Feature curve-based registration for airborne LiDAR bathymetry point clouds

Abstract

The airborne LiDAR bathymetry (ALB) system is widely used in the fields of sea–land measurement, including seafloor topographic feature description, 3D seafloor model construction, coral reef monitoring, and underwater archaeology. A manned ALB data registration method based on feature curves was proposed considering problems such as sparse ALB data features, low point cloud density, and difficult corresponding feature extraction. First, triangulation network interpolation was used to extract the isoline points of the seafloor point cloud. The curves describing the seafloor topographic trend were generated through cubic parabolic spline function interpolation. Second, the curve deformation energy function was constructed based on the curve features, and a similarity measurement was carried out on the registration curve by integrating the energy function with the longest common subsequence (LCSS) algorithm. The RANSAC and ICP algorithms were used for rough and fine registration, respectively. The effectiveness and robustness of the proposed method were evaluated using two sets of experimental data. The experimental results showed that the average distance between corresponding points in the coarse registration stage reached 0.128 m and 0.136 m. After fine registration, the average point distance between the point cloud coordinates and ground truth reached 0.073 m and 0.267 m, providing an effective and reliable solution for ALB data registration.