Developing visual sensing strategies through next best view planning
IROS'09 Proceedings of the 2009 IEEE/RSJ international conference on Intelligent robots and systems
| Hi-index | 0.00 |
We present a systematic view planning method to assist construction of 3-D models of large outdoor sites using a mobile robot platform with mounted scanner. In the first stage, we begin with a 2-D site footprint and the planner generates a minimal set of sufficient covering views. These views which incorporate constraints on the scanner, including field of view, minimum and maximum scanning distance, and grazing angle, serve as the initial set of scans which yields an approximate 3-D model of the site. In the second stage, we update this model by using a voxel-based occupancy procedure to plan and acquire the next best view. The algorithm continues to update the model sequentially until an accurate and complete 3-D model is obtained. Results are shown for a segment of the Columbia University campus. The system can also be used as a planning tool for manual construction of 3-D site models.