Real-Time Video Stabilization for Micro-Air Vehicles

Traditional methods for recovering ego-motion parameters for stabilization purposes do not perform well for micro aerial vehicles (MAV) due to poorly constrained platform motion and poor video quality. SET has designed a system to effectively address the MAV video stabilization challenge. During Phase 1, we have demonstrated a technique based on salient feature detection and matching with additional capabilities for global matching that provides robust MAV video stabilization and addresses real-time processing issues. We have demonstrated system viability for real time performance. During Phase II, we will expand on this success to further enhance stabilization capabilities and transition the system within the AFRL MAV ground station. Enhancements to the stabilization will include improvements to the underlying stabilization system, multiple selectable user displays for optimal task based utilization, and incorporation of IMU/GPS data. Specific attention will be placed on enhancing capabilities to enable wide area situational awareness from MAV video, with further enhancements to the developed and demonstrated mosaicking capabilities. Transitioning of the system will be pursued aggressively to encompass multiple deliveries of MAV stabilization systems (a baseline enhanced version as well as an IMU/GPS driven system) to be embedded within the MAV ground station during the Phase II effort.