Radiometric Calibration of Large-View-Angle Satellite Sensor at Railroad Valley Playa Site Despite Incomplete RadCalNet Datasets

The radiometric calibration network (RadCalNet) datasets have been widely applied to calibrate the satellite sensor with a narrow field of view and monitor the radiometric performance during the on-orbit life cycle since the RadCalNet datasets have very high radiometric accuracy and temporal resolution. However, only the bottom-of-atmosphere (BOA) and top-of-atmosphere spectral reflectance products are released, which cannot be directly applied for the radiometric calibration of the large-view-angle satellite sensor due to the effect of the bidirectional reflectance distribution function (BRDF) over the RadCalNet sites. Meanwhile, the released RadCalNet datasets have data incompleteness, which reduces the frequencies of radiometric calibration. Therefore, to address these problems, taking the wide field of view (WFV) of the GaoFen-6 (GF-6) satellite as the calibrated sensor, treating the Railroad Valley Playa (RVP) site as the investigated site, this article proposes a radiometric calibration method for the large-view-angle satellite sensor. The time-series MOD09GA datasets over the RVP site are collected from the Google Earth Engine platform to build the BOA BRDF model to correct the BRDF effect of the RVP site with kernel model and the random forest (RF) algorithm. The RF is also used to simulate the incomplete RadCalNet datasets. The official calibration coefficients published by the China Centre for Resources Satellite Data and Application are employed to validate the proposed radiometric calibration results. The results show the proposed method can significantly improve the accuracy of radiometric calibration results for the large-view-angle satellite sensor and implement the radiometric calibration even in case of the incomplete RadCalNet datasets. The relative errors of the eight bands of the WFV sensor are all less than 8.48% . The influences of factors in the proposed method are comprehensively discussed. The total radiometric calibration uncertainty is less than 5.54% .