2m分辨率遥感影像地表覆盖分类误差研究

The classification errors research of 2m resolution remote sensing images covering surface

以贵州省镇远县焦溪镇为研究区,以资源3号(ZY-3)2m分辨率卫星影像为主要数据源,采用面向对象分类技术提取研究区地表覆盖信息,将提取结果与普莱亚(Pleiades)影像,经目视解译并结合实地验证获取的高精度地表覆盖进行对比,分析ZY-3影像提取地表覆盖各类型的精度误差。研究结果表明:ZY-3影像相对于Pleiades数据提取平均总误差为1.56%,其中旱地和建设用地分类误差相对较大,精度误差分别是3.41%和3.08%。而水体分类误差最小,为0.70%,其余地类水田、乔木林、灌木林、草地、裸地提取误差则分别是1.28%、0.53%、0.86%、1.18%、1.47%。

Taking Jiao-Xi town located in Zhenyuan county which is in Guizhou province as the study area,based on the main data of the high resolution remote sensing image of Resources satellite three（ ZY-3） and the Pleiades,and based on the platform of e Cognition developed by Definiens Imaging company,using Object-oriented classification method,this paper respectively extracts the surface coverage from the various high resolution remote sensing image,and compares and analyzes on two extractions,and obtains the precision difference which is the extraction from ZY-3 relative to Pleiades. The results show that extraction average total error of ZY-3 image related to the Pleiades data is 1. 56%,the classification errors of upland and construction land are relatively large,and the ratio is 3. 41% and3. 08% respectively. The water has minimum classification error which is 0. 70%. And the other land use types,such as paddy fields,arbor,shrub,grassland and bare land,the extraction error is respectively 1. 28% 、0. 53% 、0. 86% 、1. 18% 、1. 47%.