RGB-D与MeanShift相结合的实时目标跟踪

来源期刊:中南大学学报(自然科学版)2019年第9期

论文作者:王广龙 田杰 朱文杰 方丹

文章页码:2163 - 2171

关键词:目标跟踪;MeanShift;RGB-D;实时跟踪

Key words:object tracking; MeanShift; RGB-D; real-time tracking

摘    要:针对MeanShift跟踪算法目标易受到周围相似背景区域干扰的问题,引入图像的RGB-D特征,研究并实现结合RGB-D与MeanShift的实时目标跟踪算法。首先,提出一种新颖的相似性度量方法优化二次型距离(advanced quadratic-form distance,AQFD);其次,基于颜色和深度特征完善了特征权重在线自适应调整机制和目标模型更新策略;最后,对该算法进行仿真分析并基于Jetson TX2平台实现。研究结果表明:本文算法在目标受到周围相似背景区域干扰情况下仍能够完成目标跟踪任务,优于先进算法Staple,具有良好的准确性与鲁棒性,并且能够应用于实时跟踪。

Abstract: Arm to solve the problem that the object is susceptible to interference from similar surrounding background in MeanShift, RGB-D feature was introduced, and thus a real-time object tracking combined RGB-D with MeanShift was proposed. Firstly, a novel similarity measurement method, advanced quadratic-form distance (AQFD),was put forward. Secondly, the online self-adaptive adjustment mechanism of feature weights and updating strategy of object model were improved. Finally, the method was simulated based on Matlab and implemented on Jetson TX2. The results show that the proposed method which is better than Staple performs well in accuracy and robustness in the case that the object is disturbed by similar surrounding background, and it can be applied to real-time tracking applications.

相关论文

  • 暂无!

相关知识点

  • 暂无!

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号