engleski

ITS Supported Parking Lot Management

In recent decades the number of vehicles in urban areas has rapidly increased causing problems such as traffic congestion and prolonged parking space search time. To overcome the problem of parking place search, large ground/underground parking lots have been built in city centres and other areas with high demand for parking space like office buildings, shopping malls, airports, etc. Driver information systems based on technologies developed within the intelligent transportation systems (ITS) have been implemented to inform the drivers about the number of free parking spaces and their distribution between different parking lots or between different levels in a particular parking lot. Such ITS based information systems use different sensors to detect car presence on a parking space like magnetic, infrared or ultrasonic sensors and belong to classical parking lot management systems. These sensors must be mounted on every single parking space and require significant installation and maintenance effort. Many parking lots also have installed cameras that cover most or all parking places used for surveillance purpose. To reduce parking lot management cost, installed cameras are being more and more used for parking lot management tasks like incoming and outgoing car counting, parking place occupancy detection, incident detection, parking lot security, and parking payment via license plate recognition. Such camera based systems have several advantages: high reliability (comparable to infrared sensors), one camera can cover a relatively large parking lot area, relatively small numbers of cameras are needed to cover the entire parking lot (number of classical sensors is equal to number of parking spaces), intelligent cameras with integrated embedded computer vision systems perform vehicle detection locally thus enabling easy integration into existing parking lot management systems, video footage can be used for additional purposes like person counting and recognition, surveillance, security, etc. This paper gives a review of existing computer vision based parking lot management systems. Implementation possibilities of additional ITS functionality (driver informing, driver navigation to a free parking space, parking lot usage statistics) are being analyzed. Problems of such systems are being investigated and possible solutions are evaluated. Comparison of vision system complexity to classical management systems is also given.

ITS ; parking lot management ; computer vision ; driver navigation system

This work has been supported by the University of Zagreb Development Fund and by the University of Zagreb Faculty of traffic and transport science.