The Scenario Import And Export Estimation Based On Multiple Target Tracking

Because of the complexity of the background environment, cross-cutting and multi-target occlusion between changes in attitude and shape of the moving object, light changes, efficient and robust multi-target tracking algorithm study is still difficult. In this thesis, the discrete continuous optimization for multiple target tracking algorithm is improved, an improved discrete continuous optimization for multiple target tracking algorithm is proposed. On the basis of target tracking, getting the sample near the image edge, so as to estimate the number and position of import and export, our work is shown as below:1. About the target tracking:For the common multiple target tracking problem in complex scenarios, an improved discrete continuous optimization for multiple target tracking algorithm is proposed. By adding new energy function, ameliorating the original energy function, and adding global optimized clustering process, our method effectively solve the tracking loss, tracking division, trajectory merger, and target identity conversion problem. The experimental results show that our method has higher tracking accuracy than traditional multiple target tracking algorithm in complex scenes.2. About the scenario import and export estimation:We use the K-means clustering algorithm to cluster the position of target sample near to the image edge. The experimental results show that, our method can estimate the scenario import and export.