Research on Improved ICP for 3D Reconstruction of Lightning Arrester

The lightning arrester is an electrical device used to prevent damage to the insulation of electrical equipment caused by over voltages, including lightning and switching impulses. Creepage distance, umbrella distance, and diameter are crucial parameters for lightning arrester protection. However, traditional manual detection methods are laborious and cumbersome. Moreover, due to the highly symmetrical structure of the lighting arrester itself, the traditional methods have difficulties such as easy reverse registration, etc. This paper proposes an approach for obtaining multi-angle point cloud maps by using robotic arm and depth camera, and a quadratic point cloud registration algorithm is presented, it incorporates both the FGR (fast global registration) registration and improved ICP (Iterative Closest Point) registration algorithms. Specifically, the FGR registration algorithm is initially employed for the approximate registration of two point-cloud, and then The KICP (ICP uses KDTREE to accelerate the ICP algorithm) is used for fine registration. The experimental results demonstrate that the average registration time is 60% faster than traditional method with high registration accuracy. Therefore, this method has good Real time performance and robustness for reconstructing lightning arrester.