Automated 3D Reconstruction of a Scene From Persistent Aerial Reconnaissance Video at High Zoom

ABSTRACT: Persistent aerial reconnaissance of an area of interest is often performed in support of ground operations. This reconnaissance is often completed with passive full motion video with the platform flying an orbit. While high-fidelity automatically-generated 3D models of the viewed scene would greatly benefit analyst understanding of the environment, frequent zoom changes and switching between color, electro-optical, and infrared sensing greatly complicates model generation. Neya Systems proposes to develop a 3D scene model generation capability that is not hindered by frequent zoom changes or changes between color, electro-optical and infrared video. A performance model will be developed that characterizes expected accuracy of the reconstruction.; BENEFIT: The benefits of three-dimension (3D) context for tactical and ISR missions are widely recognized and range from providing the spatial awareness to the Warfighter during tactical operations to providing the common spatial representation for multi-INT data fusion during manual and/or automated ISR exploitation. By providing this 3D information regardless of the video characteristics used to generate the information, our solution greatly increases the scenarios within which this capability can be used.