| 引用本文: | 万琴,朱晓林,肖岳平,孙健,王耀南,颜金娥,杨佳玉.采用RGB–D时空上下文模型的多目标遮挡跟踪算法[J].控制理论与应用,2021,38(12):2019~2030.[点击复制] |
| WAN Qin,ZHU Xiao-lin,XIAO Yue-ping,SUN Jian,WANG Yao-nan,YAN Jin-e,YANG Jia-yu.Multi-target occlusion tracking algorithm employing RGB–D spatio-temporal context model[J].Control Theory and Technology,2021,38(12):2019~2030.[点击复制] |
|
| 采用RGB–D时空上下文模型的多目标遮挡跟踪算法 |
| Multi-target occlusion tracking algorithm employing RGB–D spatio-temporal context model |
| 摘要点击 1349 全文点击 354 投稿时间：2020-10-22 修订日期：2021-11-09 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/CTA.2021.00734 |
| 2021,38(12):2019-2030 |
| 中文关键词 RGB–D 时空上下文 遮挡跟踪 时间一致性 最大后验概率(MAP) |
| 英文关键词 RGB–D spatio-temporal context occlusion tracking temporal consistency maximum a posteriori (MAP) |
| 基金项目 国家自然科学基金青年项目(62006075), 湖南省杰出青年科学基金项目(2021JJ10002), 湖南省重点研发计划项目(2021GK2024), 湖南省自然科学 基金面上项目(2020JJ4246)资助. |
|
| 中文摘要 |
| 为了提高实时RGB–D目标遮挡跟踪精确度, 解决多目标遮挡跟踪容易发生模型漂移和跟踪丢失等问题, 本
文提出一种基于RGB–D时空上下文模型的多目标遮挡跟踪算法. 首先获取多目标检测定位区域, 再通过目标时空
上下文特征提取, 建立目标时间上下文模型、目标空间上下文模型构成目标RGB–D时空上下文模型; 然后在跟踪器
判别跟踪状态时通过计算时间一致性进行颜色和深度特征自适应融合确定目标在当前帧位置; 最后, 当跟踪器判
别多目标遮挡时引入深度概率, 利用深度概率信息特征进行约束, 通过最大后验概率(MAP)关联模型有效解决目标
遮挡跟踪问题. 在公用数据集clothing store dataset和princeton tracking benchmark dataset上进行定性对比实验和定量
结果分析表明, 本文提出的算法具有良好的遮挡跟踪性能, 能较好解决多目标遮挡跟踪问题, 提高目标遮挡跟踪的
精确性和鲁棒性. |
| 英文摘要 |
| In order to improve the accuracy of real-time RGB–D target occlusion tracking and solve the problems of
model drift and tracking loss in multi-target occlusion tracking, this paper proposes a multi-target occlusion tracking algorithm
based on RGB–D spatio-temporal context model. Firstly, the multi-target detection and location region is obtained,
and then the target temporal context model and the target spatial context model are used to establish the target RGB–D
spatio-temporal context model through target spatio-temporal context feature extraction. Then, when the tracker judges
the tracking state, the color and depth features are adaptively fused by calculating the time consistency to determine the
target position in the current frame. Finally, when the tracker discriminates multi-target occlusion, the depth probability is
introduced, and the depth probability information feature is used to constrain. The maximum a posteriori (MAP) correlation
model is used to effectively solve the problem of target occlusion tracking. The qualitative comparison experiments and
quantitative results on public datasets clothing store dataset and Princeton tracking benchmark dataset show that the proposed
algorithm has good occlusion tracking performance, can better solve the problem of multi-target occlusion tracking,
and improve the accuracy and robustness of target occlusion tracking. |
|
|
|
|
|