Estimating the model parameters for remote sensing reflectance pixel by pixel: a neural network approach for optically deep waters

ABSTRACT

Remote sensing reflectance (R_rs) and inherent optical properties (IOPs) conversions are fundamental in accurate satellite measurements to guarantee the semi-analytical retrieval quality of IOPs and the biogeochemical products. Traditionally, the R_rs-IOPs conversions are determined by a quadratic polynomial function with two model parameters (G_x = 0,1). However, G_x values vary in time and location, which are attributed to the spatial and temporal variability inherent to illumination conditions, sea surface properties, and meteorological states. To improve the performance of three classical existing models used for R_rs-IOPs conversions, we designed two novel neural network models (NNG_x = 1,2) to quantitatively calculate G_x from the R_rs spectrum pixel by pixel without requirement of any auxiliary illumination and meteorological data, and then proposed for R_rs-IOPs conversions. We evaluated these approaches with numerical simulations and field measurements, and the results show that the NNG_x models are more effective in semi-analytically converting R_rs into IOPs than the three existing models. Furthermore, we applied the NNG_x models to satellite images to understand the downstream influence of the G_x values on IOPs estimates for the global oceans. We further confirm that the G_x values dramatically change for the global ocean, which is especially true for very oligotrophic gyres, coastal waters, and high latitude oceans. When we use a constant G_x for the R_rs-IOPs conversions, it leads to substantial uncertainty of up to 30% in the IOPs retrievals for China’s coastal regions. Our results suggest that it is possible to improve the data quality of IOPs for the global oceans by providing accurate pixel-level G_x values using NNG_x models.

KEYWORDS:

• Neural network
• inherent optical properties
• remote sensing reflectance
• semi-analytic inversion algorithm

Acknowledgements

The International Cooperation in Science and Technology Innovation among Governments (2019YFE0127200, Chen), National Natural Science Foundation of China (42022045, Chen), and Shan’xi Key Research and Development Program (2022ZDLSF06-09, Chen) provided financial support for this study. The MATLAB code for NNG_x model is shared in Baidu Netdisk (the website is https://pan.baidu.com/s/1k-sPJdwJ7mYn205NSq2lcw?pwd=1234, while the password is 1234).

Disclosure statement

No potential conflict of interest was reported by the author(s).