栅格DEM微地形分类的卷积神经网络法  

Micro Landform Classification Method of Grid DEM Based on Convolutional Neural Network

作　　者：周访滨[1,2] 邹联华 刘学军[3] 孟凡一 ZHOU Fangbin;ZOU Lianhua;LIU Xuejun;MENG Fanyi(School of Traffic&Transportation Engineering,Changsha University of Science&Technology,Changsha 410114,China;Key Laboratory of Special Environment Road Engineering of Hunan Province,Changsha University of Science&Technology,Changsha 410114,China;School of Geography,Nanjing Normal University,Nanjing 210023,China)

机构地区：[1]长沙理工大学交通运输工程学院,湖南长沙410114 [2]特殊环境道路工程湖南省重点实验室(长沙理工大学),湖南长沙410114 [3]南京师范大学地理科学学院,江苏南京210023

出　　处：《武汉大学学报（信息科学版）》2021年第8期1186-1193,共8页Geomatics and Information Science of Wuhan University

基　　金：国家自然科学基金(41671446,41371421);特殊环境道路工程湖南省重点实验室(长沙理工大学)开放基金(kfj140502);长沙理工大学学术学位研究生科研创新项目(CX2020SS17)。

摘　　要：针对传统规则化知识的栅格数字高程模型(digital elevation model,DEM)微地形分类方法自动化程度低、分类不完全等缺陷,构建了一种适用于栅格DEM微地形自动分类的卷积神经网络(convolutional neural network,CNN)模型。借助该模型具有自动深入学习样本数据和挖掘隐含分类信息的优势,提出了栅格DEM微地形分类的卷积神经网络方法并创建了其自动化实现流程。以山体部位分类为典型样例进行实验验证分析,实验结果统计显示:在山体部位分出的山顶、山肩、背坡、麓坡、趾坡和冲积地6类微地形中,分类精准程度最高的为冲积地,最低的为趾坡,准确率分别达到了99.64%和92.95%;栅格DEM数据的像元大小影响其分类准确率,5 m×5 m的栅格DEM比2.5 m×2.5 m和10 m×10 m更适应山体部位分类的卷积神经网络法。Objectives:Micro landform classification of grid digital elevation model(DEM)is the founda-tion of digital landform refinement application and has broad application prospects in pedology,urban plan-ning,civil engineering,military affairs,diplomacy,etc.However,problems,e.g.,low degree of automa-tion and incomplete classification,are still existed in the micro landform classification method of grid DEM based on regular knowledge.With the advantages of convolutional neural network(CNN),a CNN method for grid DEM micro landform classification is constructed and its automated implementation process and ap-proach are created.Methods:Taking the advantage of CNN that can automatically learn the sample data and mine the hidden knowledge of the data set,it can automatically learn the features of the micro landform data set and realize the automatic classification of the micro landform combining with the micro landform data.Firstly,the plane curvature,section curvature,slope and elevation of grid DEM in the experimental area are extracted as data sets.According to the decision table and prior knowledge,the hill-position of the sam-ple data set is divided into 6 types,including summit,shoulder,back-slope,foot-slope,toe-slope and allu-vium.Among which,2/3 data are used as the training set,and the remaining sample data are used as the test set.According to the data requirements of the CNN and the characteristics of the data set,a convolu-tion layer is constructed,which can automatically learn the data characteristics in training set,determines the training model,and uses the test set to evaluate the model.s accuracy.After the model meets require-ments of the accuracy,the new micro landform factor generalizes the model are entered and the results of landform classification are outputted,i.e.,the model is used to automatically classify hill-position of the grid DEM data.Results:CNN is used for grid DEM micro landform classification.The experimental re-sults show that the overall accuracy of the model test dataset established by t

关 键 词：地形分类 栅格数字高程模型 卷积神经网络 山体部位 

分 类 号：P237[天文地球—摄影测量与遥感]