A novel method for quantification of vacant parking spaces at on-street parking lots

Abstract

Intelligence transportations in crowed urban cities will surely need the thoroughly management of parking spaces, especially when the number of vehicles increases quickly, and autonomous driving vehicles will be more popular. Monitoring and real-time updating available vacant slots will be essential. In this work, we propose a camera-based monitoring solution for on-street parking spaces which are in open area, freely entering/leaving and connected with surround landscape. Our solution does not require to install physical markers at fields, however, can assess the space of a parking vehicle required on field, and report the number of available vacant slots on the field. The accuracy of the proposed algorithm has been evaluated by the data collected from an on-street parking areas nearby the university campus. It achieves the accuracy of approximate 96%, which is slightly 2% less than the results of the marker-based method. The program is lightweight that can be deployed on edge devices attached to CCTV cameras, hence, saving the bandwidth of sending data to central systems.