Using color histograms and range data to track trajectories of moving people from a mobile robot platform

People detection and tracking is still an active research field for autonomous mobile robots. Approaches that tackle the problem on mobile robots differ from those proposed for surveillance and other similar tasks. Mainly due to the motion of sensors and hence the complexity of scene interpretation, approaches proposed for fixed cameras are not appropriate. In this paper it is presented a methodology for people detection and tracking from an autonomous mobile robot. People detection and tracking initialization is made by image processing techniques, while for tracking data provided a range data camera (Kinect like sensor) and color information from the former sensor are used to track people over 3D space. Accuracy of tracker is achieved by an adaptive color histograms method based on color from RGB camera and projected over the plain view.