Trajectory analysis for parking lot vacancy detection system

Abstract

Taking a vacant parking lot in crowded metropolitan areas is a major problem especially with the huge number of vehicles. In this study, the authors provide a vision-based system for real-time management of parking spaces in the case of outdoor parking. Different real-world challenges may face these systems such as weather conditions, luminance variation, perspective distortion and inter-spaces occlusion. In this study, they propose a decisional module based on a tracking approach that determines in real time the state of the parking lots and localises the vacant parking spaces according to several extracted attributes. This decision is based on local characteristics of the parking spaces as well as the performed vehicles' events.