Yugra State University BulletinYugra State University BulletinВестник Югорского государственного университета1816-92282078-9114Yugra State University63038810.18822/byusu20240122-28MATHEMATICAL MODELING AND INFORMATION TECHNOLOGIESМАТЕМАТИЧЕСКОЕ МОДЕЛИРОВАНИЕ И ИНФОРМАЦИОННЫЕ ТЕХНОЛОГИИResearch ArticleThe comparison of cloud and shadow segmentation algorithms on satellite imagesСравнение алгоритмов сегментации облаков и их теней на космических снимкахSokolkovOleg I.СоколковОлег Игоревич
Russian Federation
sokol.oleg2012@yandex.ruUgra Research Institute of Information TechnologiesЮгорский научно-исследовательский институт информационных технологий2304202420122281604202416042024Copyright ©; 2024, Yugra State UniversityCopyright ©; 2024, Югорский государственный университет2024Yugra State UniversityЮгорский государственный университетhttps://creativecommons.org/licenses/by-sa/4.0https://vestnikugrasu.org/byusu/article/view/630388

Subject of research: the segmentation algorithms of satellite images.

Purpose of research: to compare cloud and cloud shadow segmentation algorithms.

Methods and objects of research: the calculation and comparison of efficiency metrics, labeled space images (CloudSEN12), Fmask, Kappamask, Sen2cloudless, Ukis-csmask, Mobile-Unet algorithms, Sentinel mission cloud segmentation, Sen2cor scene classification, FC-CNN.

Main results of research: the Precision, Recall, Accuracy, F1 metrics have been calculated for the algorithms under consideration. The best result was demonstrated by Mobile-Unet with a score of 0.888 on the F1 metric. The novelty of the obtained results lies in expanding the context of comparative analysis of previous similar studies: we add another algorithm to it (Ukis-csmask).

Предмет исследования: алгоритмы сегментации спутниковых снимков.

Цель исследования: сравнение алгоритмов сегментации облаков и облачных теней.

Методы и объекты исследования: вычисление и сравнение метрик эффективности, размеченные космические снимки (CloudSEN12), алгоритмы Fmask, Kappamask, Sen2cloudless, Ukis-csmask, Mobile-Unet, сегментация облаков миссии Sentinel, классификация сцены Sen2cor, FC-CNN.

Основные результаты исследования: для рассматриваемых алгоритмов вычислены метрики Precision, Recall, Accuracy, F1. Лучший результат продемонстрировал Mobile-Unet с оценкой 0,888 по метрике F1. Новизна полученных результатов заключается в расширении контекста сравнительного анализа предыдущих исследований аналогичного рода: мы добавляем к нему еще один алгоритм (Ukis-csmask).

segmentationcloudsshadowsmachine learningSentinel-2CloudSEN12Mobile-UnetFmaskKappamaskSen2corSen2cloudlessUkis-csmaskFC-CNNсегментацияоблакатенимашинное обучениеSentinel-2CloudSEN12Mobile-UnetFmaskKappamaskSen2corSen2cloudlessUkis-csmaskFC-CNNShi Qiu, Zhe Zhu, Binbin He. Fmask 4.0: Improved cloud and cloud shadow detection in Landsats 4-8 and Sentinel-2 imagery // Remote Sensing of Environment. – 2019, No. 231. – 67 p.Sergii V. Skakun, Jan Wevers, Carsten Brockmann, Georgia Doxani, et al. Cloud Mask Intercomparison eXercise (CMIX): An evaluation of cloud masking algorithms for Landsat 8 and Sentinel-2 // Remote Sensing of Environment. – 2022, Vol. 274, No. 1. – 22 p.Cesar Aybar, Luis Ysuhuaylas, Jhomira Loja, Karen Gonzales, et al. CloudSEN12, a global dataset for semantic understanding of cloud and cloud shadow in Sentinel-2 // Scientific Data. – 2022. – Vol 9, No. 1. – 17 p.Katelyn Tarrio, Xiaojing Tang, Jeffrey G Masek, Martin Claverie, et al. Comparison of cloud detection algorithms for Sentinel-2 imagery // Sci Remote Sens. – 2020. – URL: https://www.sciencedirect.com/science/article/pii/S2666017220300092Sentinel-2. ESA’s Optical High-Resolution Mission for GMES Operational Services, URL: https://sentinel.esa.int/documents/247904/349490/S2_SP-1322_2.pdfLevel-1C Algorithms and Products / Cloud Masks. URL: https://sentinel.esa.int/web/sentinel/technical-guides/sentinel-2-msi/level-1c/cloud-masksSentinel Hub’s cloud detector for Sentinel-2 imagery. URL: https://github.com/sentinel-hub/sentinel2-cloud-detector?ysclid=lil7s21ueq330887295 – Загл. с экрана.Magdalena Main-Knorn, Bringfried Pflug, Jerome Louis, Debaecker, Jerome Louis, et al. Sen2Cor for sentinel-2. In Image and Signal Processing for Remote Sensing // International Society for Optics and Photonics. – 2017. – Vol. 10427. – p. 1042704. –URL: https://www.researchgate.net/publication/320231869_Sen2Cor_for_Sentinel-2Marharyta Domnich, Indrek Sünter. KappaMask: AI-Based Cloudmask Processor for Sentinel-2 // Remote Sensing. – 2021. – Vol. 13, No. 20. – 22 p.Zongwei Zhou, Md Mahfuzur Rahman Siddiquee, Nima Tajbakhsh, Jianming Liang. UNet++: A Nested U-Net Architecture for Medical Image Segmentation // 4th Deep Learning in Medical Image Analysis (DLMIA) Workshop. – 2018. – 11 p.Junfeng Jing, Zhen Wang, Matthias Rätsch, Huanhuan Zhang. Mobile-Unet: An efficient convolutional neural network for fabric defect detection // Textile Research Journal. – 2020. – Vol. 92, No 12. – 17 p.CloudSEN12: A global dataset for cloud and cloud shadow semantic understanding. URL: https://github.com/cloudsen12/examples?ysclid=lj2iwsvofs700104227 – Загл. с экрана.M. Wieland, Yu Li, S. Martinis. Multi-sensor cloud and cloud shadow segmentation with a convolutional neural network // Remote Sensing of Environment. – 2019. – No. 230. – 22 p.Ukis-csmask. URL: https://github.com/dlr-eoc/ukis-csmask?ysclid=lj2iaa2xsz492603360 – Загл. с экрана.