基于多源遥感数据的第四系覆盖物分类方法研究:以内蒙古旗杆甸子幅1∶5万填图试点为例

详细信息    查看全文 | 推荐本文 |

英文篇名：Classification of Quaternary Coverings in Desert Grassland Shallow Cover Area Based on Multi-Source Remote Sensing Data:A Case of 1∶50 000 Pilot Geological Mapping in Qigandianzi,Inner Mongolia 作者：张艳 ; 孙杰 ; 于长春 ; 孟鹏燕 ; 郭锥 英文作者：Zhang Yan;Sun Jie;Yu Changchun;Meng Pengyan;Guo Zhui;Faculty of Information Engineering,China University of Geosciences(Wuhan);China Aero Geophysical Survey & Remote Sensing Center for Land and Resources;Hubei Institute of Land Surveying and Mapping;CRCC Harbour & Channel Engineering Bureau Group Survey &Design Institute Co.,Ltd.; 关键词：多源遥感数据 ; 第四系覆盖物 ; 分类方法 ; 荒漠草原浅覆盖区 英文关键词：multi-source remote sensing data;;quaternary covering;;classification method;;desert grassland shallow cover area 中文刊名：DZKQ 英文刊名：Geological Science and Technology Information 机构：中国地质大学(武汉)信息工程学院;中国国土资源航空物探遥感中心;湖北省国土测绘院;中铁建港航局集团勘察设计院有限公司; 出版日期：2019-03-15 出版单位：地质科技情报 年：2019 期：v.38;No.185 基金：国家重点研发计划课题“综合航空地球物理探测系统集成与方法技术示范研究”(2017YFC0602201);; 中国地质调查局项目“秦岭及天山等重点成矿区带航空物探调查”(121201203000160006) 语种：中文; 页：DZKQ201902034 页数：10 CN：02 ISSN：42-1240/P 分类号：287-296

摘要

遥感图像分类技术对于荒漠草原浅覆盖区第四系覆盖物分类具有重要意义。以内蒙古旗杆甸子幅1∶5万填图试点为例,基于ASTER、GF-2等多源遥感数据,利用植被抑制法、波段比值法、主成分分析以及纹理信息提取等多种方法,充分考虑了多光谱数据的光谱信息和高分辨率数据的形状、空间结构、纹理信息等特征,结合面向对象分类法,对研究区第四系覆盖物进行了分类,并比较分析了不同分类方法的分类效果与精度。结果表明:将波段比值、主成分分析以及纹理分析多种特征作为辅助数据参与分类,其分类效果优于基于单一ASTER数据进行的分类;通过几种不同分类方法的比较分析,发现多特征面向对象分类的总体精度最高,达到85.40%,比多特征传统监督分类的总体精度提高了约11%,分类影像上地物边界清晰。该法分类技术可以为荒漠草原浅覆盖区的地质填图提供相关技术支持。
        Remote sensing image classification is of great significance for the classification of Quaternary coverings in desert grassland shallow cover area. The 1∶50 000 pilot geological mapping in Qigandianzi, Inner Mongolia was studied as a case. In order to consider the spectral information of multi-spectral data and the shape, spatial structure, texture information of high-resolution data, the ASTER GF-2 and vegetation suppression, band ratio, principal component analysis, texture information extraction were used and combined with object-oriented classification to classify the quaternary coverings. The classification effect and precision of different classification methods were compared. The result shows that the features of band ratio, principal component analysis and texture analysis utilized in the classification as auxiliary data lead to better classification than the classification based on single ASTER data and the classification accuracy is also significantly improved. The object-oriented classification which considers information of spectrum, space and textureis superior to the traditional supervised classification effect which mainly depends on spectral information. The comparison and analysis of several different classification methods reveal that the overall accuracy of multi-feature object-oriented classification is the highest, reaching 85.40%, which is about 11% higher than that of multi-feature traditional supervised classification. The boundary of the classification image is clear. The classification technique discussed in this paper intends to provide help geological mapping in grassland desert shallow cover area.

引文

[1] Othman A A,Gloaguen R.Improving lithological mapping by SVM classification of spectral and morphological features:The discovery of a new chromite body in the Mawat Ophiolite Complex(Kurdistan,NE Iraq)[J].Remote Sensing,2014,6(8):6867-6896.
    [2] 于亚凤,杨金中,陈圣波,等.基于光谱指数的遥感影像岩性分类[J].地球科学:中国地质大学学报,2015,40(8):1416-1419.
    [3] 项甜甜,陈圣波,刘彦丽,等.基于高光谱数据的河北滦平地区岩性信息提取方法试验研究[J].遥感信息,2013,28(2):82-92.
    [4] 陈星,贾晓青,张学斌.监督分类在第四纪堆积物遥感解译中的应用:以天津武清地区为例[J].智能应用,2013,4(4):54-56.
    [5] Yu L,Porwal A,Holden E J,et al.Towards automatic lithological classification from remote sensing data using support vector machines[J].Computers & Geosciences,2012,45:229-239.
    [6] 赵建华,杨树锋,陈汉林.基于分形纹理的遥感图像岩性识别方法[J].遥感信息,2004(2):2-4.
    [7] 杨佳佳,姜琦刚,陈永良,等.基于最小二乘支持向量机和高分辨率遥感影像的大尺度区域岩性划分[J].中国石油大学学报:自然科学版,2012,36(1):60-67.
    [8] 查逢丽,马明,陈圣波,等.基于植被抑制法的植被覆盖区多光谱遥感岩性分类[J].地球科学:中国地质大学学报,2015,40(8):1403-1408.
    [9] Crippen R E,Blom R G.Unveiling the lithology of vegetated terrains in remotely sensed imagery[J].Photogrammetric Engineering and Remote Sensing,2001,67:935-943.
    [10] Gad S,Raef A.Factor analysis approach for composited ASTER band ratios and wavelet transform pixel-level image fusion:Lithological mapping of the Neoproterozoic Wadi Kid area,Sinai,Egypt[J].International Journal of Remote Sensing,2012,33(5):1488-1506.
    [11] Sandra J,Benjamin B,Richard G,et al.Remote sensing based improvement of the geological map of the Neoproterozoic Ras Gharib segment in the Eastern Desert(NE-EGYPT) using texture features[J].Journal of African Earth Sciences,2015,111:138-147.
    [12] Amer R,Kusky T,Ghulam A.Lithological mapping in the Central Eastern Desert of Egypt using ASTER data[J].Journal of African Earth Sciences,2010,56(2):75-82.
    [13] 陈波,胡玉福,喻攀,等.基于纹理和地形辅助的山区土地利用信息提取研究[J].地理与地理信息科学,2017,33(1):1-7.
    [14] Haralick R M,Shanmugan K.Texture features for image classification[J].IEEE Transaction On Systems Man and Cybernetics,1973,3(6):610-621.
    [15] 张翠芬,余健,郝利娜,等.多尺度纹理及多光谱影像协同的遥感岩性识别方法[J].地质科技情报,2017,36(4):236-243.
    [16] 张金钟.基于面向对象的遥感影像岩性分类研究[D].西安:长安大学,2016.
    [17] Makkeasorn A,Chang N B,Li J.Seasonal change detection of riparian zones with remote sensing images and genetic programming in a semi-arid watershed[J].Journal of Environmental Management,2009,90(2):1069-1080.
    [18] Theoharatos C,Economou G,Fotopoulos S.Color edge detection using the minimal spanning tree[J].Pattern Recognition,2005,38(4):603-606.