Object tracking using spatio-temporal tracklet association
一种利用时空约束轨迹片段关联的目标跟踪方法*

Object tracking is a challenging problem in visual surveillance. Noise segmentation, partial and full object occlusion may result in tracking failure. In order to solve these problems, an object tracking algorithm by associating tracklets with fundamental spatio-temporal constraints is proposed. First detect moving objects, then generate tracklet, and grow these tracklets to find the best spatial and temporal association of observations. Experiments prove the proposed method can successfully track multiple moving vehicles and persons under occlusion, noisy detections and split-merge situations.