摘要
线要素化简是地图自动综合中的重要部分之一。当前线化简算法的参数和阈值一般依赖于人工设定,且对不同的化简环境缺乏自适应学习能力。将线要素化简视作一种对局部化简单元的取舍二分类问题,从案例学习的角度出发,提出了一种新的基于支持向量机(support vector machine,SVM)的线化简方法。该方法首先以节点和弯曲为化简单元,从专家化简结果中自动获取化简案例;然后提取化简单元的特征描述项作为化简案例的属性空间,利用SVM机器学习方法进行训练,得到用于线化简的SVM分类器;最后通过SVM分类器对新的同类线要素中的化简单元作取舍分类,从而实现线化简。实验结果表明,该方法能够通过学习专家化简案例,在实际测试中较好地还原专家的化简意向,对化简单元取舍的分类正确率高,能够自适应地完成线化简。
Line simplification is one of the most important parts of automatic map generalization. Aiming at the problem that the effect of the current line simplification methods is too dependent on the artificial setting of algorithm parameters and thresholds, and lacking the adaptive learning ability for different simplification environments, this paper considers the line simplification as a binary classification problem about the selection or deletion of the simplification unit. From the perspective of case-based studying, a new line simplification method based on support vector machine(SVM) is proposed. Firstly, point and bend are taken as the simplification unit, the simplification case is automatically obtained from the simplification results of experts, and then the feature description items are extracted as the attribute space of the case sample. The line simplification classifier is trained by SVM. Finally, the SVM classifier is used to classify the simplification units of the new line elements into selection and deletion, so that the line simplification is realized. The experimental results show that this method can well reduce the simplification intention of the experts by learning the simplification case, with a high classification accuracy for the selection and deletion of simplification units, and can adaptively complete the line simplification.
作者
段佩祥
钱海忠
何海威
谢丽敏
罗登瀚
DUAN Peixiang;QIAN Haizhong;HE Haiwei;XIE Limin;LUO Denghan(Institute of Geospatial Information,Information Engineering University,Zhengzhou 450001,China;32176 Troops,Kuerle 841000,China)
出处
《武汉大学学报（信息科学版）》
EI
CSCD
北大核心
2020年第5期744-752,783,共10页
Geomatics and Information Science of Wuhan University
基金
国家自然科学基金(41571442,41171305)。
关键词
地图自动综合
线化简
案例学习
支持向量机
automatic map generalization
line simplification
case-based studying
support vector machine
