This paper presents a variant of the Iterative Closest Point (ICP) algorithm for merging multiple color point clouds generated from a mobile 3D Light Detection and Ranging (LIDAR) System. This algorithm uses hue information generated from a camera along with the coordinates of the scan points and enables high accuracy registration of point clouds. A k-d tree based nearest neighbor search associates corresponding colored points in 4-D space between data and model point clouds. Singular Value Decomposition (SVD) method solves for the rigid rotation and translation. Experimental results illustrate that 3D color point clouds accelerate the 3D map registration if the hue data and model point clouds have sufficient hue distribution and the imaging sensor robustly captures the hue.

This content is only available via PDF.
You do not currently have access to this content.