Near Real-time Privacy Protection: Automated Location-dependent Video Blurring in UAV live-streams

Abstract

In today’s world, privacy is becoming a major concern, especially with the use of drones for surveillance and recreational purposes. This paper presents a novel approach to privacy protection in UAV live-streaming by introducing an automated video blurring system that operates in near real-time, replacing time-consuming operations in the post-processing stage. Our method leverages the Scale Invariant Feature Transform algorithm to match live footage with a pre-constructed aerial template image, enabling the blurring of private properties in near real-time, allowing our UAV greater freedom of mobility whilst preserving the privacy of residents at ground level. This solution aligns with the EU’s General Data Protection Regulation (GDPR), balancing utility and privacy rights. This proposed framework has the potential to significantly aid the UAV industry by providing a practical tool for privacy preservation during aerial surveys and recreation drone flights.