|
Nilani Algiriyage, Raj Prasanna, Emma E H Doyle, Kristin Stock, & David Johnston. (2020). Traffic Flow Estimation based on Deep Learning for Emergency Traffic Management using CCTV Images. In Amanda Hughes, Fiona McNeill, & Christopher W. Zobel (Eds.), ISCRAM 2020 Conference Proceedings – 17th International Conference on Information Systems for Crisis Response and Management (pp. 100–109). Blacksburg, VA (USA): Virginia Tech.
Abstract: Emergency Traffic Management (ETM) is one of the main problems in smart urban cities. This paper focuses on selecting an appropriate object detection model for identifying and counting vehicles from closed-circuit television (CCTV) images and then estimating traffic flow as the first step in a broader project. Therefore, a case is selected at one of the busiest roads in Christchurch, New Zealand. Two experiments were conducted in this research; 1) to evaluate the accuracy and speed of three famous object detection models namely faster R-CNN, mask R-CNN and YOLOv3 for the data set, 2) to estimate the traffic flow by counting the number of vehicles in each of the four classes such as car, bus, truck and motorcycle. A simple Region of Interest (ROI) heuristic algorithm is used to classify vehicle movement direction such as \quotes{left-lane} and \quotes{right-lane}. This paper presents the early results and discusses the next steps.
|
|
|
Nilani Algiriyage, Raj Prasanna, Kristin Stock, Emma Hudson-Doyle, David Johnston, Minura Punchihewa, et al. (2021). Towards Real-time Traffic Flow Estimation using YOLO and SORT from Surveillance Video Footage. In Anouck Adrot, Rob Grace, Kathleen Moore, & Christopher W. Zobel (Eds.), ISCRAM 2021 Conference Proceedings – 18th International Conference on Information Systems for Crisis Response and Management (pp. 40–48). Blacksburg, VA (USA): Virginia Tech.
Abstract: Traffic emergencies and resulting delays cause a significant impact on the economy and society. Traffic flow estimation is one of the early steps in urban planning and managing traffic infrastructure. Traditionally, traffic flow rates were commonly measured using underground inductive loops, pneumatic road tubes, and temporary manual counts. However, these approaches can not be used in large areas due to high costs, road surface degradation and implementation difficulties. Recent advancement of computer vision techniques in combination with freely available closed-circuit television (CCTV) datasets has provided opportunities for vehicle detection and classification. This study addresses the problem of estimating traffic flow using low-quality video data from a surveillance camera. Therefore, we have trained the novel YOLOv4 algorithm for five object classes (car, truck, van, bike, and bus). Also, we introduce an algorithm to count the vehicles using the SORT tracker based on movement direction such as ``northbound'' and ``southbound'' to obtain the traffic flow rates. The experimental results, for a CCTV footage in Christchurch, New Zealand shows the effectiveness of the proposed approach. In future research, we expect to train on large and more diverse datasets that cover various weather and lighting conditions.
|
|