摘要
针对机载LiDAR(light detection and ranging)点云与航空影像的数据结构差异导致的配准难题,提出一种在二维平面基于相位信息的点云与航空影像配准方法。首先将点云投影至二维平面,在特征点提取与匹配阶段,利用FAST(features from accelerated segment test)在点云投影图与航空影像上提取特征点;然后采用Log-Gabor滤波器与影像卷积生成响应能量标签图,在此之上构造特征点描述符;最后采用最近邻算法进行匹配并设计了一种误匹配剔除流程进行误匹配剔除,由此建立点云激光点与航空影像像素之间的对应关系。为验证本文方法有效性,选取了3个场景的点云数据与航空影像进行实验,并与PSO-SIFT(position scale ori⁃entation-scale invariant feature transform)、RIFT(radiationvariation insensitive feature transform)、HAPCG(histogran of absolute phase consistency gradients)算法进行对比。结果表明,该方法的综合匹配性能在该数据中具有较高的鲁棒性与自动化水平。
Aiming at the registration problem caused by the difference of data structure between airborne light detection and ranging(LiDAR)point cloud and aerial image,this paper proposes a point cloud and aerial image registration method based on phase information in two-dimensional plane.In the stage of feature point extraction and matching,Features from accelerated segment test(FAST)is used to extract feature points from the point cloud projection map and aerial image.Then,the response energy label map is generated by using log-Gabor filter convolution with image,and the feature point descriptors are constructed on this map.Finally,the nearest neighbor algorithm is used for matching and a mismatching elimination process is designed to eliminate the mismatching.From this,the correspondence between the starting point cloud laser points and the aerial image pixels can be estab⁃lished.In order to verify the effectiveness of the proposed method,point cloud data and aerial images of three scenes are selected for experiments,and compared with position scale orientation-scale invariant feature transform(PSO-SIFT),ra⁃diation-variation insensitive feature transform(RIFT)and his⁃togran of absolute phase consistency gradients(HAPCG)algo⁃rithms.The results show that the comprehensive matching performance of the proposed method has high robustness and automation level in this data.
作者
李明
范大昭
储光涵
宋佳璇
李东子
LI Ming;FAN Dazhao;CHU Guanghan;SONG Jiaxuan;LI Dongzi(Information Engineering University,Zhengzhou 450001,China)
出处
《测绘地理信息》
CSCD
2024年第3期74-79,共6页
Journal of Geomatics
基金
国家自然科学基金(41971427)
高分遥感测绘应用示范系统(二期)(42-Y30B04-9001-19/21)
国防173基金(2021-JCJQ-JJ-1253)。
