Segmentation of physiographic features from the global digital elevation model/GTOPO30

A methodology is presented for the segmentation of certain physiographic features (mountains, basins and piedmont slopes) observed in the Great Basin Section that belongs to the Basin and Range Physiographic Province of southwest USA, from the GTOPO30 global digital elevation model. A region-growing segmentation algorithm using as seeds ridge and valley pixels and appropriate gradient-region growing criteria, was applied for the extraction of mountains and basins. Certain artifacts were recognized and corrected using local physiographic knowledge. The pixels that remained unclassified were assigned to the piedmont slope class. The results were evaluated by comparing the boundaries of the extracted features (a) to those interpreted from the digital elevation model and (b) to those depicted in existing physiographic maps, It was concluded that the features extracted were in accordance with the features depicted on maps and on the digital elevation model. In addition the methodology was applied successfully to two physiographic sections in Iran. The results demonstrated the physiographic information content of GTOPO30 and pointed out its value for applications to regional geology and geomorphology. (C) 1999 Elsevier Science Ltd. All rights reserved.