Improved 3D Point Clouds Data Registration Based on Second-order Oscillation Artificial Bee Colony Algorithm

The traditional Iterative Closest Point (ICP) algorithm is easy to fall into local optimum because of point cloud registration, and swarm intelligence optimization algorithm can effectively solve this problem. However, this will also give rise to a large amount of calculation and low search efficiency. The second-order oscillation artificial bee colony algorithm, feeding a uniformly sampled point cloud data, and extracting a simplified point cloud data based on the domain’s Intrinsic Shape Signature (ISS) radius constraint, was proposed in the 3D point cloud data registration. Simultaneously, the spatial transformation matrix parameters were then acquired through the improved second-order oscillation artificial bee colony algorithm to complete a better initial 3D point cloud data registration. Finally, the nearest neighbor search method (k-Dimension tree, k-d tree) was applied to accelerate the corresponding point search to increase ICP fine registration efficiency. The verification results show that this paper’s algorithm has superior noise resistance, higher registration accuracy, and strong robustness than the traditional registration method through the registration experiment of the point cloud library model and scene data at different initial positions.