Article
Environmental Sciences
Nivedita Sanwlani, Reshmi Das
Summary: This study derived the mineral composition of airborne dust aerosols using satellite optical measurements and compared it with chemically analyzed elemental concentrations. The results showed a high correlation between the derived mineral concentrations and the elemental concentrations, and the derived mineral concentrations were consistent with other monitoring data.
Article
Environmental Sciences
Muhammad Bilal, Md. Arfan Ali, Janet E. Nichol, Max P. Bleiweiss, Gerrit de Leeuw, Alaa Mhawish, Yuan Shi, Usman Mazhar, Tariq Mehmood, Jhoon Kim, Zhongfeng Qiu, Wenmin Qin, Majid Nazeer
Summary: Numerous studies have attempted to classify aerosols into different types using only AOD and AE, but these parameters do not provide enough information for accurate classification. A new approach called AEROSA is introduced to provide information on aerosol amount and size based on AOD and AE, without claiming specific aerosol types. AEROSA can accommodate variations in location and season, unlike the traditional GA aerosol types.
FRONTIERS IN ENVIRONMENTAL SCIENCE
(2022)
Article
Environmental Sciences
Hassan Bencherif, Aziza Bounhir, Nelson Begue, Tristan Millet, Zouhair Benkhaldoun, Kevin Lamy, Thierry Portafaix, Fouad Gadouali
Summary: This study investigates aerosol distributions and a Sahara dust-storm event that occurred in the South of Morocco in early August 2018. It shows that aerosol populations in southern Morocco are dominated by Saharan desert dust, especially during the summer season. The study also uses the HYSPLIT model to simulate air-mass back-trajectories during the event.
Article
Environmental Sciences
Filipe Aires, Eulalie Boucher, Victor Pellet
Summary: Traditional Neural Networks have been widely used in satellite remote sensing for the past 25 years. However, new neural architectures like Convolutional Neural Networks have shown great potential in high resolution image processing tasks such as classification and segmentation. The study demonstrates that CNNs can be beneficial for coarse resolution instruments under specific conditions, such as when the variable being detected has spatial coherence and instrument noise exceeds a certain threshold at the pixel scale.
REMOTE SENSING OF ENVIRONMENT
(2021)
Article
Geochemistry & Geophysics
Xu Huang, Yuxi Sun, Shanshan Feng, Yunming Ye, Xutao Li
Summary: This study proposes a novel model called ECR-CAM neural network for remote sensing scene classification tasks. The model, consisting of an encoder, classifier, reconstruction, and CAM module, can accurately locate target objects and provide precise visual explanations.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Engineering, Aerospace
Stavros Kolios
Summary: The study presents a methodological approach for hail detection using a Deep Neural Network (DNN) model trained with Meteosat multispectral infrared (IR) imagery. The accuracy of the DNN model was found satisfactory, and it was successfully applied in two case studies. The operational mode of the model allows for hail detection using modern meteorological satellites worldwide.
ADVANCES IN SPACE RESEARCH
(2023)
Article
Computer Science, Information Systems
Gabriel Villavicencio Arancibia, Osvaldo Pina Bustamante, Gabriel Hermosilla Vigneau, Hector Allende-Cid, Gonzalo Suazo Fuentelaba, Victor Araya Nieto
Summary: Chile is one of the major copper producers in the world, generating 1.7 million tons of tailings per day. Over the past two decades, the mining industry has been increasingly using thickened tailings dams (TTD) due to their advantages in water resource recovery, environmental impact reduction, and stability. A proposed intelligent system uses machine learning algorithms based on Artificial Neural Networks, Support Vector Machine, and Random Forest to estimate in-situ states and moisture content in TTD with high accuracy.
Article
Geochemistry & Geophysics
Xuyang Bai, Shurun Tan
Summary: This article proposes a novel physics-embedded artificial neural network (P-ANN) inversion algorithm for retrieving the vertical distribution of soil moisture and temperature using multichannel passive microwave observations. The P-ANN approach outperforms traditional optimization algorithms and conventional neural network approaches in dealing with layered soil retrieval. The proposed approach holds great potential for remote sensing applications and solving inverse problem challenges.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Environmental Sciences
Li Shen, Yao Lu, Hao Chen, Hao Wei, Donghai Xie, Jiabao Yue, Rui Chen, Shouye Lv, Bitao Jiang
Summary: This study introduces a building-change-detection dataset named S2Looking, which consists of large-scale side-looking satellite images and tens of thousands of annotated change instances, for training deep-learning algorithms. The dataset offers larger viewing angles, illumination variances, and complexity of rural images compared to existing datasets, and preliminary tests suggest higher level of challenges for deep-learning algorithms.
Article
Computer Science, Information Systems
Long Chen, Feilong Tang, Xu Li, Jiacheng Liu, Yanqin Yang, Jiadi Yu, Yanmin Zhu
Summary: This paper investigates how to optimize the delay in remote sensing satellite networks based on cooperation transmission. It proposes a cooperation capability model and a delay-optimal cooperation transmission scheme, which show effective and efficient performance in simulations.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Article
Environmental Sciences
Waytehad Rose Moskolai, Wahabou Abdou, Albert Dipanda, Kolyang
Summary: This paper explores the application of deep learning techniques to satellite image time series (SITS) prediction, discussing the design and evaluation elements of predictive models, various DL models, satellite image characteristics, and major applications. It also addresses the limitations and proposed solutions related to using DL for SITS prediction.
Article
Engineering, Marine
Evdokia Tapoglou, Rodney M. Forster, Robert M. Dorrell, Daniel Parsons
Summary: This study integrates machine learning and satellite remote sensing, combining C-band Synthetic Aperture Radar images with buoy data to provide detailed predictions of significant wave height (SWH) through artificial neural networks. The method allows for high-resolution spatial distribution information of wave height, demonstrating the impact of windfarm infrastructure on wave propagation and potential improvement in sea state prediction accuracy, hotspot identification, and maintenance job prioritization for wind turbines.
Article
Computer Science, Information Systems
Islombek Mirpulatov, Svetlana Illarionova, Dmitrii Shadrin, Evgeny Burnaev
Summary: Satellite data is a valuable resource for remotely solving challenging tasks like environmental monitoring and hazard evaluation. However, the quality and quantity of annotated datasets remain a sticking point in remote sensing studies. This research proposes a pipeline for improving the accuracy and resolution of markup in land-cover and land-use segmentation tasks using data from the Sentinel-2 satellite. The methodology combines classical machine learning and deep learning algorithms to achieve higher results compared to using raw inaccurate data.
Article
Computer Science, Information Systems
Zhichao Chen, Jie Yang, Zhicheng Feng, Lifang Chen
Summary: This study presents a lightweight remote sensing scene classification model called RSCNet. By using lightweight neural networks and efficient feature extraction and classification mechanisms, it achieves efficient, accurate, and real-time remote sensing scene classification.
Article
Geochemistry & Geophysics
Frank S. Marzano, Michele Iacobelli, Massimo Orlandi, Domenico Cimini
Summary: This study explores the retrieval of chlorophyll-a and total suspended matter along the coast of Italy using various methods, finding that neural network methods are more accurate at a regional scale. The study also identifies limitations of the bio-optical model in representing optical properties in different coastal areas.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
