scholarly journals DeepBedMap: a deep neural network for resolving the bed topography of Antarctica

2020 ◽  
Vol 14 (11) ◽  
pp. 3687-3705
Author(s):  
Wei Ji Leong ◽  
Huw Joseph Horgan
Keyword(s):  
Neural Network ◽  
High Resolution ◽  
Spatial Resolution ◽  
Ice Sheet ◽  
Super Resolution ◽  
Snow Accumulation ◽  
Generative Adversarial Network ◽  
Bed Topography ◽  
Bed Elevation ◽  
Elevation Model

Abstract. To resolve the bed elevation of Antarctica, we present DeepBedMap – a novel machine learning method that can produce Antarctic bed topography with adequate surface roughness from multiple remote sensing data inputs. The super-resolution deep convolutional neural network model is trained on scattered regions in Antarctica where high-resolution (250 m) ground-truth bed elevation grids are available. This model is then used to generate high-resolution bed topography in less surveyed areas. DeepBedMap improves on previous interpolation methods by not restricting itself to a low-spatial-resolution (1000 m) BEDMAP2 raster image as its prior image. It takes in additional high-spatial-resolution datasets, such as ice surface elevation, velocity and snow accumulation, to better inform the bed topography even in the absence of ice thickness data from direct ice-penetrating-radar surveys. The DeepBedMap model is based on an adapted architecture of the Enhanced Super-Resolution Generative Adversarial Network, chosen to minimize per-pixel elevation errors while producing realistic topography. The final product is a four-times-upsampled (250 m) bed elevation model of Antarctica that can be used by glaciologists interested in the subglacial terrain and by ice sheet modellers wanting to run catchment- or continent-scale ice sheet model simulations. We show that DeepBedMap offers a rougher topographic profile compared to the standard bicubically interpolated BEDMAP2 and BedMachine Antarctica and envision it being used where a high-resolution bed elevation model is required.

scholarly journals DeepBedMap: Using a deep neural network to better resolve the bed topography of Antarctica

2020 ◽  
Author(s):  
Wei Ji Leong ◽  
Huw Joseph Horgan
Keyword(s):  
Neural Network ◽  
High Resolution ◽  
Spatial Resolution ◽  
Network Architecture ◽  
Ice Sheet ◽  
Super Resolution ◽  
Snow Accumulation ◽  
Bed Topography ◽  
Bed Elevation ◽  
Elevation Model

Abstract. To better resolve the bed elevation of Antarctica, we present DeepBedMap – a novel machine learning method that produces realistic Antarctic bed topography from multiple remote sensing data inputs. Our super-resolution deep convolutional neural network model is trained on scattered regions in Antarctica where high resolution (250 m) groundtruth bed elevation grids are available. The model is then used to generate high resolution bed topography in less well surveyed areas. DeepBedMap improves on previous interpolation methods by not restricting itself to a low spatial resolution (1000 m) BEDMAP2 raster image as its prior. It takes in additional high spatial resolution datasets, such as ice surface elevation, velocity and snow accumulation to better inform the bed topography even in the absence of ice-thickness data from direct ice-penetrating radar surveys. Our DeepBedMap model is based on an adapted Enhanced Super Resolution Generative Adversarial Network architecture, chosen to minimize per-pixel elevation errors while producing realistic topography. The final product is a four times upsampled (250 m) bed elevation model of Antarctica that can be used by glaciologists interested in the subglacial terrain, and by ice sheet modellers wanting to run catchment or continent-scale ice sheet model simulations. We show that DeepBedMap offers a more realistic topographic roughness profile compared to a standard bicubic interpolated BEDMAP2 and BedMachine Antarctica, and envision it to be used where a high resolution bed elevation model is required.

scholarly journals Deriving VIIRS High-Spatial Resolution Water Property Data over Coastal and Inland Waters Using Deep Convolutional Neural Network

2021 ◽  
Vol 13 (10) ◽  
pp. 1944
Author(s):  
Xiaoming Liu ◽  
Menghua Wang
Keyword(s):  
Neural Network ◽  
High Resolution ◽  
Spatial Resolution ◽  
High Spatial Resolution ◽  
Ocean Color ◽  
Super Resolution ◽  
Inland Waters ◽  
Coastal Oceans ◽  
Chl A ◽  
Color Data

