无线电工程

针对高分一号(GF-1)多光谱影像的业务化大气校正难、辅助数据不足等问题,基于6S辐射传输模型,设计并实现了适合GF-1卫星数据大气校正算法与程序。算法以GF-1星1A级影像、元数据及传感器公开参数为输入数据,自动调用6S模型获取表观辐亮度转化为地表反射率的大气系数,得到影像的地表反射率。采用该算法对天津城区和山区的两景影像进行了大气校正,并与FLAASH校正结果进行了对比分析。结果表明,该方法能够较好地去除气溶胶散射和大气分子散射的影响,较好地还原水体、水泥地、植被和裸地等典型地物的光谱反射曲线;校正后的NDVI更有利于地物的识别,与FLAASH校正结果保持良好的一致性;在相同运行环境下,FLAASH和该方法处理时间分别为245,118 s,与FLAASH相比,该方法的大气校正效率提高了53%。研究结果对GF-1多光谱影像业务化大气校正具有一定的参考价值。

Aiming at the problems of the difficulty in the operational atmospheric correction of GF-1 multispectral image and insufficient auxiliary data, an atmospheric correction algorithm and program based on the 6 S model for GF-1 satellite data are designed and implemented.The algorithm takes GF-1 level 1 A image, metadata and sensor public parameters as input data, and automatically calls 6 S model to obtain the atmospheric coefficient of apparent radiance converted to the surface reflectance, finally obtains surface reflectance of the image.Then use the algorithm to calibrate atmospheric images of the urban area and the mountain area of Tianjin, and compare the results with FLAASH correction results.The results show that the method can well remove the effects of aerosol scattering and atmospheric molecular scattering, and effectively restore the spectral reflection curves of typical features such as water body, cement floor, vegetation, and bare ground.The corrected NDVI is more beneficial to distinguish features and maintains good consistency with the FLAASH correction results.Under the same operating environment, the processing time of FLAASH and this method is 245,118 s, respectively.Compared with FLAASH,the atmospheric correction efficiency of this method is improved by 53%.The research results have certain reference value for operational atmospheric correction of GF-1 multispectral image.