Extrema Value Judgement （DEVJ）
adapted filter
threshold selecting
