Open AccessIntegration model of POSP measurement spatial response function
Author(s) -
Xuefeng Lei,
Shi-Liang Zhu,
Zhenyang Li,
Jin Hong,
Zhenhai Liu,
Fei Tao,
Peng Zou,
Maoxin Song,
Congfei Li
Publication year - 2020
Publication title -
optics express
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.394
H-Index - 271
ISSN - 1094-4087
DOI - 10.1364/oe.393897
Subject(s) - remote sensing , radiance , mean squared error , optics , time delay and integration , sampling (signal processing) , field of view , computer science , algorithm , physics , mathematics , statistics , geology , detector
The particulate observing scanning polarimeter (POSP) measurement spatial response function (SRF) relates to the weighted contribution of each location within the measurement footprint, which is determined by the percentage of the dwell time of each location on the Earth surface to the overall sampling integration time. The SRF resulting from a combination of the equally weighted instantaneous field of view (IFOV) during integration is required for an accurate modeling. Simply using a mean value SRF assuming an equivalent weight at each sampling position instead of the actual SRF will inevitably introduce errors. Considering the data fusion between POSP and high spatial resolution sensors, a discrete integration method that takes the effect of actual weights into account is proposed in this paper. The simulation results of the integral model and the mean value model show that the larger the intensity change in the sampling area covered by the IFOV of the POSP during a single sampling, the more significant the difference between the two results. Meanwhile, the integration SRF is validated by resampling the simultaneous imaging polarization camera (SIPC) data, which is compared with POSP data acquired at the same time in an aerial experiment. The results show that the integration SRF model is more accurate to characterize the details of POSP measurement than the mean value SRF model. The proposed SRF reduces the root mean square error (RMSE) of convolved results and measurements by 5∼30% with different radiance contrast scene.