Multi-Traffic Targets Tracking Based on an Improved Structural Sparse Representation with Spatial-Temporal Constraint

• 作者: YANG Honghong SHANG Junchao LI Jingjing ZHANG Yumei WU Xiaojun
• 来源期刊: 电子学报（英文版）
• 年，卷（期）: 2022,（2）
• 所属期刊栏目: SIGNAL PROCESSING
• 页码范围: 266-276
• 页数: 11页
• 语种: 英文
• DOI: 10.1049/cje.2020.00.007

摘要

Vehicles or pedestrians tracking is an im-portant task in intelligent transportation system.In this paper,we propose an online multi-object tracking for in-telligent traffic platform that employs improved sparse representation and structural constraint.We first build the spatial-temporal constraint via the geometric rela-tions and appearance of tracked objects,then we con-struct a robust appearance model by incorporating the discriminative sparse representation with weight con-straint and local sparse appearance with occlusion analys-is.Finally,we complete data association by using maxim-um a posteriori in a Bayesian framework in the pursuit for the optimal detection estimation.Experimental res-ults in two challenging vehicle tracking benchmark data-sets show that the proposed method has a good tracking performance.