A Visual Tracking method for Mobile Robot

A novel method of color feature representation was presented. The color features based on hue-saturation-intensity color space, includes one-dimension-color-vector (ODCV) and region-mean-gray-scale (RMGS). ODCV of an object has good stability, because it is not influenced by the changes of illumination. It expresses the native color of an object. A method of moving object tracking was also presented. The method uses ODCV feature for preferential matching feature and searches the moving object in entire image frame. That trait ensures that the method has global tracking capability. The results of experiments show the method´s good ability of tracking a color rigid object with good tracking accuracy