A target tracking algorithm for vision based sea cucumber capture

Abstract

For capturing sea cucumber in the sea, an object tracking method based on Mean-Shift algorithm was proposed. Firstly, defogging algorithm was used to get restored image, and the feature of color histogram was extracted from the restored image. Then the local image was cropped at the same position and local area in the second frame. Mean-Shift algorithm was used to search the feature of color histogram in cropped area. Finally, tracking was realized by searching the searching process converged in the next frames one by one. Compared with the color histogram Mean-Shift algorithm (CHMS) algorithm, the proposed algorithm have better tracking performance and robustness.