The Visible Infrared Imaging Radiometer Suite (VIIRS) onboard the Suomi National Polar-orbiting Partnership (SNPP) satellite has been a reliable source of ocean color data products, including five moderate (M) bands and one imagery (I) band normalized water-leaving radiance spectra nLw(λ). The spatial resolutions of the M-band and I-band nLw(λ) are 750 m and 375 m, respectively. With the technique of convolutional neural network (CNN), the M-band nLw(λ) imagery can be super-resolved from 750 m to 375 m spatial resolution by leveraging the high spatial resolution features of I1-band nLw(λ) data. However, it is also important to enhance the spatial resolution of VIIRS-derived chlorophyll-a (Chl-a) concentration and the water diffuse attenuation coefficient at the wavelength of 490 nm (Kd(490)), as well as other biological and biogeochemical products. In this study, we describe our effort to derive high-resolution Kd(490) and Chl-a data based on super-resolved nLw(λ) images at the VIIRS five M-bands. To improve the network performance over extremely turbid coastal oceans and inland waters, the networks are retrained with a training dataset including ocean color data from the Bohai Sea, Baltic Sea, and La Plata River Estuary, covering water types from clear open oceans to moderately turbid and highly turbid waters. The evaluation results show that the super-resolved Kd(490) image is much sharper than the original one, and has more detailed fine spatial structures. A similar enhancement of finer structures is also found in the super-resolved Chl-a images. Chl-a filaments are much sharper and thinner in the super-resolved image, and some of the very fine spatial features that are not shown in the original images appear in the super-resolved Chl-a imageries. The networks are also applied to four other coastal and inland water regions. The results show that super-resolution occurs mainly on pixels of Chl-a and Kd(490) features, especially on the feature edges and locations with a large spatial gradient. The biases between the original M-band images and super-resolved high-resolution images are small for both Chl-a and Kd(490) in moderately to extremely turbid coastal oceans and inland waters, indicating that the super-resolution process does not change the mean values of the original images.

scholarly journals Fast and accurate learned multiresolution dynamical downscaling for precipitation

2021 ◽  
Author(s):  
Jiali Wang ◽  
Zhengchun Liu ◽  
Ian Foster ◽  
Won Chang ◽  
Rajkumar Kettimuthu ◽  
...  
Keyword(s):  
Neural Network ◽  
High Resolution ◽  
Dynamical Downscaling ◽  
Computational Cost ◽  
Super Resolution ◽  
Ground Truth ◽  
Generative Adversarial Network ◽  
Adversarial Network ◽  
Spatial And Temporal Distributions ◽  
Data Variability

Abstract. This study develops a neural network-based approach for emulating high-resolution modeled precipitation data with comparable statistical properties but at greatly reduced computational cost. The key idea is to use combination of low- and high- resolution simulations to train a neural network to map from the former to the latter. Specifically, we define two types of CNNs, one that stacks variables directly and one that encodes each variable before stacking, and we train each CNN type both with a conventional loss function, such as mean square error (MSE), and with a conditional generative adversarial network (CGAN), for a total of four CNN variants.We compare the four new CNN-derived high-resolution precipitation results with precipitation generated from original high resolution simulations, a bilinear interpolater and the state-of-the-art CNN-based super-resolution (SR) technique. Results show that the SR technique produces results similar to those of the bilinear interpolator with smoother spatial and temporal distributions and smaller data variabilities and extremes than the high resolution simulations. While the new CNNs trained by MSE generate better results over some regions than the interpolator and SR technique do, their predictions are still not as close as ground truth. The CNNs trained by CGAN generate more realistic and physically reasonable results, better capturing not only data variability in time and space but also extremes such as intense and long-lasting storms. The new proposed CNN-based downscaling approach can downscale precipitation from 50 km to 12 km in 14 min for 30 years once the network is trained (training takes 4 hours using 1 GPU), while the conventional dynamical downscaling would take 1 months using 600 CPU cores to generate simulations at the resolution of 12 km over contiguous United States.

scholarly journals A new bedrock and surface elevation dataset for modelling the Greenland ice sheet

2003 ◽  
Vol 37 ◽  
pp. 351-356 ◽  
Author(s):  
Jonathan L. Bamber ◽  
Duncan J. Baldwin ◽  
S. Prasad Gogineni
Keyword(s):  
Ice Sheet ◽  
Greenland Ice Sheet ◽  
Detailed Comparison ◽  
Radar Data ◽  
The Arctic ◽  
Small Scale ◽  
Surface Model ◽  
Rock Outcrops ◽  
Bed Topography ◽  
Elevation Model

AbstractA new digital elevation model of the surface of the Greenland ice sheet and surrounding rock outcrops has been produced from a comprehensive suite of satellite and airborne remote-sensing and cartographic datasets. The surface model has been regridded to a resolution of 5 km, and combined with a new ice-thickness grid derived from ice-penetrating radar data collected in the 1970s and 1990s. A further dataset, the International Bathymetric Chart of the Arctic Ocean, was used to extend the bed elevations to include the continental shelf. The new bed topography was compared with a previous version used for ice-sheet modelling. Near the margins of the ice sheet and, in particular, in the vicinity of small-scale features associated with outlet glaciers and rapid ice motion, significant differences were noted. This was highlighted by a detailed comparison of the bed topography around the northeast Greenland ice stream.

