Multi-modal point cloud data registration technology in complex scenarios
Due to the difficulty in obtaining multi-modal point cloud data in complex environments, the requirements for point cloud data registration and 3D model construction accuracy are becoming increasingly high, this article takes the real-life 3D modeling of Nantong Grand Theater as an example. When the position difference between the initial point cloud and the calibration point cloud is significant, using the ICP algorithm for point cloud registration can easily lead to local optimization problems. The proposed nearest point iteration (CPA-ICP) algorithm based on control point auxiliary constraints is used to register point cloud data, and the experimental comparison of the other three point cloud registration algorithms fully demonstrates, This method has high registration accuracy and efficiency, and has good reference significance for multimodal point cloud data fusion in complex scenes.
complex scenariosmulti-modal point cloudjoint directional matchingCPA-ICP algorithmdata fusion
万方数据
维普
