Conference Proceedings

Abstract

Nowadays, intelligent parking spot availability management is a must for smart cities, because nobody enjoys wasting time in traffic looking for a parking space. In this paper we propose a method to identify the free parking spots in a large parking lot by analyzing the information from the surveillance cameras. The classification result is used in a mobile and/or web application for guide the users to a specific free parking spot. The proposed solution is based on SVM (Support Vector Machines) for classification in two classes free/occupied and quantized DCT (Discrete Cosine Transform) for spots description. The experimental results show very good performance for real time parking management, with a rate over 92% of accuracy. The tests are done on four different large parking areas.