Acoustic Sensor Network-Based Parking Lot Surveillance System

Abstract

Camera-based surveillance systems are commonly installed in many parking lots as a countermeasure for parking lot accidents. Unfortunately these systems cannot effectively prevent accidents in advance but provide evidence of the accidents or crimes. This paper describes the design, implementation, and evaluation of an acoustic-sensor-network-based parking lot surveillance system. The system uses sensor nodes equipped with low-cost microphones to localize acoustic events such as car alarms or car crash sounds. Once the acoustic event is localized, the cameras are adjusted to the estimated location to monitor and record the current situation. We conducted extensive experiments in a parking lot to validate the performance of the system. The experimental results show that the system achieves reasonable accuracy and performance to localize the events in parking lots. The main contribution of our work is to have applied low-cost acoustic sensor network technologies to a real-life situation and solved many of the practical issues found in the design, development, and evaluation of the system.