D. Di Paola, C. Marotta, A. Milella, G. Circirelli, and A. Distante (Italy)
Surveillance robotics, object detection, computer vision, laser data segmentation
This paper presents a novel multimodal approach for indoor environment monitoring by a mobile robot. The proposed system consists of a vision-based and a laser based module that match independently the current scene with a stored one, looking for new or removed objects. The matching process, in both modules, uses clustering on local data to detect point clouds. A fuzzy logic inference system is then implemented to evaluate environment changes and decide whether a significant variation of the scene has occurred. Results from experimental tests demonstrate the feasibility of the proposed method in robot surveillance applications.
Important Links:
Go Back
