面向山地区域光照变化下的鲁棒景象匹配方法  

A robust scene matching method for mountainous regions with illumination variation

作　　者：王华夏[1,2] 程咏梅[1] 刘楠[1] WANG Huaxia CHENG Yongmei LIU Nan(College of Automation, Northwestern Polytechnical University, Xi＇an 710072, China College of Electronic and Information Engineering, Taiyuan University of Science and Technology, Taiyuan 030024, China)

机构地区：[1]西北工业大学自动化学院,西安710072 [2]太原科技大学电子信息学院,太原030024

出　　处：《航空学报》2017年第10期188-200,共13页Acta Aeronautica et Astronautica Sinica

基　　金：国家自然科学基金(61603364);西安市科技计划项目(CXY1436(9);CXY1350(2))~~

摘　　要：提出了一种面向山地区域光照变化下的鲁棒景象匹配导航方法。该方法针对山区景象匹配导航中光照环境不同造成的基准图与实时图像不一致问题,采取在高程数据上使用光照模型生成光照明暗图,利用数字高程图(DEM)得到水流汇集数据并生成山谷显著图,将光照明暗图与山谷显著图融合作为基准图;针对基准图与实时图中细节边缘的差异带来的误匹配问题,提出了基于形态学约束的Hausdorff距离边缘匹配算法。采用LANDSAT图像与ASTERDEM高程数据进行实验分析,结果表明提出的方法匹配正确率高且鲁棒性好。This paper presents a robust scene matching navigation method for mountainous regions with illumination varia- tion. In order to solve the problem of the inconsistency between the reference map and the real map caused by different lighting conditions in mountainous areas, the reference image is obtained by fusing the hillshade map, generated from ele- vation data using illumination model, and the valley saliency map, generated from water flow data collected by Digital Ele- vation Model （DEM）. To avoid the mismatch caused by the difference of edge between the reference and the real maps, a Hausdorff distance edge matching algorithm based on morphological constraints is proposed. The experimental results of LANDSAT image and ASTERDEM data show that the proposed method has high accuracy and good robustness.

关 键 词：景象匹配导航 光照模型 显著边缘 HAUSDORFF距离 图像融合 

分 类 号：V249.32[航空宇航科学与技术—飞行器设计]