Optimized Video Tracking for Automated Vehicle Turning Movement Counts

Abstract

This paper proposes a new method for automatically counting vehicle turning movements based on video tracking, expanding on previous work on optimization of parameters for road user trajectory extraction and on automated trajectory clustering. The counting method is composed of three main steps: an automated tracker that extracts vehicle trajectories from video data, an automated trajectory clustering algorithm, and an optimization algorithm. The proposed method was applied to obtain turning movement counts in three typical traffic engineering case studies in Canada representing industry-type conditions. These exhibited varying levels of tracking difficulty, ranging from a single-lane off-ramp to a six-movement intersection with a stop and a right-turn channel. Because of a limitation of the data set, giving flows per movement and not per lane, all sites were chosen with a single lane per movement. The 3-h morning peak period was used in the case studies. The results show an average weighted generalization error of 12% for more than 3,700 vehicles automatically analyzed for more than 8 h of video, ranging from 9.5% to 19.5%. The generalization error is on average 8.6% (and as low as 6.0% per movement) for the 3,084 uninterrupted vehicles that are in plain view of the camera. This paper describes in detail the methodology used and discusses the factors that affect counting performance and how to improve counting accuracy in further research.

Real-Time Intersection Vehicle Turning Movement Counts from Live UAV Video Stream Using Multiple Object Tracking

“…To evaluate the effectiveness of the proposed framework, a case study at a busy intersection in Hong Kong has been conducted. The overall accuracy is 91.93%, which outperforms similar approaches proposed in Bélisle et al (2017) and Bui et al (2020). Compared with Bélisle et al (2017), our proposed method utilized a more advanced deep-learning-based tracking model and the extracted trajectories are less noisy.…”

“…The overall accuracy is 91.93%, which outperforms similar approaches proposed in Bélisle et al (2017) and Bui et al (2020). Compared with Bélisle et al (2017), our proposed method utilized a more advanced deep-learning-based tracking model and the extracted trajectories are less noisy. With respect to Bui et al (2020), we employed an evolved tracker, StrongSORT, rather than DeepSORT, and trained the tracking model with a more appropriate dataset, VisDrone, rather than the default COCO dataset.…”

“…Recent literature on video image processing for TMC data collection is extensive while a majority of them focus on deeplearning-based Computer Vision (CV) technologies. In an early study, a three-step framework for TMC data collection was proposed in Bélisle et al (2017): an automated tracker to extract trajectories, an automated trajectory clustering algorithm, and an optimization algorithm. With the following development of Multiple Object Tracking (MOT) technology, Bui et al (2020) proposed a TMC data collection framework using CCTV systems and utilized the COCO dataset for YOLO training.…”

VIFECO: An Open-Source Software for Counting Features on a Video

“…Indeed, videos recorded by digital action cameras (DACs) can enhance analyses based on innovative methodologies such as mobile methods [5,6], participatory visual research [7][8][9], or non-participatory observation studies [10,11]. They are also many applications of DACs in transportation studies [12][13][14][15], environmental sciences [16,17], and organizational research [18,19].…”