Key Atmospheric Correction Technology Based on Hyperspectral Data from Xiguang-1 05 Satellite (Tianxianpeihao) – MODTRAN Model

Background Introduction

In the field of quantitative remote sensing, the key to accurately retrieving the true radiative characteristics of ground objects lies in eliminating the influence of atmospheric path radiation. Atmospheric correction methods based on the MODTRAN model, by constructing an accurate radiative transfer mechanism model, perform inverse calculations on the comprehensive radiance received at the sensor's entrance pupil, thereby eliminating interference from atmospheric absorption and scattering. This is an indispensable core step in obtaining accurate physical parameters of ground objects. This case study uses calibration data from the Xiguang-1 05 satellite (Tianxianpei) received on September 23rd at the Dunhuang calibration field in Gansu Province to retrieve the entrance pupil radiance.

Methods and Principles

Table 1. Payload Parameters of Xiguang-1 05 Satellite (Tianxianpei)

The MODTRAN model is embedded in the PcModWin software. The software uses a graphical user interface to encapsulate complex atmospheric, aerosol, and observational geometric parameters into standardized input modules. Its core idea is to simulate the transmission of electromagnetic waves from the sun to the Earth's surface and then to the sensor by establishing a physical model of atmospheric radiative transfer, thereby accurately calculating the composite radiance reaching the sensor's entrance pupil. The core atmospheric radiative transfer equation of the model is:

In the formula, I represents the radiance, S represents the source function, τ represents the optical depth, and μ represents the direction cosine, all calculated from atmospheric, aerosol, and observation geometry parameters. The formula for calculating the entrance pupil radiance is as follows:

In the formula ε_sLet B(λ,T) be the surface emissivity._s) represents blackbody radiation, T_sρ is the surface temperature, T is the transmittance, and ρ is the surface temperature._sE represents the surface reflectance._sun(λ) represents solar radiation.

The result of the entrance pupil radiance inversion interface of PcModWin is shown in the following figure:

Figure 1 PcModWin model inversion interface

Results Display

The inversion results of the entrance pupil radiance are shown in Figure 2. The inverted entrance pupil radiance curve was resampled to the band of Xiguang-1 05 satellite (Tianxianpei) as shown in Figure 3.

Figure 2 Comparison between the inverted radiance curve and the DN value curve at the test point

Figure 3. Results after inversion radiance resampling

The linear relationship between the entrance pupil radiance after resampling and the DN value at the test point was calculated, and the calibration coefficients are shown in the figure below:

Figure 4. Calculation results of absolute calibration coefficients

Analysis of the radiance inversion results and the calculation results of the absolute calibration coefficients shows that the radiance inversion results have the same trend as the DN curve of the test point, and remain the same as the DN curve of the test point at important features such as the water vapor absorption line. The absolute calibration coefficients also show the same characteristics at the same positions.