Altitude, Latitude, and Longitude Reference Database of Man-Made Obstacles

The objective of this proposal is to develop a method for the automated extraction of man-made obstacles from air and space borne digital imagery. The goal is to achieve detection accuracy of 20 feet vertical over an observation area of 1000 km2 or more. The approach taken to meeting the objective is systematic, model based and applicable to a large number of available observation platforms, near earth as well as orbital without the need to provide specialized optics. The approach employs the use of a single lens imager coupled with positional as well as attitude information. By being systematic and model based, all error sources will be modeled allowing us to develop a methodology applicable to the performance requirements. The quality of the information is expected to be sufficiently good to support the field operational use, while being straightforward in implementation to support a range of users.