Year: 2025
Type: article
Abstract: In the modern era, the integrity and safety of secure environments have become critically important. To address these challenges, an AI-powered UAV-patrolling drone system for real-time unauthorized person detection has been proposed. This drone executes predetermined flight paths, strategically covering surveillance gaps left by static CCTV cameras or human guards. The system integrates multiple state-of-the-art technologies, incorporating advanced facial recognition using Dlib-HOG, CNN, VGG-Face, Google FaceNet, and OpenFace, along with comprehensive facial analysis, providing real-time analysis of race, age, gender, and facial expressions. This technology is especially valuable for securing large venues, critical infrastructure, and high-profile events where unauthorized access poses significant risks. The system's hybrid architecture allows for optimal performance across different lighting conditions, angles, and crowd densities, setting a new standard for intelligent surveillance systems.
Authors: Khurram Iqbal, Muhammad Saad Bin Ehtisham, Muzammil Ahmad Khan, Adnan Chohan, Perfshan Erum, Khalil ur Rahman
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4413296850