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：The geocoding of aerial and satellite imagery is traditionally a tedious manual task. Therefore numerous initiatives forautomating the geocoding procedure have been started. Among them, the EU project ARCHANGEL intends to rise asystem for automated integration of Image and Map Data for Change Detection. One part of this project aims onautomatic geocoding of the different satellite imagery.The approach of matching satellite imagery and vector data introduced in this paper can be characterized by the use ofarea objects. Area features are considered to be a subset that can be extracted from both vector data and satelliteimages. Hence these features represent a senseful basement for the matching process. For the vector side of theprocessing chain a common vector format was developed which allows the import of vector data from different databases, for example the German ATKIS data base, vector maps in DXF format or GDF data. Based on this data structurefurther processing steps like extraction of polygons or selection of certain object classes or layers can be applied. Inorder to obtain a set of matchable area features the preprocessed vector data are rasterized to be fed into the matchingprocess. Within the image baseline satellite images of various sources like SPOT XS, Landsat TM, MOMS and otherscan be imported into a uniform image format. Based on this image format a segmentation process is applied to obtainarea features from the image data. The applied matching procedure calculates polygon statistics from both thepreprocessed and rasterized vector data set and the segmented satellite imagery. In our approach additionally a set oftransformation coefficients obtained from a manual tie point editing is used to support the polygon matching. Thematching routine subsequently compares the polygon area, the length of the perimeter, the position of the centroid, andthe chain frequency of pairs of polygons.
关键词：Automatic Image Registration; Polygon Matching; Feature Extraction; Data Integration
