Abstract: DEM data is widely used in many fields, but its applicability is severely restricted by the data void in production process. Therefore, it is crucial that the research on DEM data void repair. In this paper, the experimental areas of river valleys, basins, hills and mountains under four different geomorphic conditions in Taiyuan city are selected as void areas. ASTER GDEM data and direct mosaic method, inverse distance weighted interpolation method and delta surface filled method are used to reconstruct AW3D30 DEM data, and the restoration effects of different methods are compared and analyzed. The results show that: In the region of the basin, inverse distance weighted interpolation method has the best repair effect, followed by the delta surface filled method, and the direct mosaic method is weakest. However, in the valley, hill and mountain region with larger fluctuation, the delta surface filled method has better effect, and the texture features in the cavity region are obvious, and the cavity edge region is smooth, while the inverse distance weighted interpolation method is the weakest.

Key words: DEM data void repair, delta surface filled method, direct mosaic method, inverse distance weighted interpolation method, geomorphic conditions, Taiyuan