Fusion of Stereo Vision and RFID for Site Safety

摘要

Increasing safety in construction zones and enhancing security in airports are important problems that can be solved using sensor fusion and activity analysis. Identifying objects solely via computer vision (CV) is computationally costly, error prone, limited by occlusion, and sometimes impossible in practice. Radio Frequency Identification (RFID) can reliably identify tagged objects and can even localize targets at coarse spatial resolution. We provide a model based on discrete relaxation for fusing RFID and CV over both space and time to address the trajectory points correspondence problem in 3D space. We report physical outdoor experiments using commodity cameras and RFID system. Pilot results show that RFID information available at ambiguous instants in tracking reduce tracking runtime by 99%. The likelihood of producing correct object trajectories in regions partially or fully occluded to CV is also increased.