研究目的
To propose a data fusion method for atmospheric correction of satellite images that includes uncertainty information from different data sources, ensuring consistency across different satellite image tiles and sensors.
研究成果
The proposed atmospheric correction method effectively incorporates prior information and uncertainty propagation, providing consistent corrections across different sensors and image tiles. Validation against AERONET measurements shows high correlation, indicating the method's potential for accurate atmospheric correction.
研究不足
The method assumes an isotropic land surface and does not correct for adjacent and terrain effects. A continental aerosol model is used for all sites, which may introduce bias in aerosol value retrieval.
1：Experimental Design and Method Selection:
Utilizes MODIS observations for BRDF description of the earth surface and ECMWF CAMS Near-real-time data for atmospheric states prior information.
2：Sample Selection and Data Sources:
Uses Sentinel-2 TOA reflectance and MODIS MCD43 BRDF products.
3：List of Experimental Equipment and Materials:
Includes MODIS and Sentinel-2 satellite data, ECMWF CAMS Near-real-time data.
4：Experimental Procedures and Operational Workflow:
Applies PSF modeling and spectral mapping to Sentinel-2 TOA reflectance and MODIS simulated surface reflectance, then solves for atmospheric parameters using a Bayesian context.
5：Data Analysis Methods:
Uses Gaussian process emulation for radiative transfer modeling and uncertainty propagation.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
