期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：1998
卷号：XXXII Part 4
出版社：Copernicus Publications
摘要：Since the launch of high-resolution sensors, the use of satellite images as a major source of spatial information has beenthe subject of extensive research in a broad range of applications. In particular, the extraction of land cover informationfrom remotely sensed data and the use of this information as input into geographical information systems (GIS) hasreceived considerable attention over the last ten years. The successful use of GIS as a decision support tool can only beachieved, if it becomes possible to attach a quality label to the output of each spatial analysis operation. Thus the accuracyof multispectral classification gained more attention. In a GIS, the data is usually stored in terms of objects insteadof individual pixels. To this end, the classification result has to be segmented. An important aspect of this research is thepropagation of the uncertainty of a pixel belonging to a class to the uncertainty of the pixel belonging to a region. Differentapproaches for image segmentation will be presented, that take the thematic uncertainty of the pixels into account.They will be applied and verified to a small test area
