Measurement of 3D fundamental features by fusion of range, intensity and tactile images

Proposes methods of fusing range images, intensity images and tactile images. The methods are intended for application to distributed sensing of multiple mobile robots. As the features for fusion, practically important and possible to robustly extract ones, are selected: planar regions and cylindrical regions for range images, linear edges for intensity images and planar regions and linear edges for tactile images. The fusion methods to measure a 3D edge and a cylinder are constructed and formulated by utilizing a least square approach. Additionally, fusion of these features measured by multiple mobile robots is discussed. The measurement of 3D edges and cylinders are verified by experiments.