摘要
由于传统匹配方法主要基于高斯线性空间进行,所得细节与噪声程度相似,不能有效保留滑坡的边缘信息,因此影响滑坡定位的准确性和时效性。针对上述问题,设计了一种基于AKAZE特征和二进制稳健不变特征(BRISK)算法的无人机滑坡遥感影像的快速匹配算法,克服了传统特征提取与匹配方法的不足。利用AKAZE算法进行关键点探测,然后使用快速最近邻搜索库(FLANN)对关键点进行匹配,最后利用Ratio方法和单应性估计法剔除误匹配,提高关键点匹配质量。通过无人机滑坡遥感影像匹配试验,结果表明:同等条件下新算法耗时仅为传统算法的12%~48%,匹配率较传统算法增加了10%~40%,不仅能够准确地完成同名点的匹配,还提高了运行效率,为滑坡灾害的有效管理以及相关部门进行应急救援提供一定的技术支撑。
Traditional matching methods are mainly based on Gaussian linear space.The details obtained are similar to the level of noise,and the edge information can't be effectively retained,thus affect the accuracy and timeliness of efficiency.In view of the above problems,a fast matching algorithm for UAV landslide remote images based on AKAZE features and BRISK algorithm is designed,which overcomes the shortcomings of traditional matching methods,using AKAZE algorithm for key point detection.Then we use FLANN to match key points,finally use the Ratio method and the homography estimation method to eliminate mismatches and improve the quality of key point matching.Results of the remote sensing image matching test of UAV landslide show that under the same conditions,the new algorithm takes only 12%~48%of the time costed by traditional algorithm,and the matching rate is increased by 10%~40%compared with the traditional algorithm.It can not only accurately match the same keypoints,but also improve the operating efficiency.The propsed approach provides certain technical supports for the effective management of landslide disasters and emergency rescue by relevant departments.
作者
郝豪杰
刘贤赵
李朝奎
方军
HAO Haojie;LOU Xianzhao;LI Chaokui;FANG Jun(National-Local Joint Engineering Laboratory of Geospatial Information Technology,Hunan University of Science and Technology,Xiangtan 411201,China;Hunan Province Key Laboratory of Geo-Information in Surveying,Mapping and Remote Sensing,Hunan University of Science and Technology,Xiangtan 411201,China;College of Resources Environment and Safety Engineering,Hunan University of Science and Technology,Xiangtan 411201,China)
出处
《地理信息世界》
2020年第4期83-89,共7页
Geomatics World
基金
省重点研发计划(2018GK2015)资助。
