基于极线单应性的目标空间定位算法  

Target Spatial Location Algorithm Based on Epipolar Homography

作　　者：杨利斌[1] 陈滨[1] 王毅[1] YANG Libin;CHEN Bin;WANG Yi(Weapon Science and Technology Department,NAAU,Yantai 264001)

机构地区：[1]海军航空工程学院兵器科学与技术系,烟台264001

出　　处：《计算机与数字工程》2018年第10期1989-1994,2028,共7页Computer & Digital Engineering

摘　　要：基于双目立体视觉系统对目标进行空间定位的过程中,若在摄像设备获取的图像上提取的特征点对存在较大的定位误差,则得到的匹配点对无法满足极线约束,从而导致空间定位算法的失败。针对上述问题,论文提出了基于极线单应性的目标空间定位算法,将空间定位算法转化为优化问题。通过设定代价函数,求解最优的匹配点对,最后通过构造齐次方程求解得到目标的空间坐标。实验采用搭载差分GPS的无人机平台作为目标,对算法的有效性进行验证。实验结果表明,在各种系统参数下,论文提出的算法相比经典的中点式定位算法和鲁棒的极大似然算法均有更高的定位精度。In the process of target spatial location based on binocular stereo vision, the existing error of locating feature points in the image captured by cameras may lead to dissatisfaction to the epipolar constraint and failure of spatial location algorithm. In order to solve the above problem, this paper proposes an algorithm of target spatial location based on epipolar homography, which transforms the spatial location algorithm into optimization problem. By setting the cost function, the optimal matching point pairs are obtained. Finally, the spatial coordinates of the target are obtained by constructing and solving the homogeneous equation. The un- manned aerial vehicle （UAV） platform with differential GPS is used as the target in the experiments to verify the effectiveness of the algorithm. Experimental results show that the proposed algorithm has higher localization accuracy than the classical midpoint algorithm and robust maximum likelihood algorithm under various system parameters.

关 键 词：双目立体视觉 目标空间定位 极线单应性 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]