陕西省耕地质量分等省级成果汇总的多属性集簇方法

Summarizing methods of farmland classification by multi-attribute cluster in Shaanxi province

为了将1∶10000高精度县级耕地分等数据汇总形成1∶500000省级耕地分等数据,基于Arc GIS软件和Access软件,采用多属性集簇的方法对乡代码、地类代码、自然质量等、利用等及经济等属性控制条件进行集簇,采用移动、分割、增加等三种策略进行簇点视觉汇总。研究结果表明多属性集簇的方法能够满足耕地分等省级数据汇总的精度要求,提高工作效率,探索形成了科学可行的省级数据汇总方法体系,采用Arc GIS软件准确快速地构建了省、县两级耕地分等单元之间的可追溯查询关系,取得了良好的应用效果。该研究方法为实现耕地分等省级数据汇总提供理论依据和技术支撑,对耕地分等省级成果汇总和国家级成果汇总具有借鉴和参考价值。

Based on Arc GIS software and an Access database software platform, high precision of county- level farmland classification data at a scale of 1：10,000 were aggregated to create summarized a 1：500,000- scale provincial- level farmland classification data by the multiattribute cluster method. Landscape classification was conducted using country codes, land type data, classification of natural areas based on quality, landuse classification, economic classification, and multi- attributes of control conditions while using a clustering technique.Three methods are proposed to create a cluster point visual summary of the landscape as follows： 1） by moving the cluster points, 2） by dividing provincial units into county-level units,and 3） by generalizing the outlines of the county- level units. Research results： The cluster research method using multiple landscape attributes could be used to improve the efficiency and the precision of a provincial- level farmland classification system, and to explore the development of a scientifically sound and feasible method of summarizing provincial- level farmland data. Arc GIS was also used to construct a traceable query relation table between provincial- and country- level units. The results show that this method allows the development of an excellent farmland classification system. This farmland classification method provides a theoretical basis and technical support for the implementation of a provincial- level method of summarizing farmland data and the classification of farmland, even at the national level.