An adaptive parallel vectorization method for RS segmented raster map

Segmentation and vectorization are basic steps of converting segmented raster image to vector format, which is important for object-oriented analysis of remote sensing (RS) information. However, it will encounter some problems when process large RS segmented raster map´s vectorization. In this paper, an improved vectorization algorithm based on fast labeling technology is put forward as the preliminary of the main process. Then coupled with the adaptive and paralleled grain size computation model, a paralleled vectorization method is finally proposed. Experiment results show that this method can be implemented in paralleled way successfully. Moreover, besides achieving quality result of object building, it also has a high execution efficiency with minimum computing resources cost. Therefore, the key technological problem in object-oriented information computation of massive RS data has been solved successfully.