Determination Of The Best Grid Resolution Of 1:50000 DEM

Grid resolution is the basic property of grid DEM, the smaller grid resolution DEM may include more detailed topographical information, but in the same time it is required more computer capacity and increasing speed. Too much detailed data will not only bring many data redundancy which cause inconvenience for the calculation, but also affect the description of macro-topographical features. The larger grid resolution DEM is bound to loss many terrain information, which describe topography does not match the actual terrain so that can't meet the application requirements. Thus we need find appropriate grid resolution. The main research content is listed as follows:(1) Several methods of data acquisition and production for 1:50,000 DEM are compared. The author choosed the method which used contour lines to establish TIN model to produce grid DEM considering precision and the requirement of building different grid resolution of DEM.(2) According to the idea of ensuring the best grid resolution DEM based on the mean square error of slope, which was put forward by Huthinson, the author select the 1:50000 topographic maps of three typical landforms—uplands, hills, plains—as the studying area, then get the slope of the testing points on the TIN model which are used as original value by ArcGIS, calculate slope and the mean square error of slope of the testing points on the different gird resolution DEM based on VC++6.0 and GDAL (Geospatial Data Abstraction Library), analyse its changes with different grid resolution, find the inflexions of the mean square error of slope curve, finally get the best gird resolution.(3) As it is difficult to implement the best gird resolution on uplands and plains area using Hutchinson' theory, thus the author presented a solution based on topographic information entropy. In this method, the researcher gets topographic information of different grid resolution DEM by ArcGIS, calculates of terrain information entropy using Matlab, analyses terrain information entropy changes and finds terrain information entropy stable points, and then obtains the best grid resolution. The feasibility of this method was tested at hills, plains and uplands area.