基于负相关的可信赖模式匹配

Negative Correlation Based on Reliable Pattern Registration

在模式匹配计算中正相关往往被人们所关注,负相关量却常常被忽略。在实际的模式匹配算法研究中我们却发现,正相关在强调信号的同时,负相关也能同时对信号的优劣进行评价,对正相关匹配结果达到互补的作用。本文从负相关对噪声抑制的作用入手讨论负相关在可信赖模式匹配中的重要价值。本文给出的实验结果显示,负相关能够有效抑制正相关匹配过程中产生的噪声,实现高精度、高可信赖模式匹配推定。

In the correlation calculations,people only pay attention to positive correlation;negative correlation is often overlooked.During the actual research work on pattern registration algorithm,we find that positive correlation strength-ens signal,meanwhile negative correlation can also be used to evaluate the signal.Negative correlation is a complemen-tary role for positive correlation.This paper discusses the value of the negative correlation in reliable patter registration.The experimental results in this paper show that negative correlation can suppress the noises generated in positive corre-lation processing.The high precision and reliable pattern registration has been realized.