Retrieval of Cloud Optical Thickness from Sky-View Camera Images using a Deep Convolutional Neural Network based on Three-Dimensional Radiative Transfer

Ryosuke Masuda,Hironobu Iwabuchi,Alessandro Damiani,Konrad Sebastian Schmidt,Rei Kudo

doi:10.3390/rs11171962

Ryosuke Masuda, Hironobu Iwabuchi + Show 3 more

Open Access

https://doi.org/10.3390/rs11171962

Copy DOI

Abstract

Observation of the spatial distribution of cloud optical thickness (COT) is useful for the prediction and diagnosis of photovoltaic power generation. However, there is not a one-to-one relationship between transmitted radiance and COT (so-called COT ambiguity), and it is difficult to estimate COT because of three-dimensional (3D) radiative transfer effects. We propose a method to train a convolutional neural network (CNN) based on a 3D radiative transfer model, which enables the quick estimation of the slant-column COT (SCOT) distribution from the image of a ground-mounted radiometrically calibrated digital camera. The CNN retrieves the SCOT spatial distribution using spectral features and spatial contexts. An evaluation of the method using synthetic data shows a high accuracy with a mean absolute percentage error of 18% in the SCOT range of 1–100, greatly reducing the influence of the 3D radiative effect. As an initial analysis result, COT is estimated from a sky image taken by a digital camera, and a high correlation is shown with the effective COT estimated using a pyranometer. The discrepancy between the two is reasonable, considering the difference in the size of the field of view, the space–time averaging method, and the 3D radiative effect.

Highlights

Clouds primarily modulate the Earth's energy balance and climate
The purpose of this study is to develop a practical, efficient method that can be applied for the estimation of the spatial distribution of cloud optical thickness (COT) from images taken by a ground-based digital camera
The cloud spatial structure at various horizontal scales was trained by the convolutional neural network (CNN), capturing that the COT is large at the center of the cloud and small at the edges

Summary

Introduction

Clouds primarily modulate the Earth's energy balance and climate. The observation of the spatial distribution of cloud physical properties, in particular, cloud optical thickness (COT), is useful for the evaluation of clouds’ radiative effects and the diagnosis of photovoltaics, which is a promising form of renewable energy [1]. Many methods for the estimation of COT by optical measurement from satellites and the ground have been developed. Niple et al [6] estimated COT using the equivalent width of the oxygen absorption band In these methods, the COT in the zenith direction is estimated using the transmitted radiance in the zenith direction. Mejia et al [7] developed the radiance red-to-blue ratio (RRBR) method, which combines the ratio of the transmitted radiance of the red and blue channels (red-to-blue ratio; RBR) and the transmitted radiance of the red channel. They estimated the COT of each pixel from a camera image

Objectives

Results

Conclusion