An Airborne LiDAR Data Registration Method Based on Combining Multiple Information

Aiming at the problems of uneven resolution, sparse irregularity, low overlap rate and a large number of redundant ground points in airborne LiDAR point cloud data, this paper proposes an intensity-constrained registration algorithm with an initial estimation algorithm based on ground segmentation combining motion information. First, the rotation transformation relationship of adjacent frames can be calculated by the normal vector of ground point clouds segmented by an improved RANSAC (Random Sample Consensus) algorithm. Combining the rotation parameters with the translation information obtained by UAV platform provides good initial value estimation for registration, on the basis of which an intensity-constrained Generalized-ICP registration algorithm is proposed. This method makes full use of the intensity information of LiDAR to constrain the corresponding point set participating in the iteration to achieve the precise registration. The experimental results show that the proposed algorithm is able to obtain the accurate transformation relationship between laser point cloud data and has higher registration accuracy and efficiency.