Multi-Source Vertical Obstruction Generation

The objective of this Phase II SBIR development effort is to automate the process for extracting obstruction clearance features. These processes developed here will have direct applicability to other automated geospatial feature extraction problems. This development will include combining computer vision based algorithms and photogrammetric techniques to automatically extract vertical obstruction features from stereo imagery. These features included communication towers, utility poles, buildings, and roads. We will implement advanced image preprocessing, variations of the Hough transform, and 3D feature correlation techniques to allow the user to rapidly extract pertinent 3D features autonomously from a scene which historically have taken many hours using existing manual techniques. These processes will be manifested in commercial modules to Research Systems Inc. ENVI image processing software. The commercialization of these capabilities will greatly reduce the time needed to generate 3D digital terrain database for a wide variety of time sensitive government and commercial applications.