Lake Ice Detection from Sentinel-1 SAR with Deep Learning

Tom, Manu; Aguilar, Roberto; Imhof, Pascal; Leinß, Silvan; Baltsavias, Emmanuel P.; Schindler, Konrad

doi:10.48550/arxiv.2002.07040

Cited by 2 publications

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This data provides new possibilities for the semipermanent monitoring of ice coverage of large rivers and lakes (Van Leeuwen et al, 2018). Earlier research demonstrated that remote sensing data can be used to detect, classify and monitor ice on sea or on large lakes (Tom et al, 2020;Howell et al, 2021;Zhang et al, 2021;Lohse et al, 2020;Li et al, 2021). River ice monitoring in Arctic regions has been conducted by several authors: Weber et al (2003) used active data, Altena et al (2021) applied different optical data sets, Goldberg et al (2020) worked with low resolution RS data to predict ice jams, and Zakharova et al (2021) used altimetry to classify river ice.…”

Section: Introductionmentioning

confidence: 99%

River ice monitoring of the Danube and Tisza rivers using Sentinel-1 radar data

van¹,

Sípos²,

Lábdy³

et al. 2022

Geographica Pannonica

View full text Add to dashboard Cite

Due to extreme weather, occasionally Hungary's main rivers and lakes grow an ice cover causing severe damage to infrastructure and increased flood hazard. During cold periods in 2017 and 2022, a dangerous layer of ice developed on the main rivers in the country. Since river ice is rare in this region, no permanent ice monitoring system is in operation. Due to their all weather capabilities, active remote sensing instruments provide a good opportunity to monitor ice coverage. ESA's Sentinel-1 radar satellites acquire data with a relatively high spatial and temporal resolution. A method was developed to provide ice coverage information at a regular interval; depending on the satellite revisit, at least once every 5 days, but often also on a daily basis. In 2017, maps were created for sections along the Danube and in 2022 for another section of the Tisza river. The ice coverage was calculated with a spatial resolution of 10 metre and visualised with a spatial density of 100 metre along the rivers. The mapping procedure provides visual information to give a fast overview of the spatial extent of ice coverage and quantitative , tabular information for operational activities to mitigate the damage due to ice packs and ice jams.

show abstract

Section: Introductionmentioning

confidence: 99%

River ice monitoring of the Danube and Tisza rivers using Sentinel-1 radar data

van¹,

Sípos²,

Lábdy³

et al. 2022

Geographica Pannonica

View full text Add to dashboard Cite

show abstract

“…around 5cm (see Rott et Nagler 1994, Nagler et al 2016, thereby yielding a much lower radar backscatter, which is comparable to open water. Very recently, Tom et al 2020 have used a deep learning approach (DeepLabV3+ Convolutional Neural Network (CNN)) to classify water and ice over three Swiss lakes, using webcam and Sentinel-2 observations to provide a ground truth for the training phase, and for the evaluation of the network's performance. Taking advantage of the spatial correlation of information, a CNN should be able to interpret texture and contextual information with much more complexity than machine learning methods, and improve classification in windy conditions.…”

Section: Introductionmentioning

confidence: 99%

“…In fact, a human operator can, in the vast majority of cases, discriminate between water, land and ice, looking only at the RGB bands. This is why satellite imagery is generally used to provide ground truth for training classifiers using SAR data as done by a human operator in Tom et al 2020. However, because of the aforementioned unreliability of satellite optical data for accurately tracking the lake freeze-up and break-up dates, it has seen much less consideration in literature than SAR.…”

Section: Introductionmentioning

confidence: 99%

On water and ice classification from Sentinel-2 imagery using machine learning

Jugier

Cremese

Fournier

et al. 2022

Preprint

View full text Add to dashboard Cite

Accurate and dynamic mapping of water and ice surfaces is directly useful to navigation and lake ice cover monitoring to study climate change. Water and ice maps are also useful for various scientific applications such as atmospheric correction of satellite imagery, remote sensing of water quality, and as input data for hydrological, weather and climate models. The existing literature shows that multi-spectral satellite imagery, as provided by Sentinel-2 and Landsat-8, provides a very effective means to discriminate between water, land, and ice. However, most studies focus either on very specific cases (a specific lake for instance), or on general cases but without complex and yet very frequent cases such as turbid waters and salt lakes which can be confused with snow and ice. The Copernicus High-Resolution Snow and Ice Monitoring Service provides an operational Sentinel-2 ice and water classification product at 20m resolution but with a lot of confusion on the aforementioned cases. Using a database of 31 fully hand-labelled Sentinel-2 L2A atmospherically corrected images, and machine learning SVM and RandomForest methods, the current study shows that the classification of land, water, ice, snow, turbid waters, salt lake categories can be achieved with

show abstract

Lake Ice Detection from Sentinel-1 SAR with Deep Learning

Cited by 2 publications

References 0 publications

River ice monitoring of the Danube and Tisza rivers using Sentinel-1 radar data

River ice monitoring of the Danube and Tisza rivers using Sentinel-1 radar data

On water and ice classification from Sentinel-2 imagery using machine learning

Contact Info

Product

Resources

About