基于历史矢量数据的土地覆被样本自动选择方法  

Automatic Sample Selection Method for Land Cover Based on Historical Vector Data

作　　者：孙姝娟 李民录 韩丽丽 SUN Shujuan;LI Minlu;HAN Lili(School of Architecture and Civil Engineering,Chengdu University,Chengdu 610106,China;Beijing PIESAT Information Technology Co.,Ltd.,Beijing 100195,China)

机构地区：[1]成都大学建筑与土木工程学院,四川成都610106 [2]北京航天宏图信息技术股份有限公司,北京100195

出　　处：《成都大学学报（自然科学版）》2020年第2期209-214,共6页Journal of Chengdu University（Natural Science Edition）

摘　　要：针对目前遥感技术在土地覆被中的应用需要依靠人工采集样本的现状,提出一种基于历史矢量数据的土地覆被分类样本自动选择方法:分别对遥感影像进行分割、对历史矢量数据进行筛选和缓冲处理,然后对分割结果和缓冲处理后的矢量数据进行空间分析和判断得到初始样本,再通过分类训练剔除错误样本得到最终样本,并以四川省某区域为试验区,对方法进行了测试和分析.结果表明,自动选择方法提取的样本准确率均在95%以上,相对于人工采集样本大大提高了效率;基于选择的样本对影像进行分类,结果精度也较高,能够有效地用于土地覆被专题图的快速更新.Aiming at the current situation that the application of remote sensing technology in land cover relies on manually-collected samples,an automatic sample selection method for land cover is proposed based on historical vector data.Firstly,remote sensing images are segmented and historical vector data is filtered and buffered,then spatial analysis is performed on the data to get initial samples,and at last incorrect samples are removed and correct samples are reserved.A study area in Sichuan Province is selected to conduct experiments by using the proposed method in order to test and analyze the method.The results show that the accuracy of samples obtained by automatic sample selection method is above 95%.This method greatly improves the efficiency compared with that of the manual sample collection.The accuracy of image classification based on these samples is precise,indicating that the method in this paper could be effectively used for the rapid updating of thematic land cover maps.

关 键 词：土地覆被 分类 样本自动选择 历史矢量 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]