基于ORB+GMS算法的航拍图像匹配方法的研究  

Study on aerial images matching based on ORB+GMS algorithm

作　　者：雷远华 曾德培 Lei Yuanhua;Zeng Depei(China Energy Engineering Group Hunan Electric Power Design Institute Co.,Ltd.,Changsha 410007,China)

机构地区：[1]中国能源建设集团湖南省电力设计院有限公司,湖南长沙410007

出　　处：《矿山测量》2021年第2期43-46,80,共5页Mine Surveying

摘　　要：ORB算法匹配效率较高、鲁棒性较好,网格运动统计算法可保留更多内点,故结合使用ORB和网格运动统计算法并在航拍图像上验证匹配效果。经实验证明:与ORB+RANSAC算法相比,执行效率提高0.2 s左右,内点比约提高11%,内点数量大幅度增加,所提算法可应用在对效率、稳定性和内点数量要求较高的航拍场景。ORB algorithm had the higher matching efficiency and the better robustness,and the grid motion statistics algorithm could retain more internal points,so the matching effect was verified on the aerial images by using ORB and the grid motion statistics algorithm.Experimental results showed that,comparing with the ORB+RANSAC algorithm,the execution efficiency was increased by about 0.2 s,the internal point ratio was increased by about 11%,and the number of the internal points was greatly increased.Algorithm proposed in the paper could be used in the aerial photography scenarios with the high requirements for efficiency,stability and internal point number.

关 键 词：航拍图像匹配 ORB算法 网格运动统计算法 

分 类 号：P231[天文地球—摄影测量与遥感]