Searching And Tracking Algorithm Of The Aeria Target Based On Background Modeling

Considering characteristics of aerial background constrast, this paper proposes a accurate and quick algorithm of search and tracking for aerial target based on background modeling. Firstly, according to the characteristics of aerial background, this paper selects Gaussian mixture background modeling to model the background and proposes a fast improved algorithm of Gaussian mixture background modeling for the problems of the traditional Gaussian mixture model. This algorithm combines the characteristics between the single Gaussian model and Gaussian mixture model, and improves the algorithm in terms of establishing model, updating the parameter and deleting invalid Gaussian distribution. Meanwhile in order to improve accuracy of the aerial target search, it combines the background modeling, three-frame difference method and connected domain analysis method to make the algorithm search aerial targets accurately based on a quickly modeling background. Secondly, tracking algorithm on the basis of searching target extracts the multi-feature of Blob information to track based on the Kalman filter tracking framework, and this algorithm improves the robustness of the aerial target tracking. Finally, the experimental results show that the proposed algorithm can track the aerial target accurately, steadily and quickly, which has a very important practical significance in modern defense technology.