Cost Effective Smart Parking System on Campus

Abstract

IoT technology has recently been applied to many fields and has contributed to improving the convenience and quality of life of humans. In this study, we focus on developing an accurate, cost-efficient solution to the congestion, gas-wasting, stress-inducing struggle that exists for commuters while parking at a university. Our proposed system simply uses a single Raspberry Pi and camera module stationed above a parking lot and applies deep learning techniques to identify a car and non-car object to detect vacant and non-vacant parking spots. The system delivers parking-related information to commuters through an Android app in real-time. We have tested our system in a simulated environment and found that the system provides an average of 98% confidence in distinguishing between cars and non-cars and detecting whether a parking space is vacant or occupied.