On the Generalization Ability of Data-Driven Models in the Problem of Total Cloud Cover Retrieval
Total Cloud Cover (TCC) retrieval from ground-based optical imagery is a problem that has been tackled by several generations of researchers. The number of human-designed algorithms for the estimation of TCC grows every year. However, there has been no considerable progress in terms of quality, most...
Published in: | Remote Sensing |
---|---|
Main Authors: | , , , , , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
MDPI AG
2021
|
Subjects: | |
Online Access: | https://doi.org/10.3390/rs13020326 https://doaj.org/article/fc50fc8eee494c24b72d8a590a41c655 |
id |
ftdoajarticles:oai:doaj.org/article:fc50fc8eee494c24b72d8a590a41c655 |
---|---|
record_format |
openpolar |
spelling |
ftdoajarticles:oai:doaj.org/article:fc50fc8eee494c24b72d8a590a41c655 2024-01-07T09:42:02+01:00 On the Generalization Ability of Data-Driven Models in the Problem of Total Cloud Cover Retrieval Mikhail Krinitskiy Marina Aleksandrova Polina Verezemskaya Sergey Gulev Alexey Sinitsyn Nadezhda Kovaleva Alexander Gavrikov 2021-01-01T00:00:00Z https://doi.org/10.3390/rs13020326 https://doaj.org/article/fc50fc8eee494c24b72d8a590a41c655 EN eng MDPI AG https://www.mdpi.com/2072-4292/13/2/326 https://doaj.org/toc/2072-4292 doi:10.3390/rs13020326 2072-4292 https://doaj.org/article/fc50fc8eee494c24b72d8a590a41c655 Remote Sensing, Vol 13, Iss 2, p 326 (2021) total cloud cover all-sky camera algorithms assessment neural networks machine learning data-driven approach Science Q article 2021 ftdoajarticles https://doi.org/10.3390/rs13020326 2023-12-10T01:44:33Z Total Cloud Cover (TCC) retrieval from ground-based optical imagery is a problem that has been tackled by several generations of researchers. The number of human-designed algorithms for the estimation of TCC grows every year. However, there has been no considerable progress in terms of quality, mostly due to the lack of systematic approach to the design of the algorithms, to the assessment of their generalization ability, and to the assessment of the TCC retrieval quality. In this study, we discuss the optimization nature of data-driven schemes for TCC retrieval. In order to compare the algorithms, we propose a framework for the assessment of the algorithms’ characteristics. We present several new algorithms that are based on deep learning techniques: A model for outliers filtering, and a few models for TCC retrieval from all-sky imagery. For training and assessment of data-driven algorithms of this study, we present the Dataset of All-Sky Imagery over the Ocean (DASIO) containing over one million all-sky optical images of the visible sky dome taken in various regions of the world ocean. The research campaigns that contributed to the DASIO collection took place in the Atlantic ocean, the Indian ocean, the Red and Mediterranean seas, and the Arctic ocean. Optical imagery collected during these missions are accompanied by standard meteorological observations of cloudiness characteristics made by experienced observers. We assess the generalization ability of the presented models in several scenarios that differ in terms of the regions selected for the train and test subsets. As a result, we demonstrate that our models based on convolutional neural networks deliver a superior quality compared to all previously published approaches. As a key result, we demonstrate a considerable drop in the ability to generalize the training data in the case of a strong covariate shift between the training and test subsets of imagery which may occur in the case of region-aware subsampling. Article in Journal/Newspaper Arctic Arctic Ocean Directory of Open Access Journals: DOAJ Articles Arctic Arctic Ocean Indian Remote Sensing 13 2 326 |
institution |
Open Polar |
collection |
Directory of Open Access Journals: DOAJ Articles |
op_collection_id |
ftdoajarticles |
language |
English |
topic |
total cloud cover all-sky camera algorithms assessment neural networks machine learning data-driven approach Science Q |
spellingShingle |
