Files

Download

Download Full Text (518 KB)

Abstract

Drones are becoming more relevant and could potentially become a part of everyday life. From Amazon delivering packages to light displays used in the Super Bowl half-time show, drones are taking the place of people and inefficient systems. They are being used by military forces for scouting and relaying information from a distance, and by normal people every day to get an aerial view of a park or just for the entertainment of flying. This project develops a system for identifying and recognizing faces in real-time using a drone, specifically a Parrot ‘Bebop’, which is flying autonomously. The drone company, Parrot, provides an API for independent development that we used to program an automated flight pattern and to reformat the video for optimal face recognition. The video stream is sent from the drone via its personal WIFI hotspot to the GPU server to handle the computationally expensive power required by face recognition algorithms. The video is then displayed to the user monitor and an easy-to-use interface is provided for users to add people to the recognized face database. Our final product has the potential to act as “eye in the sky” for security in crowded areas and can eliminate human error in tense life-threatening police and military situations.

Publication Date

2017

Disciplines

Computer Engineering | Engineering

Faculty Advisor/Mentor

Dr. Alberto Cano

VCU Capstone Design Expo Posters

Rights

© The Author(s)

Date of Submission

May 2018

Real-time Face Recognition Drone Surveillance System

Share

COinS