scholarly journals Super Resolution Digital Image Correlation (SR-DIC): an Alternative to Image Stitching at High Magnifications

2021 ◽  
Author(s):  
R. S. Hansen ◽  
D. W. Waldram ◽  
T. Q. Thai ◽  
R. B. Berke
Keyword(s):  
High Resolution ◽  
Digital Image Correlation ◽  
Spatial Resolution ◽  
Digital Image ◽  
Super Resolution ◽  
Image Correlation ◽  
Low Resolution ◽  
Strain Measurements ◽  
Resolution Image ◽  
High Resolution Image

Abstract Background High-resolution Digital Image Correlation (DIC) measurements have previously been produced by stitching of neighboring images, which often requires short working distances. Separately, the image processing community has developed super resolution (SR) imaging techniques, which improve resolution by combining multiple overlapping images. Objective This work investigates the novel pairing of super resolution with digital image correlation, as an alternative method to produce high-resolution full-field strain measurements. Methods First, an image reconstruction test is performed, comparing the ability of three previously published SR algorithms to replicate a high-resolution image. Second, an applied translation is compared against DIC measurement using both low- and super-resolution images. Third, a ring sample is mechanically deformed and DIC strain measurements from low- and super-resolution images are compared. Results SR measurements show improvements compared to low-resolution images, although they do not perfectly replicate the high-resolution image. SR-DIC demonstrates reduced error and improved confidence in measuring rigid body translation when compared to low resolution alternatives, and it also shows improvement in spatial resolution for strain measurements of ring deformation. Conclusions Super resolution imaging can be effectively paired with Digital Image Correlation, offering improved spatial resolution, reduced error, and increased measurement confidence.

scholarly journals A GENERATIVE ADVERSARIAL NETWORK APPROACH FOR SUPER-RESOLUTION OF SENTINEL-2 SATELLITE IMAGES

2020 ◽  
Vol XLIII-B1-2020 ◽  
pp. 9-14
Author(s):  
F. Pineda ◽  
V. Ayma ◽  
C. Beltran
Keyword(s):  
High Resolution ◽  
Satellite Images ◽  
Resolution Enhancement ◽  
Super Resolution ◽  
Structural Similarity ◽  
Generative Adversarial Network ◽  
Acceptable Result ◽  
Adversarial Network ◽  
High Resolution Satellite Images ◽  
Sentinel 2

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.

scholarly journals Large-scale atmospheric forcing and topographic modification of precipitation rates over High Asia – a neural-network-based approach

2015 ◽  
Vol 6 (1) ◽  
pp. 61-81 ◽  
Author(s):  
L. Gerlitz ◽  
O. Conrad ◽  
J. Böhner
Keyword(s):  
Neural Network ◽  
High Resolution ◽  
Spatial Resolution ◽  
Large Scale ◽  
Local Scale ◽  
Predictor Variables ◽  
High Mountain ◽  
Target Area ◽  
Monthly Precipitation ◽  
Highly Nonlinear

Abstract. The heterogeneity of precipitation rates in high-mountain regions is not sufficiently captured by state-of-the-art climate reanalysis products due to their limited spatial resolution. Thus there exists a large gap between the available data sets and the demands of climate impact studies. The presented approach aims to generate spatially high resolution precipitation fields for a target area in central Asia, covering the Tibetan Plateau and the adjacent mountain ranges and lowlands. Based on the assumption that observed local-scale precipitation amounts are triggered by varying large-scale atmospheric situations and modified by local-scale topographic characteristics, the statistical downscaling approach estimates local-scale precipitation rates as a function of large-scale atmospheric conditions, derived from the ERA-Interim reanalysis and high-resolution terrain parameters. Since the relationships of the predictor variables with local-scale observations are rather unknown and highly nonlinear, an artificial neural network (ANN) was utilized for the development of adequate transfer functions. Different ANN architectures were evaluated with regard to their predictive performance. The final downscaling model was used for the cellwise estimation of monthly precipitation sums, the number of rainy days and the maximum daily precipitation amount with a spatial resolution of 1 km2. The model was found to sufficiently capture the temporal and spatial variations in precipitation rates in the highly structured target area and allows for a detailed analysis of the precipitation distribution. A concluding sensitivity analysis of the ANN model reveals the effect of the atmospheric and topographic predictor variables on the precipitation estimations in the climatically diverse subregions.

scholarly journals A high-resolution synthetic bed elevation grid of the Antarctic continent