total cloud cover all-sky camera algorithms assessment neural networks machine learning data-driven approach Science Q Mikhail Krinitskiy Marina Aleksandrova Polina Verezemskaya Sergey Gulev Alexey Sinitsyn Nadezhda Kovaleva Alexander Gavrikov On the Generalization Ability of Data-Driven Models in the Problem of Total Cloud Cover Retrieval |
topic_facet |
total cloud cover all-sky camera algorithms assessment neural networks machine learning data-driven approach Science Q |
description |
Total Cloud Cover (TCC) retrieval from ground-based optical imagery is a problem that has been tackled by several generations of researchers. The number of human-designed algorithms for the estimation of TCC grows every year. However, there has been no considerable progress in terms of quality, mostly due to the lack of systematic approach to the design of the algorithms, to the assessment of their generalization ability, and to the assessment of the TCC retrieval quality. In this study, we discuss the optimization nature of data-driven schemes for TCC retrieval. In order to compare the algorithms, we propose a framework for the assessment of the algorithms’ characteristics. We present several new algorithms that are based on deep learning techniques: A model for outliers filtering, and a few models for TCC retrieval from all-sky imagery. For training and assessment of data-driven algorithms of this study, we present the Dataset of All-Sky Imagery over the Ocean (DASIO) containing over one million all-sky optical images of the visible sky dome taken in various regions of the world ocean. The research campaigns that contributed to the DASIO collection took place in the Atlantic ocean, the Indian ocean, the Red and Mediterranean seas, and the Arctic ocean. Optical imagery collected during these missions are accompanied by standard meteorological observations of cloudiness characteristics made by experienced observers. We assess the generalization ability of the presented models in several scenarios that differ in terms of the regions selected for the train and test subsets. As a result, we demonstrate that our models based on convolutional neural networks deliver a superior quality compared to all previously published approaches. As a key result, we demonstrate a considerable drop in the ability to generalize the training data in the case of a strong covariate shift between the training and test subsets of imagery which may occur in the case of region-aware subsampling. |
format |
Article in Journal/Newspaper |
author |
Mikhail Krinitskiy Marina Aleksandrova Polina Verezemskaya Sergey Gulev Alexey Sinitsyn Nadezhda Kovaleva Alexander Gavrikov |
author_facet |
Mikhail Krinitskiy Marina Aleksandrova Polina Verezemskaya Sergey Gulev Alexey Sinitsyn Nadezhda Kovaleva Alexander Gavrikov |
author_sort |
Mikhail Krinitskiy |
title |
On the Generalization Ability of Data-Driven Models in the Problem of Total Cloud Cover Retrieval |
title_short |
On the Generalization Ability of Data-Driven Models in the Problem of Total Cloud Cover Retrieval |
title_full |
On the Generalization Ability of Data-Driven Models in the Problem of Total Cloud Cover Retrieval |
title_fullStr |
On the Generalization Ability of Data-Driven Models in the Problem of Total Cloud Cover Retrieval |
title_full_unstemmed |
On the Generalization Ability of Data-Driven Models in the Problem of Total Cloud Cover Retrieval |
title_sort |
on the generalization ability of data-driven models in the problem of total cloud cover retrieval |
publisher |
MDPI AG |
publishDate |
2021 |
url |
https://doi.org/10.3390/rs13020326 https://doaj.org/article/fc50fc8eee494c24b72d8a590a41c655 |
geographic |
Arctic Arctic Ocean Indian |
geographic_facet |
Arctic Arctic Ocean Indian |
genre |
Arctic Arctic Ocean |
genre_facet |
Arctic Arctic Ocean |
op_source |
Remote Sensing, Vol 13, Iss 2, p 326 (2021) |
op_relation |
https://www.mdpi.com/2072-4292/13/2/326 https://doaj.org/toc/2072-4292 doi:10.3390/rs13020326 2072-4292 https://doaj.org/article/fc50fc8eee494c24b72d8a590a41c655 |
op_doi |
https://doi.org/10.3390/rs13020326 |
container_title |
Remote Sensing |
container_volume |
13 |
container_issue |
2 |
container_start_page |
326 |
_version_ |
1787422888099315712 |