室内遮挡多人跟踪

Tracking Multi-People under Indoor Occlusion

提出一种在固定摄像机室内遮挡条件下的多人跟踪算法,包括基于改进的独立成分分析模型的运动前景分割、数据关联与合并检测和基于贪心搜索的遮挡目标定位三部分。将方法应用于室内多目标遮挡,结果表明该算法对发生遮挡的目标数量、目标被遮挡程度以及目标运动模式均无约束。

Proposes multi-people tracking algorithm for a fixed camera under indoor occlusion,including the segmentation of movement foreground based on the improved independent component analysis model,based on data association of the merger-split and based on occlusion targeting of the greedy search.The method is applied to indoor occlusion of multi-objective.The result shows that the algorithm is no constraints about how many objects occluded,how much degree occluded,and target motion mode.