土地利用现状数据缩编是为了满足土地管理工作中不同级别的行政区需求对不同尺度的土地利用数据进行的简化、分类、归纳、符号化等工作,缩编应能使不同尺度的数据表达的土地利用信息保持最大程度的一致,缩编的质量直接影响缩编结果可用性。本文提出了缩编结构相似度指数、缩编布局相似度指数、缩编图斑适宜度指数辅助传统的相关性统计指标,用来衡量土地利用现状图缩编结果的质量,总结出基于GIS的栅格数据空间分析、空间统计功能实现土地利用现状数据缩编质量评价的技术路线和方法。以南京市为例,计算出土地利用数据从1∶5万缩编到1∶10万的各项缩编质量评价指数,验证了方法的可行性。

Land use data are the core of land management. The scales of land-use data will be different at different levels of administrative districts, so it is necessary to transform data between different scales through generalization by methods of simplification, classification, induction, symbolization. Generalization should enable the expression of different scales of land use information to maintain the greatest degree of consensus. Many achievements have been made in map generalization of land use data during the past years and have been focused on transformation method, but they are little concerned about the generalization results quality. According to our study, the generalization quality has direct impact on the availability of generalization results, and quality of generalization can be evaluated with the function of GIS spatial statistical and spatial analysis of the raster datum. The relevance of the traditional statistical indicators and generalization structural similarity index, generalization layout similarity index and land block suitable index can be used to measure the quality of generalization results of land use map. In this paper, detailed introduction about the formula, method and process of generalization quality evaluation of land use data is given supported by IDRISI software. The case study is related to the production quality of land use data generalization from Nanjing city, Jiangsu Province. It presents an implementation of a new land use generalization methodology for land use data production. Using the data of land use map of 2005, from the analysis of data, simple statistical correlation with the traditional index such as Cramer's V index is 0.8319, Kappa is 0.9151, generalization quality is good, but the Whole Structural Similarity Index is 0.87, with Whole Layout Similarity Index for generalization being 0.82, and Land Block Suitable Index for 4 Ha being 0.74, There is a certain excessive generalization to land block. According to these quality assessment index values of land use data generalization results of Nanjing land use data which performed from the 1:5 million to 1:10 million, the generalization results meet the basic requirements in addition to some excessive generalization. This quality assessment method of land use data generalization is effective and it gives an opportunity for evaluating the consistency of various scales for different purposes automatically.