An improve moving target tracking algorithm based on CBWH

Aiming at classic Mean Shift based on tracking algorithm can not track target when the speed changed, the paper proposed an improved algorithm based on corrected background weighted histogram(CBWH), the algorithm used two iterations on target, first time used mean shift to iterate and calculated the similarity coefficient, second iteration used twice the current frame center minus the previous frame center as the center and calculated the similarity coefficient, then comparing the two coefficients and choosing the bigger as the center to track the target. Based on different scenarios of the experiment show that the algorithm is compared with the original algorithm has better robustness and accuracy.