>> 自然科学版期刊 >> 2021 >> 2021年02期 >> 正文
融合支持向量机和面向对象方法的矿区土地利用信息提取
时间: 2021-03-10 次数:

霍光杰, 胡乃勋, 陈涛,.融合支持向量机和面向对象方法的矿区土地利用信息提取[J].河南理工大学学报(自然科学版),2021,40(2):70-75.

Huo G J, HU N X, CHEN T,et al.Mining land use information extraction based on combining supportvector machine and object-oriented method[J].Journal of Henan Polytechnic University(Natural Science) ,2021,40(2):70-75.

融合支持向量机和面向对象方法的矿区土地利用信息提取

霍光杰1,2, 胡乃勋3, 陈涛3, 甄娜1,2

1.河南省地质环境监测院,河南 郑州 450000;2.河南省地质环境保护重点实验室,河南 郑州 450006;3.中国地质大学 地球物理与空间信息学院,湖北 武汉 430074

摘要:为了提高矿区土地利用信息遥感分类和提取精度,本文采用多源数据融合技术,利用高分二号卫星数据,融合面向对象思想和支持向量机方法,对河南省禹州市的采矿区进行以露天采场为主的矿区土地利用信息提取。结果表明,融合支持向量机和面向对象方法的矿区信息提取总体精度为86.44% Kappa系数为0.83,优于融合K近邻和面向对象的方法,表明该方法在矿区信息提取中有理想的精度,可为矿区的环境监测和科学管理提供可靠的技术支撑。

关键词:矿区信息提取;面向对象;支持向量机;土地利用

doi:10.16186/j.cnki.1673-9787.2020070002

基金项目:国家自然科学基金资助项目(61601418 );河南省财政项目(豫财预〔2014134号,〔2015128号,〔201644号) 第一作者简介:霍光杰(1978―),男,河南开封人,高级工程师,主要从事地质环境监测和地质环境信息化方面的研究工作。Email 751874636@ qq.com

收稿日期:2020/07/01

修回日期:2020/09/03

出版日期:2021/03/15

Mining land use information extraction based on combining supportvector machine and object-oriented method

Huo Guangjie1,2, HU Naixun3, CHEN Tao3, ZHEN Na1,2

1.Geological Environment Monitoring Institute of Henan Province Zhengzhou  450000 Henan China;2.Key Laboratory of Geological Environment Protection of Henan Province Zhengzhou  450006 Henan China;3.Institute of Geophysics and Geomatics China University of Geosciences Wuhan  430074 Hubei China

Abstract:In order to improve the the accuracy of remote sensing classification and extraction of land use information in mining areas by using multi-source data fusion technology and GF-2 satellite data the land use information of open pit mining area in Yuzhou City Henan Province was extracted based on combining support vector machine and object-oriented method. The results showed that the overall accuracy of mining area information extraction based on combing support vector machine and object-oriented method was 86. 44% and the Kappa coefficient was 0. 83 which was better than that of the combined K-nearest neighbor algorithm and object-oriented method. It showed that this method had ideal accuracy in mining area information extraction and could provide reliable technical support for environmental monitoring and scientific management of mining areas.

Key words:mining area information extraction;object-oriented;support vector machine;land use

 融合支持向量机和面向对象方法的矿区土地利用信息提取_霍光杰.pdf

最近更新