针对目标在跟踪过程中受环境变化影响(光照、遮挡等)使其跟踪发生偏移的问题,提出一种从目标粗匹配到粒子群算法精确定位的等级关联结构的多目标跟踪算法。与现有跟踪算法相比,在粗匹配阶段粒子随机产生过程中融入了上下文信息,提高了目标匹配的准确度,降低了错误跟踪的目标数;对于在粒子群精确定位阶段有显著偏差的目标位置, 采用Metropolis-Hastings采样算法进行纠正,同时完成模板更新,从而保证了目标跟踪的准确性。实验结果表明,该算法在目标被遮挡的情形下能够准确地跟踪被遮挡的目标。

To cope with the drift problem of tracking caused by environmental changes (such as illumination variations and occlusions), we propose a multi-target tracking algorithm with a hierarchical associative structure, which first coarsely matches the targets and then accurately locates them using Particle Swarm Optimization (PSO). Compared with the state-of-the-art tracking algorithms, context information is integrated into the generation of the particles during the coarse matching stage in this paper, thus enhancing the accuracy of target matching as well as reducing the number of false-tracked targets. To ensure the tracking accuracy, the targets' locations with prominent deviations in the phase of accurate tracking are rectified via Metropolis-Hastings algorithm; meanwhile, the targets' templates are updated. Experimental results show that the proposed algorithm can track the occluded targets more accurately under the occlusions.