【摘要】 以Landsat8数据和"珞珈一号"夜光数据为主要数据源,将遥感与地理信息科学技术相结合,对合肥建成区进行提取。首先将Landsat8相应波段分别与自身第8波段进行Gram-Schmidt变换处理,并对处理获取的新波段进行波段计算,获取土壤调节植被指数(SAVI)、改进的归一化差异水体指数(MNDWI)和改进的归一化裸露指数(MNDBI),分别代表植被、水体、建成区三类主要的土地覆盖类型,将三指数波段进行合成,通过监督分类获取整个合肥市的建成区分类图。最后运用"珞珈一号"夜光数据采用二分迭代法获取的建成区边界去除建成区分类图中远离城区的建成区和城区周边被错分为建成区的裸地或农田,获得精准的合肥市建成区。结果表明GS变化后的"三指数法"影像监督分类的精度有了较大的提升,可达93.39%,Kappa系数达到0.8670,同时"珞珈一号"夜光数据可有效地去除建成区周边裸地,使得精度提高到95.10%,Kappa系数也达到了0.9010。更多还原

【Abstract】 With Landsat8 data and “Luojia No.1” luminous data as the main data sources, remote sensing, geographic information science and technology are combined together to extract the built-up area of Hefei.First, the corresponding band of Landsat8 and its own 8 th band are processed by Gram-Schmidt transformation, and the new band obtained by the processing is calculated to obtain the soil adjustment vegetation index(SAVI).The modified normalized difference water index(MNDWI) and the modified normalized difference bare-land index(MNDBI) represent three main land cover types of vegetation, water bodies, and built-up areas.The three index bands are combined to obtain the built-up area classification map of the entire Hefei city through supervision and classification.Finally, using the “Luojia No.1” luminous data, the boundary of the built-up area obtained by the dichotomous iterative method is used to remove the built-up area far from the urban area and the bare land or farmland that is mistakenly divided into the built-up area in the built-up area classification map to obtain an accurate Hefei city built-up area.The results show that the accuracy of the “three-index method” image supervised classification after the GS change has been greatly improved, reaching 93.39%,and the Kappa coefficient reaching 0.8670;at the same time, the luminous data of “Luojia No.1” can effectively remove naked surrounding areas in built-up areas.The accuracy has been improved to 95.10%,and the Kappa coefficient has also reached 0.9010.更多还原