Confidence Estimation of ToF and Stereo Data for 3D Data Fusion. Stima della Confidenza delle Misure Ottenute da Sensori ToF e da Sistemi Stereo per la Fusione di Dati 3D

This thesis focuses on the analysis of ToF and stereo vision systems, with the goal of extracting reliable confidence measures associated to the computed depth maps.The two families of sensors are described and after an analysis on practical issues some confidence measures are provided.Then, a framework for 3D data fusion with confidence information is presented and evaluated over a dataset of real world data.Results show that the proposed approach outperforms the performance of the two systems