Automatic registration for long range laser data in the stop-and-go mode

Abstract

In this work we are focusing of the registration of Long Range scans taken in the SAG mode, we find an automatic registration algorithm based on mapping 3D scans to 2D images. After mapping the scans to images, we use SURF matching algorithm for automatic matching points detection. Rough registration parameters can be estimated via Least Squares method based on the tie points in the two scans. To refine these registration parameters, the Iterative Closest Point (ICP) approach is used with the scans. The overall method of automation of registration is efficient in terms of speed, and accuracy, especially with many scans.