基于资源三号卫星影像的岩性监督分类研究
|
摘要
监督分类法在遥感影像分类中应用广泛,结合已有地质资料和实测地质剖面信息,将研究区岩性分为苏维依组泥岩和粗砂岩,库姆格列木群泥岩、砾岩和含砾粗砂岩,巴什基奇克组粗砂岩和砾岩,巴西改组细砂岩等8个类型,运用ENVI提供的常用监督分类方法——最大似然法,对研究区的遥感影像进行岩性分类研究。并利用混淆矩阵分别对不同影像的分类结果进行精度评价,结果表明ZY-3影像的分类总体精度达到了78.65%,Kappa系数为0.760 6,能够较好地进行岩性信息提取,具有较高的可靠性,可在地质调查工作中推广,能够有效地减少面积性地质测量工作,特别适用于岩石裸露程度较高的艰苦工作区。
The supervised classification method is widely applied in remote sensing image classification. In this paper,combined with the existing geological data and the measured geological section information,the lithology in study area is divided into eight groups,i. e. the mudstone and the gritstone in E2-3 s,the mudstone,the conglomerate and the gravelbearing gritstone in E1-2 k,the gritston and the conglomerate in K2 b and the fine sandstone in K1 b,and then the lithology classification is carried out in maximum likelihood,the most common supervised classification method in ENVI. Lastly,the confusion matrix is used to make the classification accuracy evaluation. The results show that the overall accuracy is up to78. 65% and Kappa coefficient is 0. 760 6. The method is suitable to the lithological information abstraction. Meanwhile,it is verified that the abstracted litholigic information combined with the supervised classification of image in auxiliary of the accurate measured geological section information has the high reliability,which can be popularized in the geological survey and effectively reduce the area geological survey work,especially for the hard region with the relatively high situation of bare rock.
The supervised classification method is widely applied in remote sensing image classification. In this paper,combined with the existing geological data and the measured geological section information,the lithology in study area is divided into eight groups,i. e. the mudstone and the gritstone in E2-3 s,the mudstone,the conglomerate and the gravelbearing gritstone in E1-2 k,the gritston and the conglomerate in K2 b and the fine sandstone in K1 b,and then the lithology classification is carried out in maximum likelihood,the most common supervised classification method in ENVI. Lastly,the confusion matrix is used to make the classification accuracy evaluation. The results show that the overall accuracy is up to78. 65% and Kappa coefficient is 0. 760 6. The method is suitable to the lithological information abstraction. Meanwhile,it is verified that the abstracted litholigic information combined with the supervised classification of image in auxiliary of the accurate measured geological section information has the high reliability,which can be popularized in the geological survey and effectively reduce the area geological survey work,especially for the hard region with the relatively high situation of bare rock.
引文
[1]李德仁.我国第一颗民用三线阵立体测图卫星——资源三号测绘卫星[J].测绘学报,2012,41(3):317-322.
[2]唐新明,张过,祝小勇,等.资源三号测绘卫星三线阵成像几何模型构建与精度初步验证[J].测绘学报,2012,41(2):191-198.
[3]张过.缺少控制点的高分辨率卫星遥感影像的几何纠正[D].武汉:武汉大学,2005.
[4]王文杰,蒋卫国,王维,等.环境遥感监测与应用[M].北京:中国环境科学出版社,2011.
[5]彭望琭,白振平,刘湘南,等.遥感概论[M].第三版.北京:高等教育出版社,2010.
[6]邓书斌.ENVI遥感图像处理方法[M].北京:科学出版社,2010.
[2]唐新明,张过,祝小勇,等.资源三号测绘卫星三线阵成像几何模型构建与精度初步验证[J].测绘学报,2012,41(2):191-198.
[3]张过.缺少控制点的高分辨率卫星遥感影像的几何纠正[D].武汉:武汉大学,2005.
[4]王文杰,蒋卫国,王维,等.环境遥感监测与应用[M].北京:中国环境科学出版社,2011.
[5]彭望琭,白振平,刘湘南,等.遥感概论[M].第三版.北京:高等教育出版社,2010.
[6]邓书斌.ENVI遥感图像处理方法[M].北京:科学出版社,2010.