结合多级阈值与SOM的土地利用规划碎图斑识别  

Identification of fragmented polygons in land use planning maps by combining multi-level thresholds and SOM

作　　者：帅莹瑛 林浩顺[3] 毛政元[1,2] SHUAI Yingying;LIN Haoshun;MAO Zhengyuan(Academy of Digital China(Fujian),Fuzhou University,Fuzhou 350108,China;Key Laboratory of Spatial Data Mining and Information Sharing of Ministryof Education,Fuzhou University,Fuzhou 350108,China;Information Center of Land and Resources of Fujian Province,Fuzhou 350001,China)

机构地区：[1]福州大学数字中国研究院(福建),福州350108 [2]福州大学空间数据挖掘与信息共享教育部重点实验室,福州350108 [3]福建省国土资源信息中心,福州350001

出　　处：《测绘科学》2021年第4期186-191,共6页Science of Surveying and Mapping

基　　金：国家自然科学基金项目(41801324);福建省自然科学基金(面上)项目(2019J01244)。

摘　　要：针对土地利用规划地图数据更新过程中因地块调入调出产生的碎图斑的识别问题,该文在深入分析样本数据特征的基础上提出一种结合多级阈值与自组织特征映射(SOM)神经网络的碎图斑识别方法。首先通过面积和最大内角阈值排除不符合碎图斑特征的多边形;再将矩形度、圆形度、最小内角以及延展度作为特征向量输入训练好的SOM聚类模型,基于多边形特征相似性识别碎图斑,较好地弥补了目前制图实践中大量采用的统一面积阈值法指标选取单一、主观性强的不足,实现了碎图斑的准确识别。以福州市长乐区和福清市土地利用规划调整建设用地管制区数据为样本的验证实验表明,该文方法的识别精度与稳定性(针对不同区域数据的适应性)均优于目前制图实践中大量采用的统一面积阈值法。In order to solve the problem of identifying the fragmented polygons caused by land parcel adjustment during updating land use planning maps, this paper presented a solution by combining multi-level thresholds and self-organizing feature map(SOM) neural networks based on intensive analysis of multi-dimensional features of fragmented polygons in sample maps. It employed the area and maximum internal angle thresholds to exclude polygons that did not conform to fragmented features. And it input the feature vector consisting of rectangularity, circularity, minimum internal angle and extensibility of polygons into the trained SOM clustering model,then identified fragmented polygons according to their feature similarity.This method could make up for the deficiency of single index selection and strong subjectivity of the unified area threshold method which was widely used in the mapping practice,and realized the accurate identification of fragmented polygons.To evaluate the effectiveness of this method,it was applied to the land use planning map of Changle and Fuqing,Fuzhou.Experimental results showed that the accuracy and stability(adaptability to land use planning map in different regions)of the proposed method were both better than the unified area threshold method which had been widely used in cartographic practice.

关 键 词：土地利用规划 碎图斑 自组织特征映射 聚类 

分 类 号：P208[天文地球—地图制图学与地理信息工程]