Registration of LIDAR point clouds using image features
Conference Paper
Overview
Identity
Additional Document Info
View All
Overview
abstract
An optimal linear translation and attitude estimation (OLTAE) algorithm is proposed to register 3dimensional point clouds based on the image features associated with the individual data sets. In LIDAR applications, such images are created by projecting the point cloud data on to an image plane. Physically, this image is the return light intensity observed by the LIDAR imager that is usually available to the analyst for post processing. Associated image features are extracted from the corresponding images by utilizing the recent advances in computational vision and image processing. Features thus obtained have unique descriptors that automate the matching process and ease the solution of the so-called correspondence problem. Corresponding matched features from the images are then used as vector measurements for the 3 dimensional point cloud registration algorithm. As a byproduct the algorithm is shown to provide the uncertainties associated with the translation vector and the pose orientation estimate. The methods developed in the paper are subsequently applied to measurement sets obtained from a Light Detection and Ranging (LIDAR) system for spacecraft proximity operation emulation applications.
