[1]王希鹏,李永,李智,等.融合图像深度的抗遮挡目标跟踪算法[J].郑州大学学报(工学版),2021,42(05):19-24.
点击复制

融合图像深度的抗遮挡目标跟踪算法()
分享到:

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
42
期数:
2021年05期
页码:
19-24
栏目:
出版日期:
2021-09-10

文章信息/Info

Title:
Anti-occlusion target tracking algorithm ba<x>sed on image depth—ncig2020
作者:
王希鹏李永李智张妍
文献标志码:
A
摘要:
由于视频信息的局限性,在遮挡情况下的目标跟踪依然是一个很难解决的问题。针对目标跟踪过程中的遮挡问题,提出将图像深度引入单目标跟踪。首先应用单目图像深度估计算法对图像进行深度估计,获取图像的深度信息;其次,将基于孪生网络的目标跟踪算法与图像深度相结合,构建遮挡判别模块,利用目标深度信息的变化判断遮挡情况;最后,对目标跟踪器的候选框重新排序避免被遮挡物干扰。实验结果表明,该算法能有效地应对遮挡情况对跟踪性能的影响,在跟踪成功率和精确度上均高于其他对比算法。
Abstract:
Due to the limitation of video information, target tracking in the case of occlusion is still a difficult problem to solve. Aiming at the problem of occlusion in the target tracking process, it is proposed to introduce image depth into single target tracking. Firstly, the monocular image depth estimation algorithm is used to estimate the depth of the image to obtain the depth information of the image secondly, the target tracking algorithm ba<x>sed on the siamese network is combined with the image depth to construct an occlusion discriminating module, which uses the change of the target depth information to determine the occlusion situation finally, the candidate fr<x>ames of the target tracker are reordered to avoid interference by obstructions. Experimental results show that the algorithm can effectively deal with the influence of occlusion on tracking performance, and it is better than other comparison algorithms in tracking success rate and accuracy.
更新日期/Last Update: 2021-10-11