Skyspy

Abstract

Urban areas frequently experience traffic congestion and safety concerns due to unauthorized parking in restricted zones. Traditional parking enforcement methods, which heavily rely on man- ual inspections, are labor-intensive, error-prone, and inefficient. To tackle these challenges, the SkySpy project proposes an autonomous drone-based system that utilizes advanced computer vi- sion and machine learning techniques to monitor and manage unauthorized parking in urban areas. The primary objective of SkySpy is to design and implement a quadcopter drone equipped with a high-resolution camera, GPS, and deeplearning algorithms. The drone autonomously patrols predefined routes, scanning for vehicles parked in no-parking zones. Upon detecting such vehicles, it captures high-resolution images and uses Optical Character Recognition (OCR) technology to read and record the vehicles license plate. The system then communicates this information in real-time to traffic authorities, enabling the automatic issuance of fines to violators. This reduces reliance on manual parking enforcement, making the process more efficient and accurate. The project integrates various hardware components, including a Raspberry Pi for image pro- cessing, a Pixhawk flight controller for drone navigation, and an infrared camera for nighttime surveillance. Additionally, the system features GPS modules for location tracking and radio telemetry modules for communication. The software stack includes OpenCV for image process- ing, YOLO for object detection, TensorFlow or PyTorch for deep learning, and MAVLink for drone communication protocols. SkySpy addresses the growing need for efficient traffic and parking management in expanding urban areas. The autonomous nature of the drone allows for continuous, real-time surveillance, reducing human error and labor costs. The system is initially designed for daytime monitoring but has the potential to be extended for nighttime use by incorporating infrared imaging. While the projects current scope is limited to detecting parking violations, its underlying technology could be adapted for broader applications such as general traffic monitoring, surveillance, or emergency response. In conclusion, SkySpy is a pioneering project that combines cutting-edge drone technology, machine learning, and computer vision to offer a scalable solution for urban traffic management. By automating the detection and reporting of unauthorized parking, SkySpy provides a more ac- curate, efficient, and sustainable alternative to traditional enforcement methods. The systems potential applications extend beyond parking enforcement, creating opportunities for use in other critical areas of urban management and public safety. keywords: Quadcopter; Object detection; GPS;Realtime;Pixhawk;OCR (Optical Character Recognition)