摘要
地貌形态分类对于地貌研究、土壤制图、滑坡防治等诸多地学研究及应用领域都有着重要作用。数字地形分析领域的研究者将地貌形态分类知识与地理计算相结合,已开发出许多地貌形态自动分类方法。该文将现有的地貌形态自动分类方法分为3类进行讨论:基于聚类的方法、基于规则知识的自动分类方法和基于典型样点的自动分类方法。基于聚类的方法对地貌形态分类专家知识缺乏考虑,聚类结果常难以明确对应到目标地貌形态类型;基于规则知识的方法常需要显式给出分类规则,应用难度较大,扩展性较差;基于典型样点的方法具有较好发展前景,但还有待改善对隐含专家知识的利用程度。各类方法目前对复合形态类型均难以有效提取。对于可能的方法改进方向,该文从空间结构特征信息、隐性专家分类知识的来源两方面进行了讨论。
Accurate map of landform classification is one of the important basic data for many geographical researches and appli-cations(such as geomorphological research,digital soil mapping,landslide treatment engineering,etc.).Currently many methods of landform automatic classification have been designed through ct^mbining geoct^mputation and domain knowTledge on landform classification.However,their application results(especially for composite landform types)are still unsatisfactory due to the character is tics of landform types,such as semantic vagueness,multi-scale and spat ial structure,w hich is hard to be quantified.In this paper,existing landform classification methods are classified to be three types,including clustering-based methods,rule-based methods and typica-sample-based methods.The clustering-based methods lack consideration on expert classification knowledge.This situation makes the results from this type of method often cannot well match the target taxonomy of landform type.The rule-based methods need users to formalize the expert knowledge as explicit rules,which makes them not only hard to apply in practice but also less extensible to other areas.The typical-sample-based methods,although having good potential in ap-plication,still need more improveiment on the utilization of implicit expert knowTledge.In general,current automatic classification methods of landform are difficult to effec:tively extract composite landform types.Possible improvements for landform classifica-tion methods are further discussed from two aspects,including consideration of information on spatial structure and the source of implicit knowledge on landform classification.
作者
王彦文
秦承志
WANG Yan-wen;Q1N Cheng-zhi(StateKey Laboratory of Resources and Environmental 1 Tiformatuon System, Institute of Geographic Sciences and Natural Resources, CAS, Beijingl00101;College of Resources and Environment, University of Chinese A cademy of Sciences, Beijing 100049;Jiangsu Center for Collaborative Innovation in Geographical Inform.ation Resource Development and Application, School of Geograj) hic Science, Nanjing Normal Ihdversity, Nanjing 210023, China)
出处
《地理与地理信息科学》
CSCD
北大核心
2017年第4期16-21,共6页
Geography and Geo-Information Science
基金
国家自然科学基金项目(41422109
41431177)
资源与环境信息系统国家重点实验室自主创新项目
