基于稀疏表示模型目标跟踪算法-物理电子学专业论文.docx

万方数据 万方数据 Abstract In recent years, object tracking technology has obtained extensive and in-depth research and development in the domestic and overseas, and it also has broad application prospects in many areas. Currently, object tracking is still facing many difficulties and challenges, including the object shape changes, background interference, illumination changes, occlusion, etc. At the same time, the tracking process needs to meet the requirements of accuracy and real-time. Generally speaking, a typical video tracking system in the particle filter framework consists of the following components: object initialization, the input video image, the appearance model, the dynamic model, the observation model, object localization and the template updating model. Object tracking algorithm proposed in this paper are built in the particle filter framework, the main work is as follows: A tracking algorithm based on improved sparse representation prototypes is studied. The advantage of this model is the ability to quickly perform iterative solution to improve the tracking speed. Based on this, it improves the dynamic model of the object through introducing several state parameters to achieve an effective tracking in the case of object size change and rotation change. It describes the components of this algorithm in detail, including dynamic model, particle structure, observation model, template update. Experiments verify the effectiveness of the algorithm. A tracking algorithm based on incremental 2D subspace learning is studied. This algorithm directly extracts the two-dimensional image of the object, and then puts the image into the training set. It obtains the mean, the eigenvalues, the row-projected eigenvectors and the column-projected eigenvectors of training set through incremental 2D-PCA algorithm. Experiments verify the effectiveness of the algorithm, but in the situation of severe occlusion, this algorithm has a serious flaw. Two tracking algorithms based on spa