Image-Based Visual Servoing for Automated Radar Control and UAV Tracking

Overview

This project develops a sophisticated image-based visual servoing system that combines advanced computer vision algorithms with radar control mechanisms to achieve real-time UAV tracking and automated radar positioning for surveillance applications. The research focuses on implementing an automated radar alignment system using state-of-the-art object detection and tracking algorithms to maintain continuous lock on unmanned aerial vehicles (UAVs) through integrated hardware-software control systems.

Key Achievements

• YOLOv8 + DeepSORT Integration: Implemented and optimized the detection-tracking pipeline, achieving 91% tracking recall and 13% improvement in processing speed through model pruning and optimized batching techniques.

• PLC-based Actuation System: Designed and integrated a programmable logic controller (PLC) based actuation interface that enables continuous UAV lock-in capabilities with seamless target handover mechanisms.

• End-to-End Pipeline Development: Authored comprehensive Python scripts, deployment configurations, and evaluation frameworks for training, inference, and benchmarking the complete visual servoing system.

Technical Stack

Computer Vision: YOLOv8, DeepSORT
Programming: Python
Control Systems: PLC-based actuation
Performance: 91% tracking recall, 13% speed optimization

Research Impact

The developed system demonstrates significant improvements in automated surveillance capabilities, combining deep learning-based object detection with robust tracking algorithms and hardware integration for practical defense and security applications.