摘要
高景一号是国内首个具备高敏捷、多模式成像能力的商业卫星星座,为探讨高景一号高空间分辨率遥感数据对特高压山区施工造成水土流失信息提取能力,基于面向对象图像分析方法,通过多次试验确定分割尺度,对影像光谱特征、指数特征、纹理特征和对象面积特征分析优选,建立研究区水土流失信息提取规则集,实现对水土流失信息空间分布的准确提取。为凸显高景一号遥感影像面向对象方法相对于像元级分类提取水土流失信息更具有优势,利用决策树算法对研究区进行了像元级水土流失信息提取。结果表明:面向对象方法提取特高压建设施工引起的水土流失信息精度高,且面向对象方法优于像元级水土流失信息提取精度。
Super View-1 is the first commercial satellite constellation in China with high agility and multi-mode imaging capability. In order to discuss the extraction ability of Super View-1 high spatial resolution remote sensing data on soil erosion caused by construction of ultra-high voltage in mountainous areas,the object-oriented image analysis method is used to determine the segmentation scale through multiple experiments and analyze and optimize the spectral features,exponential features,texture features and covering characteristics of the object.Then a set of rules on soil erosion information extraction in the study area have been established and a precise extraction of water and soil loss information have been achieved. For the purpose of highlighting the advantages of object-oriented method of the Super View-1 remote sensing image on extracting the soil and water loss information comparing to the pixel-level classification method,a decision-tree algorithm is used to extract the soil erosion information in the study area. The results show that the object-oriented method extracts soil erosion information caused by construction of UHV construction with higher accuracy and can get precise soil erosion information than the pixel-level extraction method.
作者
丁永福
张金德
刘皓
王新元
何宣虎
戚国辉
宋盼盼
刘新
DING Yong-fu;ZHANG Jin-de;LIU Hao;WANG Xin-yuan;HE Xuan-hu;QI Guo-hui;SONG Pan-pan;LIU Xin(State Grid DC Construction Company,Belting 100052;Unisplendour Software System Corporation Limited,Beijing 100084)
出处
《矿产勘查》
2018年第6期1291-1298,共8页
Mineral Exploration
关键词
高景一号
特高压
山区施工
水土流失
面向对象
SuperView- 1
ultra-high voltage
mountain construction
soil erosion
object-oriented
