研究利用CBERS02B卫星的CCD传感器数据反演陆地气溶胶的方法。采用的方法是暗像元法。具体步骤为: 根据地面采集的植被光谱数据, 结合CCD传感器特点, 建立浓密植被(暗像元)红蓝波段(CCD传感器的第三和第一波段)反射率与地表反射率之间的关系, 确定了暗像元识别的阈值, 讨论气溶胶光学厚度对暗像元识别的影响以及消除这种影响的方法; 利用6S进行辐射传输运算, 构建查找表; 根据CBERS02B卫星的CCD传感器数据, 从查找表插值得到气溶胶光学厚度, 并进行了算法的误差分析。用广西南宁市及北京地区附近的两景数据进行了实际的反演试验, 使用MODIS的气溶胶产品与反演结果进行比对。结果显示, CBERS02B卫星的CCD传感器数据能够较好的反演陆地气溶胶。

In this paper, the retrieval of AOD (aerosol optical depth) over land surfaces from CCD (Charge Coupled Device) data of CBERS (China Brazil Earth Resources Satellite) 02B was studied. The inversive method is modified DDV (dark dense vegetation) algorithm: (1) based on the spectral data of vegetation and the characteristics of the CCD sensor, the relationship between blue and red band surface reflectances and the NDVI (Normalized Difference Vegetation Index) thresholds for dense vegetation (dark pixel) were worked out, and the atmospheric effect and its correction were also analyzed. (2) The LUT (look up table) for different bands was pre-computed using the 6S atmospheric radiative transfer code. (3) The AODs was retrieved through interpolating in LUT in accordance with radiances obtained by CCD, and the errors of this modified DDV method was also analyzed and discussed. Finally, the modified DDV method was successfully applied to Nanning and Beijing areas. The comparison of the retrieved AODs with those of MODIS shows a good agreement.