ELIFAN, an algorithm for the estimation of cloud cover from sky imagers

Abstract

Abstract. In the context of a network of sky cameras installed on atmospheric multi-instrumented sites, we present an algorithm named ELIFAN, which aims to estimate the cloud cover amount from full-sky visible daytime images with a common principle and procedure. ELIFAN was initially developed for a self-made full-sky image system presented in this article and adapted to a set of other systems in the network. It is based on red-to-blue ratio thresholding for the distinction of cloudy and cloud-free pixels of the image and on the use of a cloud-free sky library, without taking account of aerosol loading. Both an absolute (without the use of a cloud-free reference image) and a differential (based on a cloud-free reference image) red-to-blue ratio thresholding are used. An evaluation of the algorithm based on a 1-year-long series of images shows that the proposed algorithm is very convincing for most of the images, with about 97 % of relevance in the process, outside the sunrise and sunset transitions. During those latter periods, however, ELIFAN has large difficulties in appropriately processing the image due to a large difference in color composition and potential confusion between cloud-free and cloudy sky at that time. This issue also impacts the library of cloud-free images. Beside this, the library also reveals some limitations during daytime, with the possible presence of very small and/or thin clouds. However, the latter have only a small impact on the cloud cover estimate. The two thresholding methodologies, the absolute and the differential red-to-blue ratio thresholding processes, agree very well, with departure usually below 8 % except in sunrise–sunset periods and in some specific conditions. The use of the cloud-free image library gives generally better results than the absolute process. It particularly better detects thin cirrus clouds. But the absolute thresholding process turns out to be better sometimes, for example in some cases in which the sun is hidden by a cloud.

Cézeaux-Aulnat-Opme-Puy De Dôme: a multi-site for the long-term survey of the tropospheric composition and climate change

“…A more detailed segregation of BL/FT air masses using a set of measurements and model outputs shows that the size distribution actually keeps the air-mass-type signature even in air masses transported in the FT for more than 75 h (Farah et al, 2018). The frequency of new particle formation events is on average 30 % of the measurement days (Rose et al, 2013;Venzac et al, 2007), which is a relatively high frequency compared to most European low-troposphere environments (Manninen et al, 2010). A key question is to characterize the vertical extension of these new particle formation events.…”

On the Generalization Ability of Data-Driven Models in the Problem of Total Cloud Cover Retrieval

“…Most of these algorithms are designed by experts introducing their understanding of the physical processes resulting in the all-sky imagery similar to the one presented in 1. Given all-sky imagery acquired, in most simple cases, an index is calculated pixel-wise, e.g., red-to-blue ratio (RBR) in the series of papers of Long et al [32,33,41] or in following studies [42][43][44], or the ratio B−R B+R in [36,37,45], or even a set of indices [46]. Then, an empirical threshold is applied for the classification of pixels dividing them into two classes: "cloudy" and "clear sky" ones.…”

“…Then, these human-engineered features are used in some sort of an algorithm for the computation of an index or multiple indices which then are aggregated over the image to a quantitative measure. The algorithm may be considered simple [32,33,36,38,48], adaptive [39,47] or designed to be complex to some extent in attempt of taking some advanced spatial features into account [34,35,41,43,44] or in attempt of correcting the distortion of imagery or other features of imagery that are not consistent with the initial researcher's assumptions [36,43,44,46]. However, all these methods rely on the aggregation step at some point that is commonly implemented as a variation of thresholding.…”

Cloud Occurrence Frequency at Puy de Dôme (France) Deduced from an Automatic Camera Image Analysis: Method, Validation, and Comparisons with Larger Scale Parameters

“…For example, the software is able to define the area of interest by horizon masking. A more optimized CF evaluation was also performed by the software ELIFAN [18], which is used in this study.…”

“…We have also compared the COF deduced from the camera analysis to the CF calculated with the algorithm ELIFAN [18] applied to the all-sky camera images taken at the Cézeaux site from December 2015 to June 2019 ( Figure 8). The ELIFAN CF parameter presents larger variability compared to the local camera COF but follows a similar seasonal evolution, with larger mean values in winter (0.67) than in summer (0.44).…”