基于光谱混合的大气吸收通道发射率图像模拟

发布时间：2019-04-04 11:19

【摘要】：提出了一种基于光谱混合的大气强吸收通道地表发射率图像模拟方法,旨在为全链路图像模拟提供地表基础数据。首先,对数据源图像进行端元选择,获取图像端元光谱;然后,通过光谱匹配将地表测量光谱替代图像端元光谱进行丰度反演;最后,基于线性光谱混合模型,结合替代端元在吸收通道的地表测量发射率,实现发射率图像模拟。利用该方法,以Landsat 8 OLI多光谱图像为数据源,并假设JHU和USGS光谱库数据为数据源的地表测量光谱,进行了模拟试验和精度分析。鉴于真实的吸收通道地表发射率图像无法获得,精度分析采用了比较基于该方法模拟的OLI通道反射率图像与真实反射率图像的方式。获得的模拟图像总体均方根误差分别为0.045和0.049,说明模拟方法可行且具有较高的精度。
[Abstract]:A spectral mixing-based surface emissivity image simulation method for atmospheric strong absorption channels is proposed in this paper, which aims to provide basic surface data for all-link image simulation. Firstly, the end-element spectrum of the data source image is selected to obtain the end-element spectrum of the image, and then, the surface measurement spectrum is substituted for the end-element spectrum of the image by spectral matching for the inversion of the abundance. Finally, the emissivity image simulation is realized based on the linear spectral mixing model, combined with the emissivity measurement of the substitute end element in the absorption channel. Using this method, using Landsat 8 OLI multispectral image as data source and assuming JHU and USGS spectral database data as data source, the simulation experiment and precision analysis are carried out. In view of the fact that the real absorption channel surface emissivity image can not be obtained, the method of comparing the OLI channel reflectivity image simulated based on this method with the real reflectivity image is adopted in the precision analysis. The root mean square error (RMS) of the simulated image is 0.045 and 0.049 respectively, which shows that the simulation method is feasible and has high accuracy.
【作者单位】：中国国土资源航空物探遥感中心;中国科学院遥感与数字地球研究所;
【基金】：国家杰出青年科学基金项目(41325004) 国家科技高分辨率对地观测系统重大专项(30-Y20A06-9003-15/16、11-Y20A32-9001-15/17) 国家自然科学基金项目(41271370)
【分类号】：TP751

【相似文献】