抄録
Tracking a target robustly by vision is very difficult for mobile robot running on irregular terrain in natural environment, because the image deformation caused by rolling and pitching of the camera, as well as relative movement between the target and the camera, affect the tracking ability greatly in this task. One approach to cope with such problems is matching the target image with many affine transformed candidate images while tracking. But when the number of candidate images get larger, such approach is not available to real-time task due to the computational cost. In this paper we propose a new Robustness Analysis for Tracking (RAT) to improve the tracking ability. RAT is the analysis based on feature of the object image, where three parameters: ‘Detectability’, ‘Robustness for Depth (RBD)’ and ‘Robustness for Rotation (RBR)’ are defined. Much more robust templates can be found by analyzing the object image using RAT before the tracking task is performed. The experimental results will be shown to verify the effectiveness of this method.
