Spaceborne camera site calibration experiment based on automatic calibration field

Radiometric calibration is the prerequisite and foundation for the quantitative application of satellite remote sensing data. Accurate calibration of on-orbit radiometric performance directly affects the quantitative accuracy and reliability of subsequent agricultural monitoring, resource exploration, and ecological environment remote sensing products. This paper presents a field calibration experiment of a spaceborne hyperspectral camera based on an automatic calibration field. Utilizing a long-term stable, unattended reflectivity measurement system and synchronous atmospheric parameter monitoring equipment within the automatic calibration field, key parameters such as surface reflectivity, atmospheric optical thickness, and water vapor content were simultaneously acquired during satellite transit. Employing a reflectivity-based calibration method combined with the MODTRAN5 atmospheric radiative transfer model, absolute radiometric calibration of the spaceborne hyperspectral camera in the visible-near-infrared band was successfully achieved.

The experiment selected the Railroad Valley Playa (RVUS) automated calibration field, listed on the RadCalNet website, as the test site. This site is an internationally recognized high-precision radiometric calibration reference field, with advantages such as uniform surface coverage, stable spectral characteristics, and continuous and reliable long-term observation data. Its supporting unattended monitoring system can effectively avoid human intervention errors and fully meet the requirements of high-precision radiometric calibration.

Based on the satellite transit window, this experiment selected satellite transit data at 18:30 UTC on September 17, 2025 for site calibration. The L1B image of the RVUS automatic calibration field on that day is shown in Figure 1, where the red area represents the atmospheric data monitoring range. Two types of core data were simultaneously acquired through the site monitoring system: first, the surface reflectance curve of a 1km×1km monitoring area (corresponding to 28×28 pixels in the satellite image); and second, a complete atmospheric parameter dataset including solar azimuth, aerosol thickness, water vapor content, and ozone concentration (see Table 1).

Figure 1. Satellite-acquired RVUS automatic calibration field imagery

Table 1. Atmospheric parameters of RVUS station at 18:30 (UTC) on September 17, 2025.

The atmospheric parameters in Table 1 and the measured surface reflectance curve (see Figure 2) were input into the MODTRAN5 model. The key model parameter settings are shown in Table 2. The radiance spectrum at the entrance pupil of the satellite camera was simulated by forward modeling, and the radiance spectrum curve was obtained (see Figure 3). The simulated radiance spectrum was resampled to the actual band range of the satellite camera, and the mean DN value of the calibration field region in the camera image was extracted to establish the fitting relationship between the entrance pupil radiance and the DN value (see Figure 4). By calculating the ratio of the two, the absolute radiometric calibration coefficients of each band of the satellite hyperspectral camera were finally obtained (see Figure 5), completing the core calibration calculation.

Figure 2. Surface reflectance curve at 18:30 (UTC) on September 17, 2025.

This study successfully completed the site-specific absolute radiometric calibration of a spaceborne hyperspectral camera using the Railroad Valley Playa (RVUS) automatic calibration field. The experimental results clearly established a quantitative correspondence between entrance pupil radiance and image DN values, accurately obtained radiometric calibration coefficients for each band, and verified the effectiveness and accuracy of the calibration method.

Experiments show that the reflectivity-based calibration method based on automatic calibration fields is simple to operate, stable in the long term, and requires no manual intervention. It can significantly reduce the cost and difficulty of traditional site calibration, providing an efficient and sustainable on-orbit calibration approach for commercial remote sensing satellites.