一种应用于卫星中红外遥感的辐射传输简化计算方案

A simplified parameterization scheme of mid-infrared radiative transfer for satellite remote sensing

作为可见光-近红外和热红外的连接谱段,中红外(MIR)兼具短波辐射的反射特性和长波辐射的发射特性,对于温度探测、目标识别等应用场景有十分重要的意义。目前,MIR谱段辐射传输(RT)计算仍然面临背景参数输入多且不易获取的问题。基于此,提出了一种应用于卫星MIR遥感的RT简化计算方案。使用MODTRAN模型(MM),对MIR白天RT过程中所涉及的背景地表和大气参数定量模拟分析,得到影响RT的关键参数,据此形成一套基于MM的参数简化方案(SS)。在MODIS 6个MIR通道的中心波长处,对SS和MM计算的总辐亮度做对比验证,RMSE均小于0.004399,具有较高精度。SS仅依赖于表面温度、表面发射率、大气廓线类型、水汽和云光学厚度,不需要输入气溶胶、CO_(2)和O_(3)数据。相比MM,SS将输入参数从8类降低到5类,计算的时间效率提升了9.02%。在有限的计算资源下,本文所提的SS能够为MIR图像快速处理、大场景遥感仿真等应用领域提供支撑。

As the connecting spectral band of visible-near infrared and thermal infrared,mid-infrared(MIR)combines the reflective characteristics of short-wave radiation and the emissive characteristics of long-wave radiation,which is of great importance for application scenarios such as temperature detection and target identification.At present,the radiative transfer(RT)calculation of the MIR spectrum still faces the problem of many input background parameters which are not easy to obtain.Based on this,a simplified parameterization scheme of MIR RT for satellite remote sensing is proposed.Using the MODTRAN model(MM),the background surface and atmospheric parameters involved in the MIR daytime RT process were quantitatively simulated and analyzed to obtain the key parameters affecting RT,from which an MM-based parameter simplified scheme(SS)was developed.The total radiance calculated by SS and MM is compared and validated at the center wavelength of 6 MIR channels of MODIS,and the RMSE is less than 0.004399 with high accuracy.The SS only relies on surface temperature,surface emissivity,atmospheric profile type,water vapor and cloud optical thickness,and does not require input of aerosol,CO_(2)and O_(3)data.Compared with the MM,the SS reduces the input parameters from 8 categories to 5 categories,and the calculation time efficiency increases by 9.02%.With limited computational resources,the SS proposed in this paper can provide support for the application areas such as fast processing of MIR images and remote sensing simulation of large scenes.