Guest Introduction: The Changing Shape of Computer Vision in the Twenty-First Century
International Journal of Computer Vision
| Hi-index | 0.00 |
Recently, airborne video surveillance platforms have gained greater acceptance for use in a variety of DoD missions due to their utility, affordability and autonomy. While a variety of airborne collectors and unmanned aerial vehicles may be equipped for video surveillance to support diverse mission needs, ground processing systems cannot handle this high data rate medium without some degree of autonomous processing to simplify and extend the exploita-tionprocess for video imagery.In this paper, we outline a novel approach for near-real-time video registration based on sensor model pa-rameter adjustments and the application of a Kalman filter. The goal of our Precision Video Registration (PVR) development is to register video with a reference image to provide accurate 3-D geolocations. Our sensor-based 3-D treatment is unique since most registration approaches employ only simple image-to-image mappings, such as affine transformations. In our approach, we explicitly model the projections between the 3-D world and 2-D images and perform registration in 3-D with greater accuracy and fidelity. PVR performance results show significant accuracy improvement over unregistered frame geolocation, and autonomously generated video mosaics appear smooth and seamless.