点线特征约束的LiDAR点云与航空影像自动配准

Automatic registration of airborne LiDAR data and aerial images constrained by point and line features

提出了一种点线特征约束的机载激光雷达(LiDAR)点云与航空影像自动配准方法.将LiDAR点云内插生成强度图和数字表面模型(DSM),利用航空影像和强度图自动匹配得到点特征;基于现有方法从点云和影像中提取直线段特征,并依据随机采样一致性原理精化影像同名直线段匹配结果.以反投影距离为测度交会得到摄影测量直线特征,在物方空间内根据方向、长度比和距离测度实现点云影像同名直线特征匹配.最后进行点线特征约束下的航空影像联合区域网平差,解算影像内外方位元素,并采用迭代循环策略逐步精化配准结果.实验结果表明:提出的方法精度优于0.5倍的点云平均点间距,实现了城区LiDAR点云与航空影像的高精度自动配准,其结果可以满足点云影像融合相关应用的要求.

An automatic registration method of LiDAR data and aerial images constrained by point and line features was proposed. Firstly, intensity images and DSM were generated by LiDAR point clouds interpolation and point features were obtained by matching between aerial images and intensity images automatically. Then, line features were extracted from point clouds and the images, and the matching results of the conjugate image line features were refined according to the principle of random sampling consistency. The photogrammetric lines were reconstructed based on the distance between back projection and observations in images, followed by matching of the conjugate line between point clouds and images according to the direction, the ratio of length and the distance measurement in the object space. Finally, the aerial image block adjustment constrained by point and line features was performed to solve interior and exterior orientation parameters of images, and the interative loop strategy was adopted to refine results. Experimental results show that the accuracy of the proposed algorithm is better than 0.5 times of the average distance.High accuracy of aerial image registration with LiDAR piont clouds is realized, which could meet the requirements of applications related to fusion of LiDAR point clouds and aerial images.