高光谱图像大气校正自动化与快速处理

Automation and Fast Processing of Atmospheric Correction for Hyperspectral Data

全谱段宽幅高分辨率推扫式光谱成像仪作为我国新一代航空高光谱成像仪已进入应用校飞阶段,文章针对其TB/日数量级高光谱图像快速处理问题,对大气校正过程的参数自动化设置方法和并行加速方法进行研究。在传统基于辐射传输模型的大气校正基础上,分析了可见近红外和短波红外通道是否波段合并、重合波段优选、大气类型选择、是否水汽反演、水汽吸收波段选择等方面对反射率反演精度的影响,实现了参数优化自动设置,并开发了并行化大气校正算法。以吉林榆树和辽宁辽中的数据进行验证,结果表明,反演反射率与参考真值反射率一致性高,同时处理速度比串行处理大大提高,可为高光谱反射率数据产品的业务化生产提供有力工具。

The Wide Swath and High Resolution Airborne Pushbroom Hyperspectral Imager is a new generation of airborne hyperspectral imager in China.Aiming at the problem of rapid processing of hyperspectral image with large amount of data,the method of parameter automation and parallel acceleration for atmospheric correction is studied.As a new generation of aerial hyperspectral imager in China,the Wide Swath and High Resolution Airborne Pushbroom Hyperspectral Imager has entered the application and flight calibration stage.Based on the atmospheric correction of radiation transmission model,the influences of various factors on the accuracy of reflectance inversion are analyzed,and the parallel atmospheric correction is carried out.After experiments with the imaging hyperspectral data,the reflectance of the inversion and the measured reflectance of the ground are quantitatively evaluated.The results show that the inversion reflectance has high consistency with the reference true reflectance,and the processing speed is much higher than that of serial processing,which can provide a powerful tool for the business production of hyperspectral reflectance data products.