Orthometric Processing of Very High Spatial Resolution Frame Imagery Using Distributed Techniques

Abstract : Very high spatial resolution frame imagery (pixel areas <<1.0 m(exp 2) is becoming readily available from a number of government-owned and commercial sensors. These systems have and will be mounted on a variety of rotary and fixed-wing aircraft, and because of their small size, are ideally suited as organic sensors for Unmanned Aerial Vehicle (UAV) platforms. Because of the relatively small ground coverage of these systems, a large number of individual frames must be quickly and accurately georegistered to cover a typical area of interest. In a test conducted at Yuma, AZ during February 2004, a total of 88 frames of stereo imagery with 0.01 m(exp 2) spatial resolution were used to cover approximately 4 km along a roadway. Using a recently developed distributed processing system, the 88 frames were mosaicked and georegistered to meter level accuracy in less than 52 minutes. Incorporating state-of-the art hardware, the same number of frames could be used to cover more than 20 km on the ground, providing a valuable data source to multiple applications including targeting, route reconnaissance, and improvised explosive device (IED) detection.