AUTHOR=Lin Haojia , Yuan Zhilu , He Biao , Kuai Xi , Li Xiaoming , Guo Renzhong 

TITLE=A Deep Learning Framework for Video-Based Vehicle Counting

JOURNAL=Frontiers in Physics

VOLUME=Volume 10 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2022.829734

DOI=10.3389/fphy.2022.829734

ISSN=2296-424X

ABSTRACT=Traffic surveillance can be used to monitor and collect the traffic condition data of road network, which plays an important role in a wide range of applications in Intelligent Transportation Systems (ITS). Detecting and counting vehicles in traffic videos accurately and rapidly is one of the key problems. Traditional video-based vehicle detection methods, such as background subtraction, frame difference, and optical flow, have some limitations in accuracy or efficiency. In this paper, deep learning is applied for vehicle counting in traffic videos. First, to solve the problem of the lack of annotated data, a method of vehicle detection based on transfer learning is proposed. Then, on the basis of vehicle detection, a vehicle counting method based on fusing virtual detection area and vehicle tracking is proposed. Finally, for the possible situation of vehicle missing detection and false detection, a missing alarm suppression module and a false alarm suppression module are designed to improve the accuracy of vehicle counting. The results show that the proposed deep learning-based model can achieve lane-level vehicle counting rapidly without enough annotated data, and the accuracy of vehicle counting can reach up to 99%. In terms of computational efficiency, this method has high real-time performance and can be used for real-time vehicle counting.