Coarse automated registration of visual imagery with three-dimensional light detection and ranging models
Prakash Duraisamy, Kamesh R. Namuduri, Stephen C. Jackson, Bill P. Buckles · Journal of Electronic Imaging · 2013
An automated, fast, robust registration of visual images with three-dimensional data generated from light detection and ranging (LiDAR) sensor is described. The method is intended to provide a coarse registration of the images, suitable for input into other algorithms or where only a rough registration is required. Our approach consists of two steps. In the first step, two-dimensional lines are extracted as features from both the visual and the LiDAR data. In the second step, a novel and efficient search is performed to recover the external camera parameters, and thus the camera matrix (as a calibrated camera is assumed). An error metric involving line matchings is used to guide the search. We demonstrate the performance of our algorithm on both synthetic and real-world data.