Multi-Target Vehicle Tracking Algorithm Based on Improved DeepSORT
Open Access
Abstract
In this paper, we address the issues of insufficient accuracy and frequent identity switching in the multi-target tracking algorithm DeepSORT by proposing two improvement strategies. First, we optimize the appearance feature extraction process by training a lightweight appearance extraction network (OSNet) on a vehicle re-identification dataset. This makes the appearance features better suited for the vehicle tracking model required in our paper. Second, we improve the metric of motion features by using the original IOU distance metric or GIOU metrics. The optimized tracking algorithm using GIOU achieves effective improvements in tracking precision and accuracy. The experimental results show that the improved vehicle tracking models MOTA and IDF1 are enhanced by 4.6% and 5.9%, respectively. This allows for the stable tracking of vehicles and reduces the occurrence of identity switching phenomenon to a certain extent.Keywords
Funding Information
- Autonomous Region Key Research and Development Program Project (2022B01015, XJJTZKX-FWCG-202401-0045, GKJ2023XTWL04)
- Technology research and development project of Xinjiang Communications Investment (Group) Co., Ltd. (2022B01015, XJJTZKX-FWCG-202401-0045, GKJ2023XTWL04)
- Ganquanbao Economic DevelopmentZone Science and Technology Program Project (2022B01015, XJJTZKX-FWCG-202401-0045, GKJ2023XTWL04)
This publication has 27 references indexed in Scilit:
- Short-term traffic flow prediction based on hybrid decomposition optimization and deep extreme learning machinePhysica A: Statistical Mechanics and its Applications, 2024
- PNAS-MOT: Multi-Modal Object Tracking With Pareto Neural Architecture SearchIEEE Robotics and Automation Letters, 2024
- Performance evaluation for multi-target tracking with temporal dimension specificsChinese Journal of Aeronautics, 2023
- Focal and efficient IOU loss for accurate bounding box regressionNeurocomputing, 2022
- Real-time multiple object tracking using deep learning methodsNeural Computing and Applications, 2021
- An Edge Traffic Flow Detection Scheme Based on Deep Learning in an Intelligent Transportation SystemIEEE Transactions on Intelligent Transportation Systems, 2020
- Unsupervised Deep Representation Learning for Real-Time TrackingInternational Journal of Computer Vision, 2020
- A moving vehicle tracking algorithm based on deep learningJournal of Ambient Intelligence and Humanized Computing, 2020
- A labeled random finite set online multi-object tracker for video dataPattern Recognition, 2019
- Online multi-object tracking combining optical flow and compressive tracking in Markov decision processJournal of Visual Communication and Image Representation, 2019