Quantitative evaluations of stable and adaptive tracking using the updating HS-histogram method

In this paper, we proposed a robust detecting and tracking method for mobile robot by using the Updating Hue and Saturation-histogram method (UHS-his), we also proposed standard quantitatively evaluation parameters for a general detecting and tracking method. Hue and Saturation are elements of HSV color space. The UHS-his method can be used indoors and outdoors as the robust detecting and tracking method of the person following robot. To use the UHS-his method as the robust detecting and tracking method for a person following robot, the method is also required to be stable for tracking and adaptive for fast moving. Therefore, the standard quantitative evaluation parameters of stable and adaptive tracking are necessary. There are no standard evaluation methods for detecting and tracking by a color camera. We proposed two parameters for the standard quantitative evaluation of detecting and tracking by an image processing. One is "Tracking Error" to evaluate the stability of the general detecting and tracking method. The other is "Limit Speed" to evaluate the adaptivity of the method. The UHS-his method is compared with the existing tracking method by using the proposed standard quantitative evaluation parameters. The results of evaluations show that Tracking Error of the UHS-his method is up to 1/5 of the existing tracking method and Limit Speed of the UHS-his method is up to 10 times fast of the existing tracking method. The experimental results suggest that the UHS-his method is stable for tracking and adaptive for fast moving.