An improved camshift algorithm based on RGB histogram equalization

Classical Camshift algorithm (CCA) has been widely used in the fields of visual tracking, image smoothing and segmentation currently. In view of the fact that CCA is easily affected by the object which is similar to the tracking target in hue, an improved Camshift algorithm is presented in this paper. In this algorithm, the histogram equalization is applied simultaneously to R, G, B channels in order to extend the hue differentiation between pixels. The experiment results show that the proposed algorithm performs well when tracking target has the similar hue with the other object and provides higher tracking accuracy.