Research And Application Of Terrain Segmentation Method Based On Semantics

Digital elevation model(DEM)is an important part of national spatial data infrastructure,which contains abundant geomorphological information,and has a wide range of applications in land management,construction planning,disaster prediction and so on.However,the grid data in DEM is taken as the mainstream data currently,lacking local regional characteristics,and with the increase of artificial reconstruction areas,the terrain structure tends to be more complicated,making the understanding of terrain more difficult.Therefore,in order to strengthen the semantic understanding of terrain,the terrain segmentation methods based on semantics are studied in this paper,and then the segmentation results are taken as objects,and the topological relationship among terrain objects is established in the study of the submergence analysis algorithm based on terrain objects.The main contents of this paper are as follows:(1)A terrain semantic segmentation algorithm based on marked watershed is proposed,which divides terrain units by combining the idea of positive and negative terrain and marked watershed.The experiments show that the algorithm not only achieves the semantic segmentation of terrain and obtains terrain units with a single semantic attribute,but also effectively avoids the problems of parallel boundaries and missegmentation existing in the traditional slope unit extraction methods.(2)A terrain semantic segmentation algorithm based on DEM texture and watershed is proposed.By combining DEM texture feature information and image gray level information,a hybrid gradient is constructed to further improve the accuracy of segmentation edges of terrain semantic units.The experiments show that the algorithm improves the accuracy of edge location on the basis of effectively extracting terrain semantic units.(3)Taking terrain semantic unit as the basic object,the topographic data is established,and the submergence analysis algorithm based on terrain objects is proposed.The experiments show that this algorithm can dynamically simulate the submergence situation in the study area,and effectively solve the problems of the difficulty in judging the flow direction of flat ground and the difficulty in realizing the submergence analysis of artificial terrain in current submergence analysis.