多特征融合匹配的霍夫森林多目标跟踪

针对目标遮挡、形变等复杂环境中多目标跟踪准确性低的问题，提出了一种多特征融合匹配的霍夫森林多目标跟踪算法．首先，该算法根据目标检测响应进行初步关联，在线选取正负样本，通过融合颜色直方图、方向梯度直方图特征以及光流信息构建目标的特征模型;然后利用霍夫森林学习，形成可靠的长轨迹;最后采用多特征融合的轨迹匹配算法，引入颜色直方图的相似性度量和基于Gabor滤波器的特征点匹配两种方式，形成加权融合的概率矩阵，将长轨迹逐级关联为目标的完整轨迹．实验表明，该算法在多个复杂环境的视频序列中，可以有效解决目标形变、相互遮挡等问题，能实现多目标的鲁棒性跟踪．

Multi-target tracking algorithm based on the multi-feature fusion matching and Hough forest

In order to solve the problem of low accuracy due to target occlusion and deformation in multi-target tracking, this paper proposes a multi-target tracking algorithm based on the multi-feature fusion matching and Hough forest. First, we select positive and negative samples online according to primary association among detection responses and construct the feature model of the target with color histogram, histogram of oriented gradient (HOG) and optical flow information. Then, longer trajectory associations are generated based on the online learned Hough forest framework. Finally, a trajectory matching algorithm based on multi-feature fusion is proposed, and we introduce two methods of similarity measure in color histogram and feature points matching based on the Gabor filter to generate the probability matrix with the weighted factor. Therefore, it can further form the complete trajectories of the targets by associating them gradually. Experimental results show that the proposed algorithm can effectively solve the problems of target deformation and mutual occlusion in the video sequences of complex environments, and realize the robust tracking of multiple targets.