Fusing Sentinel-2 Satellite Images and Aerial RGB Images

Author(s):  
Jakob Sigurdsson ◽  
Magnus O. Ulfarsson ◽  
Johannes R. Sveinsson
Author(s):  
Thomas Fisher ◽  
Harry Gibson ◽  
Gholamreza Salimi-Khorshidi ◽  
Abdelaali Hassaine ◽  
Yutong Cai ◽  
...  

Over a billion people live in slums, with poor sanitation, education, property rights and working conditions having direct impact on current residents and future generations. A key problem in relation to slums is slum mapping. Without delineations of where all slum settlements are, informed decisions cannot be made by policymakers in order to benefit the most in need. Satellite images have been used in combination with machine learning models to try and fill the gap in data availability of slum locations. Deep learning has been used on RGB images with some success but since labeled satellite images of slums are relatively low quality and the physical/visual manifestation of slums significantly varies within and across countries, it is important to quantify the uncertainty of predictions for reliable application in downstream tasks. Our solution is to train Monte Carlo dropout U-Net models on multispectral 13-band Sentinel-2 images from which we can calculate pixelwise epistemic (model) and aleatoric (data) uncertainty in our predictions. We trained our model on labelled images of Mumbai and verified our epistemic and aleatoric uncertainty quantification approach using altered models trained on modified datasets. We also used SHAP values to investigate how the different features contribute towards the model’s predictions and this showed that certain short-wave infrared and red-edge image bands are powerful features for determining the locations of slums within images. Having created our model with uncertainty quantification, in the future it can be applied to downstream tasks and decision-makers will know where predictions have been made with low uncertainty, giving them greater confidence in its deployment.


Author(s):  
Sergey V. Pyankov ◽  
Nikolay G. Maximovich ◽  
Elena A. Khayrulina ◽  
Olga A. Berezina ◽  
Andrey N. Shikhov ◽  
...  

2021 ◽  
Vol 13 (8) ◽  
pp. 1505
Author(s):  
Klaudia Kryniecka ◽  
Artur Magnuszewski

The lower Vistula River was regulated in the years 1856–1878, at a distance of 718–939 km. The regulation plan did not take into consideration the large transport of the bed load. The channel was shaped using simplified geometry—too wide for the low flow and overly straight for the stabilization of the sandbar movement. The hydraulic parameters of the lower Vistula River show high velocities of flow and high shear stress. The movement of the alternate sandbars can be traced on the optical satellite images of Sentinel-2. In this study, a method of sandbar detection through the remote sensing indices, Sentinel Water Mask (SWM) and Automated Water Extraction Index no shadow (AWEInsh), and the manual delineation with visual interpretation (MD) was used on satellite images of the lower Vistula River, recorded at the time of low flows (20 August 2015, 4 September 2016, 30 July 2017, 20 September 2018, and 29 August 2019). The comparison of 32 alternate sandbar areas obtained by SWM, AWEInsh, and MD manual delineation methods on the Sentinel-2 images, recorded on 20 August 2015, was performed by the statistical analysis of the interclass correlation coefficient (ICC). The distance of the shift in the analyzed time intervals between the image registration dates depends on the value of the mean discharge (MQ). The period from 30 July 2017 to 20 September 2018 was wet (MQ = 1140 m3 × s−1) and created conditions for the largest average distance of the alternate sandbar shift, from 509 to 548 m. The velocity of movement, calculated as an average shift for one day, was between 1.2 and 1.3 m × day−1. The smallest shift of alternate sandbars was characteristic of the low flow period from 20 August 2015 to 4 September 2016 (MQ = 306 m3 × s−1), from 279 to 310 m, with an average velocity from 0.7 to 0.8 m × day−1.


Author(s):  
Viacheslav V. Krylenko ◽  
◽  
Marina V. Krylenko ◽  
Alexander A. Aleynikov ◽  
◽  
...  

The study of the relief of large coastal accumulative forms, based on modern technologies, is rele-vant for solving many applied problems. Coastal and underwater bars, shoals, banks are characteristic elements of large coastal accumulative forms’ geosystems. Previously existing methods of relief re-searches, especially underwater, were labor-intensive and expensive. Accordingly, the development and implementation of new methods of geographical research are necessary. The Dolgaya Spit, includ-ing its underwater shoal and the Elenin Bank, is one of the largest accumulative forms of the Sea of Azov. The purpose of our work was to obtain new information on the relief structure and the shoreline dynamics of the Dolgaya Spit based on the use of new research methods. Digital models of surface and underwater relief were built on the basis of processing Sentinel-2 satellite images and data from unmanned aerial photography. The subsequent analysis allowed identify regularities that reflect the current and previous hydro-lithodynamic conditions that determined the transformation of the Dolgaya Spit during its evolution. The fulfilled studies confirmed the possibility of successful use of modern remote methods for studying the relief of coastal accumulative forms.


Author(s):  
F. Pineda ◽  
V. Ayma ◽  
C. Beltran

Abstract. High-resolution satellite images have always been in high demand due to the greater detail and precision they offer, as well as the wide scope of the fields in which they could be applied; however, satellites in operation offering very high-resolution (VHR) images has experienced an important increase, but they remain as a smaller proportion against existing lower resolution (HR) satellites. Recent models of convolutional neural networks (CNN) are very suitable for applications with image processing, like resolution enhancement of images; but in order to obtain an acceptable result, it is important, not only to define the kind of CNN architecture but the reference set of images to train the model. Our work proposes an alternative to improve the spatial resolution of HR images obtained by Sentinel-2 satellite by using the VHR images from PeruSat1, a peruvian satellite, which serve as the reference for the super-resolution approach implementation based on a Generative Adversarial Network (GAN) model, as an alternative for obtaining VHR images. The VHR PeruSat-1 image dataset is used for the training process of the network. The results obtained were analyzed considering the Peak Signal to Noise Ratios (PSNR) and the Structural Similarity (SSIM). Finally, some visual outcomes, over a given testing dataset, are presented so the performance of the model could be analyzed as well.


Limnetica ◽  
2020 ◽  
Vol 39 (1) ◽  
pp. 373-386
Author(s):  
Xavier Sòria-Perpinyà ◽  
Marcela Pereira-Sandoval ◽  
Antonio Ruiz-Verdú ◽  
Juan M. Soria ◽  
Jesús Delegido ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document