惯性组合导航系统中基于BRISK的快速景象匹配算法  

Rapid scene matching algorithm based on BRISK for inertial integrated navigation system

作　　者：许允喜[1,2] 蒋云良[1] 陈方[1] 

机构地区：[1]湖州师范学院信息与工程学院,浙江湖州313000 [2]浙江大学信息与电子工程系,浙江杭州310027

出　　处：《光电子．激光》2012年第8期1589-1596,共8页Journal of Optoelectronics·Laser

基　　金：国家自然科学基金(60872057);浙江省自然科学基金(R1090244;Y1101237;Y1110944);浙江省公益技术应用研究项目(2011C23132);湖州市自然科学基金(2011YZ07)资助项目

摘　　要：针对景象匹配/惯性组合导航系统对景象匹配算法精确性、实时性和鲁棒性的要求,提出惯性组合导航用基于二进制鲁棒不变尺度关键点(BRISK)特征描述符的高精度快速景象匹配算法。首先,结合惯性导航的误差特性设置BRISK特征检测子及描述子参数,提取参考图和实测图的BRISK特征和描述符;然后,根据最近邻汉明距离准则获取匹配的BRISK特征点对;最后,利用RANSAC方法和最小二乘算法(LSA),求得位置和航向参数。实验结果表明,本算法具备亚pixel级精度、ms级实时性以及优越的抗噪声、抗光照变化和抗旋转尺度变换能力。Considering the requirements of accuracy, real-timeness, robustness in scene matching/ inertial integrated navigation system,the scene matching algorithm {or inertial integrated navigation with high precision and high-speed based on binary robust invariant scalable keypoints （BRISK） is proposed. First, we set parameters of the BRISK detector and descriptor combined with the error characteristics of inertial navigation,extract BRISK features and calculate BRISK descriptors of the reference image and the real image. Then, BRISK descriptors are matched based on the nearest neighbor Hamming distance crite- rion. Finally, we adopt RANSAC and the least square algorithm （LSA） for getting aircraft position and course deviation. Experimental results show that the proposed algorithm has wonderful performance with precision at sub-pixel level, real-timeness at millisecond level and the excellent ability of resistance to changes in illumination, noise, rotation and scale transformation.

关 键 词：惯性导航 景象匹配 二进制鲁棒不变尺度关键点(BRISK) 局部特征描述符 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]