结合特征组与分层策略的遥感图像分类研究  被引量:3

Research on remote sensing image classification combined feature groups and hierarchy strategy

在线阅读下载全文

作  者:潘建平[1] 夏鑫 阳春花[2] PAN Jianping;XIA Xin;YANG Chunhua(School of Civil Engineering,Chongqing Jiaotong University,Chongqing 400074,China;Surveying and Mapping Institute of Chongqing,National Bureau of Surveying and Mapping Geographic Information,Chongqing 400074,China)

机构地区:[1]重庆交通大学土木工程学院,重庆400074 [2]国家测绘地理信息局重庆测绘院,重庆400074

出  处:《测绘科学》2019年第1期104-111,共8页Science of Surveying and Mapping

基  金:卫星测绘技术与应用国家测绘地理信息局2017年重点实验室项目(KLSMTA-201705)

摘  要:针对藏南地区地表覆盖信息获取困难、生产效率低等问题,该文提出一种结合特征组与分层策略的高空间分辨率遥感图像逐层分类方法。依据地理国情普查地表覆盖信息分类标准,结合研究区遥感图像实际的地表覆盖状况,充分挖掘高空间分辨率遥感图像的地物特征信息,建立该地区地物图像特征组,并采用分层策略减弱地物间的相互影响程度,实现研究区遥感图像逐层分类。实验结果表明,文中提出的逐层分类方法能有效提高地表覆盖类别解译效率,分层成果有利于进一步的精细分类。Aiming at the problems of difficulties in obtaining information from surface coverage and low production efficiency in the area of Zangnan,the article proposed a high spatial resolution remote sensing image classification method combined feature groups and hierarchy strategy.According to the classification criteria of land cover information for geographical national census,this paper combined with the actual surface cover of remote sensing image in study area,fully excavated the feature information of remote sensing images with high spatial resolution to establish the classification feature group in the region,then developed a hierarchical feature extraction strategy to reduce the degree of interrelationship between ground objects,and to realize the layer-by-layer classification of remote sensing images in the study area.Finally,the results showed the classification method and strategy proposed in this paper could effectively improve the efficiency of classification,which could be better for the fine classification.

关 键 词:特征组 分层策略 逐层分类 高空间分辨率 地理国情普查 

分 类 号:P228[天文地球—大地测量学与测量工程]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象