The research on the mean shift algorithm for target tracking

3rd International Conference on Advances in Energy, Environment and Chemical Engineering

能源学;生态环境科学;化学工业

Cao, Honghong^1

Software Engineering College, Chongqing University of Posts and Telecommunications, Chongqing

400065, China^1

关键词: Affine transformations;    Comparing experiments;    Local optima;    Mean shift algorithm;    Moving objects;    Particle swarm optimization algorithm;    Scale invariant feature transforms;    Tracking process;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/69/1/012167/pdf
DOI  :  10.1088/1755-1315/69/1/012167

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

The traditional mean shift algorithm for target tracking is effective and high real-time, but there still are some shortcomings. The traditional mean shift algorithm is easy to fall into local optimum in the tracking process, the effectiveness of the method is weak when the object is moving fast. And the size of the tracking window never changes, the method will fail when the size of the moving object changes, as a result, we come up with a new method. We use particle swarm optimization algorithm to optimize the mean shift algorithm for target tracking, Meanwhile, SIFT (scale-invariant feature transform) and affine transformation make the size of tracking window adaptive. At last, we evaluate the method by comparing experiments. Experimental result indicates that the proposed method can effectively track the object and the size of the tracking window changes.