MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641096137 A) filed by Velammal Institute Of Technology; Mr. Nithin Kumar R; Mr. Kannan G; Dr. Venkadesh R; Vignesh S; Balamathavan M; Raaja Annamalai M; and Parvathareddy Ruthwik Reddy on August 08, 2026, for An Ai-Based Autonomous Aerial Surveillance Drone System With Real-Time Object Detection, Tracking, Suspicious Activity Identification And Alert Generation.
Inventors include Mr. Nithin Kumar R; Mr. Kannan G; Dr. Venkadesh R; Vignesh S; Balamathavan M; Raaja Annamalai M; and Parvathareddy Ruthwik Reddy.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: This project presents the design and development of an Autonomous Surveillance Drone with a Tracking System that integrates Artificial Intelligence techniques to perform intelligent monitoring of public spaces. The rapid proliferation of urban spaces, combined with increasing safety concerns and the limitations of conventional fixed surveillance infrastructure, has created a pressing need for more adaptive, mobile, and intelligent surveillance solutions.The proposed system leverages a consumer-grade camera-equipped drone as a mobile aerial platform that continuously captures live video of designated monitoring zones. The captured video is processed in real time on a laptop-based server using state-of-the-art deep learning frameworks. Object detection is performed using the You Only Look Once (YOLO) model, which identifies persons and vehicles with high confidence and speed. Detected objects are then handed to a tracking sub-system based on DeepSORT (Deep Simple Online and Realtime Tracking) or ByteTrack algorithms, which assign persistent unique identities to each detected individual across consecutive video frames. A rule-based suspicious activity detection module analyses the tracked motion parameters, including displacement magnitude, velocity, and movement trajectory, to identify behaviours that deviate from normal patterns. When such anomalous behaviour is detected, the system transitions to an active tracking and alert state. A semi- autonomous control decision module computes positional offsets and generates directional guidance commands to assist the drone in following the suspect. Concurrently, an alert generation module captures a visual snapshot and location data, which are communicated to a simulated police monitoring interface. The system is designed to be cost-effective, practically deployable for academic prototyping and demonstration purposes, and scalable for future real-world deployment. The entire processing pipeline is implemented in Python using OpenCV for computer vision operations, the Ultralytics YOLO library for detection and tracking, and Flask for backend communication. Testing was performed using pre-recorded drone footage and simulated outdoor scenarios, demonstrating the system's ability to detect, track, and generate alerts reliably. This project contributes a practical, working prototype that showcases the potential of combining unmanned aerial vehicles (UAVs) with AI-powered computer vision to build next-generation surveillance systems that are intelligent, responsive, and capable of operating with minimal human intervention.
Disclaimer: Curated by HT Syndication.
