AUTOMATIC REGISTRATION OF GROUND-BASED LIDAR DATA SETS BY USING 3D LINE FEATURES

摘要

In this paper the authors show how to extract the 3D line features from ground-based LiDAR data sets and develop the algorithms in which 3D line feature matching for registering overlapping scenes can be established by means of geometric constraints in an automatic course. The proposed working scheme consists of three major kernels including 3D line feature extraction, 3D line feature matching and the registration of ground-based LiDAR data sets. 3D line features are extracted by means of image process techniques and then input into the matching engine where geometric constraints are imposed and conjugated line features among stations are picked. The registration of consecutive overlapping data sets is performed through the two-step model of 3D spatial similarity transformation. The whole processing chains in this study are featured in a highly automatic fashion. The conducted experiments show that the proposed method seems to be the alternative or the complementary for LiDAR data sets registration apart from the point, surface and other features that are often realized as feature primitives for LiDAR data.