Evaluation and Application of Roadside LiDAR-based and Vision-based Multi-model Trajectory Data
Loading...
Authors
Guan, Fei
Issue Date
2022
Type
Thesis
Language
Keywords
Alternative Title
Abstract
Trajectory data has attracted more and more interest in the transportation industry because of the spatiotemporal information it contains. Early-stage trajectory data has the limitation of low frequency and small detection penetration. Recently, a new type of multi-model all-traffic trajectory data has become a new trend, which contains the high-frequency trajectories (10Hz) of all road users (vehicles, drivers, pedestrians, bicyclists, etc.) within the detection range. This kind of trajectory data can be better used for microscopic traffic analysis due to its higher accuracy, higher frequency, and full detection penetration, such as investigating driving behavior, detecting unsafe conflicts, and assessing risk levels for specific road sections/intersections, etc. Currently, two popular roadside sensors for collecting multi-model all-traffic trajectory data are LiDAR and Camera (computer vision-based). This paper evaluates trajectory data generated from these two sensors in the same intersection under the same time period in terms of detection range, trajectory length, volume counting, and speed measuring in different lighting conditions and provides a summary of the advantages and disadvantages of the types of data, which could be a reference for researchers, engineers, and other trajectory data users to use when considering which sensor to pick. This paper also presents a wildlife crossing behavior study based on trajectory data.