| Περίληψη: | The availability of new high resolution satellite data brings with it the need for new image analysis methods. Traditional pixel-oriented algorithms do not give credit to the spatial coherence of high resolution imagery. In particular, for detection and classification of man-made structures, object-based procedures are much more appropriate. The use of object features such as scale, compactness, orientation and texture, in addition to spectral characteristics, extends the possibilities of remote sensing satellite image analysis considerably. This book describes recent progress in object-based image interpretation, and also presents many new results in its application to verification of nuclear non-proliferation. A comprehensive workflow and newly developed algorithms for object-based high resolution image (pre-) processing, feature extraction, change detection, classification and interpretation are developed, applied and evaluated. The entire analysis chain is demonstrated with high resolution imagery acquired over Iranian nuclear facilities.
|
|---|
