Target tracking algorithm based on improved image hash and multiple search areas

Moving target tracking is one of the important research topics in the field of computer vision. Aiming at the problem of tracking failure caused by illumination intensity change and similar background interference, a target tracking algorithm based on improved image hash and multi search area is proposed. Firstly, the image hash algorithm is improved, and the image hash algorithms based on cyclic comparison and threshold segmentation are designed, which are applied to track the target to obtain two hash sequences; Secondly, multiple search areas with different sizes are taken near the target position in the previous frame, and two candidate frames most similar to the target are selected by comparing the Hamming distance in the search area; Finally, the final target frame is obtained by fusing the two candidate frames, and whether to update the target is judged according to the threshold. The experimental results show that the algorithm can track the target stably and accurately under the illumination change and background interference, and the tracking success rate is 95.3%.