论文部分内容阅读
针对国产风云三系列中分辨率卫星快速有效地进行水体识别的问题,基于2011—2016年数据进行了辽宁省主要湖泊水库的光谱分析,提出了晴空条件和有云情况下分别采用归一化水体指数方法和通道值与归一化水体指数相结合方法进行湖泊水库的提取。结果表明:FY3B/MERSI湖泊光谱曲线具有水体光谱特征,具体表现为8通道的数据反射率最高,18通道数据反射率最低。通过与49景TM数据水体提取结果进行对比,FY3B/MERSI数据水体提取的面积精度达到85%以上,总体分类精度达到90%以上,Kappa系数在0.56~0.95之间。提出了简便、快捷的计算模型,为国产卫星数据的业务应用提供了初步方法。
Based on the data of 2011-2016 for spectral analysis of major lakes and reservoirs in Liaoning Province, aiming at the problem of rapid and effective recognition of water bodies in three series of medium-resolution satellites made in China, the authors propose that normalized water bodies Exponential method and channel value and normalized water body index method for lake reservoir extraction. The results show that the spectral curve of FY3B / MERSI lakes has the spectral characteristics of water body. The results show that the data reflectance of channel 8 is the highest, and that of channel 18 is the lowest. Compared with the result of 49 TM TM data, the area accuracy of FY3B / MERSI data extraction was more than 85%, the overall classification accuracy was more than 90%, and the Kappa coefficient was between 0.56 and 0.95. A simple and quick calculation model is proposed, which provides a preliminary method for the business application of domestic satellite data.