基于直方图均衡化的景象匹配导航方法研究  

Research on Scene Matching Navigation Based on Histogram Equalization

作　　者：王涛[1] 岳亚洲[1] 

机构地区：[1]西安飞行自动控制研究所,西安710065

出　　处：《电光与控制》2011年第7期32-35,共4页Electronics Optics & Control

基　　金：航空科学基金(20080818004);九七三项目基金资助(2010CD731806)

摘　　要：针对基准图与实时图之间灰度分布差异对匹配结果的影响,将直方图调整与灰度相关算法相结合,提出了一种基于直方图均衡化预处理的景象匹配导航方法。首先对基准图与实时图的直方图进行调整,减少二者的灰度分布差异,然后采用积相关算法计算匹配结果。分析了景象匹配主要验证方法的优缺点,提出了一种实用的验证方法。利用该验证方法,在各种干扰条件下进行了大样本采样验证,并且采用相对突出因子对匹配结果进行评估。大量试验结果表明基于直方图均衡化预处理的景象匹配导航方法可以有效地改善相关曲面的形态,突出相关曲面的尖锐程度,提高匹配概率。According to effect of the gray distribution difference between the reference image and real-time image on matching probability,a Scene Matching Navigation（SMN） method based on histogram equalization was proposed by combining histogram modification with gray-level correlation.First,histograms were adjusted to reduce the difference of gray distribution between the reference image and real-time image.Then cross correlation was used to obtain the result of matching.The main methods for checking the scene matching were analyzed,and a practical testing method was proposed.By using the testing method,simulations were made under various disturbance conditions to verify the effect of histogram equalization and to evaluate the matching results.The simulation results showed that the histogram equalization based SMN can improve the pattern of correlation surface,sharp the peak of correlation surface,and increase the matching probability.

关 键 词：自主导航 景象匹配 直方图均衡化 匹配概率 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] TP751.1[自动化与计算机技术—检测技术与自动化装置]