Topographic Feature Line Extraction Of Regular Grid DEM

The Digital Elevation Model(DEM)is discrete data that can be used to describe the true surface morphology in its digital form.In addition,information on the structural features of various topographic features is also included,such as point data(bottom valleys,saddle points,etc.),line data(ridge lines,valley lines,etc.),and surface data.Among them,the ridge line and the valley line are typical topographic feature lines,which form the dividing line of topographic changes.From another perspective,they are also watersheds and waterlines.It has great significance in studying topography and landforms.One of the hot topics in geoscience research is how to accurately extract the ridge line and the valley line from the topography.The regular grid DEM has the advantages of simple structure,rapid processing on the computer,etc.Therefore,the study of the ridge line and valley line of the paper is based on the regular grid DEM.In the extraction methods of the ridge line and the valley line,the paper combines the literature review and experimental research and analysis to improve the surface flow physics method to extract feature lines.The main research contents are:(1)Innocent DEM generation: The surface of the DEM is not smooth,and there are often recessed areas.These areas are called pseudogerminations.Caused by this phenomenon are: First,the error caused by the DEM data acquisition and processing,the second is the existence of some real surface morphology(such as reservoirs,dams,etc.).Therefore,before the feature line extraction,the filling process needs to be performed.When the traditional surface water physical model method extracts the characteristic line,it uses the culling and filling equal methods to process these false earthworms.There is a certain problem that there is a true surface form in the pseudo-defective land.These areas are often treated as pseudo-defectives together,resulting in the incompatibility of the extracted feature line results with the real terrain.In order to obtain a flawless DEM and to make the extracted feature lines more accurate,the thesis designs an effective and flawless DEM generation method,which combines the characteristics of multi-source data(DEM,TM data).First,DEM data is used to extract depressions,and the authenticity of the depressions is initially identified based on the characteristics of the depressions(defective areas);Then TM images are used to extract water information from the study area.This is due to the fact that most of the land in the depression is actually a real surface such as a reservoir or a dam,and the extracted water body information is used as the real ground.Finally,the original DEM data,data extracted from the depression,and data extracted from the water body are comprehensively analyzed and processed.Get an innocent DEM that matches the real terrain.(2)Determination of the best neighborhood window: When the feature line is extracted,there is a certain deviation in the position of the feature line extracted by the traditional method.In response to this problem,some scholars have proposed using the positive and negative terrain to correct it.Although the problem is solved to some extent,in the extraction of positive and negative terrains,the size of the DEM resolution and the selection of the neighborhood window have an important influence on the extraction of the positive and negative terrains,thus affecting the extraction of the feature lines.Therefore,under the condition that DEM resolution is determined,the paper analyzes how to obtain the best neighborhood window,that is,the basic idea is to seek the maximum value of the difference in the information of the different analysis windows,using the mean-variable-point analysis method.Get the best neighborhood analysis window for the study area.(3)Based on the above two steps,the feature line extraction experiment was performed on the study area,and the feature line results extracted from the paper method were compared with the feature line results extracted from the existing methods.It is concluded that the feature lines extracted by the paper method have better coherence and fewer break lines.