A Fast Registration Method for MEMS LiDAR Point Cloud Based on Self-Adaptive Segmentation

The Micro-Electro-Mechanical System (MEMS) LiDAR point cloud in autonomous vehicles has a large deflection range, which results slow registration speed and poor applicability. To maximize speed, an improved Normal Distribution Transform (NDT) method that integrates density features been proposed. First, the is reduced using modified voxel filter pass-through filter. Next, Intrinsic Shape Signature (ISS) algorithm utilized to analyze extract key points; Four-Point Congruent Set (4PCS) then employed calculate initial pose under constraints of set complete coarse registration. Finally, self-adaptive segmentation model constructed by K-D tree obtain points, NDT combined with this form SSM-NDT algorithm, used for fine Each was compared on vehicle dataset PandaSet actual collected datasets. show novel increases at least 60% takes into account good accuracy strong anti-interference.