2016 ◽  
Author(s):  
Felicity S. Graham ◽  
Jason L. Roberts ◽  
Ben K. Galton-Fenzi ◽  
Duncan Young ◽  
Donald Blankenship ◽  
...  
Keyword(s):  
High Resolution ◽  
Ice Sheet ◽  
Radar Data ◽  
Detailed Knowledge ◽  
Pass Filter ◽  
Low Pass Filter ◽  
Bed Elevation ◽  
Antarctic Continent ◽  
Ice Sheet Dynamics ◽  
The Antarctic

Abstract. Digital elevation models of Antarctic bed topography are heavily smoothed and interpolated onto low-resolution (> 1 km) grids as our current observed topography data are generally sparsely and unevenly sampled. This issue has potential implications for numerical simulations of ice-sheet dynamics, especially in regions prone to instability where detailed knowledge of the topography, including fine-scale roughness, is required. Here, we present a high-resolution (100 m) synthetic bed elevation terrain for the whole Antarctic continent. The synthetic bed surface preserves topographic roughness characteristics of airborne and ground-based ice-penetrating radar data from the Bedmap1 compilation and the ICECAP consortium. Broad-scale features of the Antarctic landscape are incorporated using a low-pass filter of the Bedmap2 bed-elevation data. Although not intended as a substitute for Bedmap2, the simulated bed elevation terrain has applicability in high-resolution ice-sheet modelling studies, including investigations of the interaction between topography, ice-sheet dynamics, and hydrology, where processes are highly sensitive to bed elevations. The data are available for download at the Australian Antarctic Data Centre (doi:10.4225/15/57464ADE22F50).

Super-Resolution Based on Generative Adversarial Network for HRTEM Images

2021 ◽  
pp. 2154027
Author(s):  
Fuqi Mao ◽  
Xiaohan Guan ◽  
Ruoyu Wang ◽  
Wen Yue
Keyword(s):  
Deep Learning ◽  
High Resolution ◽  
Information Structure ◽  
Rapid Development ◽  
Super Resolution ◽  
Training Data ◽  
Learning Technology ◽  
Hrtem Image ◽  
Generative Adversarial Network ◽  
Dual Regression

As an important tool to study the microstructure and properties of materials, High Resolution Transmission Electron Microscope (HRTEM) images can obtain the lattice fringe image (reflecting the crystal plane spacing information), structure image and individual atom image (which reflects the configuration of atoms or atomic groups in crystal structure). Despite the rapid development of HTTEM devices, HRTEM images still have limited achievable resolution for human visual system. With the rapid development of deep learning technology in recent years, researchers are actively exploring the Super-resolution (SR) model based on deep learning, and the model has reached the current best level in various SR benchmarks. Using SR to reconstruct high-resolution HRTEM image is helpful to the material science research. However, there is one core issue that has not been resolved: most of these super-resolution methods require the training data to exist in pairs. In actual scenarios, especially for HRTEM images, there are no corresponding HR images. To reconstruct high quality HRTEM image, a novel Super-Resolution architecture for HRTEM images is proposed in this paper. Borrowing the idea from Dual Regression Networks (DRN), we introduce an additional dual regression structure to ESRGAN, by training the model with unpaired HRTEM images and paired nature images. Results of extensive benchmark experiments demonstrate that the proposed method achieves better performance than the most resent SISR methods with both quantitative and visual results.

scholarly journals Automatic Target Recognition for Synthetic Aperture Radar Images Based on Super-Resolution Generative Adversarial Network and Deep Convolutional Neural Network

2019 ◽  
Vol 11 (2) ◽  
pp. 135 ◽  
Author(s):  
Xiaoran Shi ◽  
Feng Zhou ◽  
Shuang Yang ◽  
Zijing Zhang ◽  
Tao Su
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Target Recognition ◽  
Super Resolution ◽  
Automatic Target Recognition ◽  
Synthetic Aperture ◽  
Sar Image ◽  
Generative Adversarial Network ◽  
Adversarial Network ◽  
Feature Characterization

Aiming at the problem of the difficulty of high-resolution synthetic aperture radar (SAR) image acquisition and poor feature characterization ability of low-resolution SAR image, this paper proposes a method of an automatic target recognition method for SAR images based on a super-resolution generative adversarial network (SRGAN) and deep convolutional neural network (DCNN). First, the threshold segmentation is utilized to eliminate the SAR image background clutter and speckle noise and accurately extract target area of interest. Second, the low-resolution SAR image is enhanced through SRGAN to improve the visual resolution and the feature characterization ability of target in the SAR image. Third, the automatic classification and recognition for SAR image is realized by using DCNN with good generalization performance. Finally, the open data set, moving and stationary target acquisition and recognition, is utilized and good recognition results are obtained under standard operating condition and extended operating conditions, which verify the effectiveness, robustness, and good generalization performance of the proposed method.

Sign in / Sign up

Export Citation Format

Share Document