Abstract—In large parking areas such as those at mega shopping malls or stadiums, drivers always have difficulty to find vacant car- park lots especially during peak periods or when the parking lots are almost full. A solution to reduce the drivers’ searching time for vacant car-park lots will greatly save time, reduce cost and improve the traffic flow in the car park areas. In this paper, a research project which was developed to acquire car-park occupancy information using integrated approach of image processing algorithms is presented. Motivation for developing this system came from the fact that minimum cost is involved because image processing technique is used rather than sensor-based techniques. Security surveillance cameras which are readily available in most car parks can be used to acquire the images of the car park. This solution is much cost effective than installing sensor on each parking lot. This project is called as Car-Park Occupancy Information System (COINS), and it was tested using simulation model and also in real-case scenarios. Keywords—Image processing, car park occupancy, object detection, information system. I. INTRODUCTION t is very common to have huge crowds in places like mega shopping malls and stadiums during peak periods. In shopping malls, discounts and season sales offered by merchants can attract thousands of customers to come during the sale periods. Most of the customers travel to the shopping malls with their own vehicles and it is not surprising to see that car parks are always full during these periods. However, research into drivers' parking behavior indicated that this does not seem to stop many drivers from queuing at their favorite car park for significant periods [1]. New generation information services have been proposed or developed to replace traditional "Full/Spaces" sign at the entrance. These come in the form of parking information via mobile phones, personal digital assistants (PDAs), RDS-TMC, navigation systems and Urban Traffic Management and Control (UTMC) [1]. Parking guidance system such as the one developed at Shinjuku, Japan [2] and Tapiola, Finland [3] are examples of systems developed to guide drivers in finding vacant car parks. Manuscript received 23 November 2007. D.B.L. Bong is a Lecturer with Faculty of Engineering, Universiti Malaysia Sarawak, 94300 Kota Samarahan, Sarawak, Malaysia. (phone: 60- 82-583313; fax: 60-82-583409; email: bbldavid@feng.unimas.my). K.C. Ting is a Postgraduate Student with Faculty of Engineering, Universiti Malaysia Sarawak, 94300 Kota Samarahan, Sarawak, Malaysia. K.C. Lai is a Postgraduate Student with Faculty of Engineering, Universiti Malaysia Sarawak, 94300 Kota Samarahan, Sarawak, Malaysia. There are mainly four categories of car park guidance systems using different technologies - counter-based, wired sensor-based, wireless sensor-based and image-based. Counter-based systems use sensors to count the number of vehicles entering and exit a car park area. This can be gate-arm counters and induction loop detectors located at the entrances and exits [3]. This system can give information on the total number of vacant lots in a closed car park area, but does not help much in guiding the driver to the exact location of the vacant lots. Wired sensor-based system, on the other hand, is using detection sensors such as ultrasonic sensors which are installed at each parking lot. These sensors are wired to a central control unit that store and manage the parking occupancy information. This information is then relayed to display panels at strategic locations in the car park. The display panels provide information, direction and guide the drivers to vacant parking lots. Yu and Liu [4] used vortex berth detector in the parking guidance system. MCU is used to process the sensor signal and parking information is transmitted through CAN bus interface to the control centre. Wolff et al used magnetic field sensors and wired-based concept to test on their simulation model [5]. One shortcoming of the wired sensor systems is long and complicated wiring is required from parking lots to the central control unit. Also, the cost for developing this system is high because large amount of sensor units are required and long distance of wiring from parking lots to the control centre. With the advancement of wireless technologies, wireless- based methods have been employed in parking guidance systems. Wireless sensors nodes are deployed and each parking lot is equipped with one node. Tang, Zheng and Cao [6] developed such a system using Crossbow Mote products and the extended Crossbow XMesh network architecture. The sensor board is equipped with the sensors of light, temperature, acoustic and a sounder. Benson et al used DSYS25z sensing nodes which is an ATMega 128L micro- controller based system with RF transceiver and antenna. Wang, Zhou and Li adopted HMC1021 anisotropic magneto- resistive magnetic field sensor from Honeywell together with microprocessor and transceiver for the same purpose [7]. There are also other researches using wireless-based methods such as in [8] and [9]. Examples of start-up companies that are actively working in this field are ParkingCarma [10], Streetline Networks [11] and vehicleSense [12]. University of California at Berkeley [13] and EPFL Switzerland [14] carried out Integrated Approach in the Design of Car Park Occupancy Information System (COINS) D.B.L. Bong, K.C. Ting and K.C. Lai I IAENG International Journal of Computer Science, 35:1, IJCS_35_1_02 ______________________________________________________________________________________ (Advance online publication: 19 February 2008)