A Lightweight Low-Altitude Drone Detection And alert system

Project Code :TCMAPY2122

Objective

The primary objective of this project is to develop an efficient and accurate drone detection system that operates in low-altitude environments. The system aims to detect drones using state-of-the-art machine learning algorithms like YOLO11n and YOLO12n, ensuring high detection accuracy. Another key objective is to ensure the system is lightweight and can perform detection and provide alerts without high computational requirements, making it suitable for various deployment scenarios. The system also aims to provide a user-friendly interface, allowing users to register, log in, and easily interact with the detection results. Additionally, the system should offer real-time detection through live video feeds and store detection history for future reference. The objectives also include evaluating the models’ performance using the mAP metric and comparing the results between different YOLO models. Ultimately, this project seeks to provide a reliable and scalable solution for drone detection, enhancing security and safety in areas at risk from unauthorized drone activity.

Abstract

The increasing use of drones in various applications necessitates the development of efficient detection systems for their identification and alerting, especially in low-altitude environments. This project focuses on creating a lightweight and accurate drone detection and alert system using machine learning algorithms. By leveraging YOLO11n and YOLO12n models, the system is designed to detect drones from images with high precision. The system includes user-friendly features such as image detection, live detection with alerts, and history tracking. The dataset used for training and validation consists of images labeled with drone data, allowing the model to learn distinguishing features for effective classification. The performance of the models was evaluated using the mean Average Precision (mAP) metric, achieving impressive results with mAP@0.5 values of 0.9806 for YOLO12n and 0.9782 for YOLO11n. The backend is built using Python with Flask, and the frontend is developed using HTML, CSS, and JavaScript. SQLite is used for storing user information and detection history. The system is intended for environments where accurate drone detection is crucial, offering a reliable and accessible solution.

Keywords: Drone detection, YOLO models, machine learning, real-time alerts, low-altitude, image classification, Flask, SQLite, object detection, system performance.

NOTE: Without the concern of our team, please don't submit to the college. This Abstract varies based on student requirements.