Multi-object tracking in traffic environments: A systematic literature review
Diego M. Jiménez-Bravo, Álvaro Lozano Murciego, André Sales Mendes, Héctor Sánchez San Blas, Javier Bajo
Abstract
The use of computer vision techniques to detect objects in images has grown in recent years. These techniques are especially useful to automatically extract and analyze information from an image or a sequence of them. One of the problems addressed by computer vision is multi-object tracking over frames sequences. To know the path and direction of objects can be crucial for some areas like traffic control and supervision; by doing that the system can be able to reduce traffic jams or redirect vehicles over less condensed areas. These algorithms include several aspects to have in mind in order to start a new development or research in this area, for instance, is important to review the current state-of-the art techniques, the hardware requirements, the main evaluation metrics, the commonly used datasets, among others. Therefore, the objective of this research is to present a systematic literature review which analyzes the recent works developed in the area of multi-object tracking in traffic environments. This paper reviews the techniques, hardware, datasets, metrics, and open lines of research in this area.