Improved Feature Extraction from High-Resolution Remotely Sensed Imagery using Object Geometry

摘要

Information extraction from high spatial resolution imagery is sometimes hampered by the limited number ofspectral channels available from these systems. Standard supervised classification algorithms found in commer-cial software packages may misclassify different features with similar spectral characteristics; leading to a highoccurrence of false positives. An additional step in the information extraction process was developed incorporat-ing the concept of object geometry. Objects are defined as a contiguous group of pixels identified as belonging toa single class in the spectral classification. Using results from the spectral classification, a supervised approachwas developed using genetic programming to select and mathematically combine feature-specific shape descrip-tors from a larger set of shape descriptors, to form a new classifier. This investigation focused on extraction ofresidential housing from QuickBird and IKONOS imagery of the Mississippi Gulf Coast before and immediatelyafter hurricane Katrina. Use of genetic programming significantly reduced false positives caused by asphaltpavement and isolated roofing material scattered throughout the image.