The Study On Path Search And Landform Classification From DEM

Path search and landform classification from Digital Elevation Model(DEM)data that are hot topic problems of digital terrain analysis in various study areas such as geography,geomorphology and geographical information science have profound application value in human production and life practice.It's important that consider the impact of terrain on the trfficability to search for the appropriate path for procession in accordance with direction of forward motion in complex natural terrain.In order to solve this practical issue,a path search algorithm is proposed along the direction.In addition,Because the landform affects the configuration of resources,it is necessary to study the method of automatic division of basic landform types to scientifically understand the geomorphological features and make rational use of natural resources.The main work of the paper is divided into the following two aspects:(1)Considering the terrain flatness,terrain and other factors on the impact of the trafficability,the method measures the trafficability of the area with slope,relief and roughness extracted by DEM data.On this basis,the path skeleton line of the accessible area is established and transformed into Directed graph.Then using Dijkstra algorithm finds multiple paths consistent with direction of forward motion from path topology.(2)At present,the result of the classification show that the accuracy has yet to be improved.Therefore,in order to improve the accuracy of topography classification.With geomorphologic shape and geomophogensis for the principle,the method first determines the topographic factors from mathematical implications and geography significances to construct geomorphological classification samples.Using trained convolution neural network to classify the plains and the mountains part of Qinghai-Tibet Plateau,Guangdong and Guangxi hills,then adopting decision tables to classify further.