Development of UAV -based Automated Vehicle Recognition System for Parking Enforcement

摘要

This paper presents the development of an Unmanned Aerial Vehicle (UAV) based system that is designed to perform vehicle recognition for parking enforcement purpose. The proposed system is intended to be used to monitor and enforce parking fee payment in a multi-entrance and multi-exit open parking lot. This system is useful when the condition is not suitable to implement single-entry and single-exit barrier gate at the parking lot. The system is developed based on an UAV equipped with a camera to be deployed from a ground station and patrols the parking lot through a fixed flight path with a pre-defined waypoints mission to capture all vehicles' license plate that were parked in the parking lot. The images were then downloaded from the UAV for license plate detection and recognition via Google Cloud Vision. Postprocessing was performed to increase the detection accuracy of license plates. The detected license plates and their respective timestamp were logged in the database for parking fee charging purpose. To automate the process, the UAV control logic and graphical user interface was designed and developed using Microsoft Visual C# and DJI Windows SDK.