SAR and InSAR for Flood Monitoring: Examples With COSMO-SkyMed Data

Refice, Alberto; Capolongo, Domenico; Pasquariello, G.; DaAddabbo, Annarita; Bovenga, F.; Nutricato, R.; Lovergine, Francesco P.; Pietranera, L.

doi:10.1109/jstars.2014.2305165

Cited by 111 publications

(60 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…If data are available during the maximum flooding phase, it is possible to accurately map the affected area using high-resolution SAR images, such as those acquired by the TerraSAR-X (Giustarini et al, 2013) and COSMO-SkyMed (CSKM) (Refice et al, 2014) satellites. In particular, the identification of the flooded area is performed by analysing the SAR backscattering, which shows low values in water-covered areas.…”

Section: Sar Datamentioning

confidence: 99%

“…Copernicus Emergency Management Service (©European Union, 2012Union, -2017). With satellite remote sensing data, it is possible to map flood effects over vast areas at different spatial and temporal resolutions using multispectral (Brakenridge et al, 2006;Gianinetto et al, 2006;Nigro et al, 2014;Wang et al, 2012;Yan et al, 2015;Rahman and Di, 2017) or Synthetic Aperture Radar (SAR) images (Boni et al, 2016;Mason et al, 2014;Schumann et al, 2015;Refice et al, 2014;Pulvirenti et al, 2011;Clement et al, 2017;Brivio et al, 2002). A good description of the main methodologies that are used to map floods with satellite data has been published by Fayne…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Low cost, multiscale and multi-sensor application for flooded area mapping

Giordan

Notti

Villa³

et al. 2018

Nat. Hazards Earth Syst. Sci.

View full text Add to dashboard Cite

Abstract. Flood mapping and estimation of the maximum water depth are essential elements for the first damage evaluation, civil protection intervention planning and detection of areas where remediation is needed.In this work, we present and discuss a methodology for mapping and quantifying flood severity over floodplains. The proposed methodology considers a multiscale and multisensor approach using free or low-cost data and sensors. We applied this method to the November 2016 Piedmont (northwestern Italy) flood. We first mapped the flooded areas at the basin scale using free satellite data from low-to mediumhigh-resolution from both the SAR (Sentinel-1, COSMOSkymed) and multispectral sensors (MODIS, Sentinel-2). Using very-and ultra-high-resolution images from the lowcost aerial platform and remotely piloted aerial system, we refined the flooded zone and detected the most damaged sector. The presented method considers both urbanised and nonurbanised areas. Nadiral images have several limitations, in particular in urbanised areas, where the use of terrestrial images solved this limitation. Very-and ultra-high-resolution images were processed with structure from motion (SfM) for the realisation of 3-D models. These data, combined with an available digital terrain model, allowed us to obtain maps of the flooded area, maximum high water area and damaged infrastructures.

show abstract

Section: Sar Datamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Low cost, multiscale and multi-sensor application for flooded area mapping

Giordan

Notti

Villa³

et al. 2018

Nat. Hazards Earth Syst. Sci.

View full text Add to dashboard Cite

show abstract

“…If data are available during the maximum flooding phase, it is possible to accurately map the affected area using high resolution SAR images as those acquired by the TerraSAR-X (Giustarini et al, 2013) and COSMO-SkyMed 165 (Refice et al, 2014) satellites. In particular, the identification of the flooded area is performed by analysing the SAR backscattering, which generally shows low values in water-covered areas.…”

Section: Sar Datamentioning

confidence: 99%

“…Advances in remote sensing and geotechnology have introduced the possibility, in last years, of having rapid maps and models during or little time after a flood event (Copernicus Emergency Management Service (© European Union, 2012). With satellite remote sensing data, it is possible to map flood effects over wide areas at different spatial and temporal resolution using multispectral (Brakenridge, 35 et al, 2006;Gianinetto et al,2006;Nigro et al, 2014;Wang et al;Yan et al, 2015;Rahman and Di, 2017) or Synthetic Aperture Radar (SAR) images (Boni et al 2016;Mason et al 2014;Guy et al 2015;Refice et al 2014;Pulvirenti et al;Clement et al, 2017;Brivio et al;2002). A good description of main methodologies used to map flood with satellite data has been published by Fayne et al (2017).…”

mentioning

confidence: 99%

Low cost, multiscale and multi-sensor application for flooded areas mapping

Giordan¹,

Notti²,

Villa³

et al. 2017

Preprint

View full text Add to dashboard Cite

Flooded areas mapping and estimation of maximum water height are important elements for a first damages evaluation, civil 15 protection interventions planning and detection of areas where remedial are more needed.In this work, a methodology for mapping and quantifying flood severity over plain areas is presented and discussed. The proposed methodology considers a multiscale and multi-sensor approach using free or low cost data/sensors. We applied this method to November 2016 Piemonte (NW Italy) flood. We first mapped flooded areas at basin scale using free satellite data from low to medium-high resolution using both SAR (Sentinel-1, Cosmo-Skymed) and multispectral sensors (MODIS, 20 Sentinel-2). Using very-and ultra-high resolution images from the low-cost aerial platform and Remotely Piloted Aerial System, we refined the flooded area and we detected the most damaged. The presented method considers both urbanized and not urbanized areas. Nadiral images have several limitations in particular in urbanized areas, where the use of terrestrial images solved this limitation. Very-and ultra-high resolution images have been processed with Structure from Motion (SfM) for the realization of 3-D models. These data, combined with available digital elevation model, allowed us to obtain maps of flooded 25 area, maximum water high and damaged infrastructures.

show abstract

“…For example, in the studies on the automated segmentation from magnetic resonance images [19][20][21], the number of training examples is very huge (up to millions), the classes are strongly imbalanced, and generating accurate statistical solution is not trivial. In addition, data imbalance in huge data sets is also reported in other applicative domains, such as marketing data [22], oil spill detection or land cover changes from remote sensing images [16,27], text classification [18] and scene classification [35]. In these areas, very large data sets have to be handled and the minority class is the one of interest, consequently two problematic issues add on: the computational complexity dependent on the size of the data set and the need to pursue a fairly high rate of correct detections in the minority class.…”

Section: Introductionmentioning

confidence: 99%

Parallel selective sampling method for imbalanced and large data classification

D’Addabbo

Maglietta

2015

Pattern Recognition Letters

View full text Add to dashboard Cite

a b s t r a c tSeveral applications aim to identify rare events from very large data sets. Classification algorithms may present great limitations on large data sets and show a performance degradation due to class imbalance. Many solutions have been presented in literature to deal with the problem of huge amount of data or imbalancing separately. In this paper we assessed the performances of a novel method, Parallel Selective Sampling (PSS), able to select data from the majority class to reduce imbalance in large data sets. PSS was combined with the Support Vector Machine (SVM) classification. PSS-SVM showed excellent performances on synthetic data sets, much better than SVM. Moreover, we showed that on real data sets PSS-SVM classifiers had performances slightly better than those of SVM and RUSBoost classifiers with reduced processing times. In fact, the proposed strategy was conceived and designed for parallel and distributed computing. In conclusion, PSS-SVM is a valuable alternative to SVM and RUSBoost for the problem of classification by huge and imbalanced data, due to its accurate statistical predictions and low computational complexity.

show abstract

SAR and InSAR for Flood Monitoring: Examples With COSMO-SkyMed Data

Cited by 111 publications

References 39 publications

Low cost, multiscale and multi-sensor application for flooded area mapping

Low cost, multiscale and multi-sensor application for flooded area mapping

Low cost, multiscale and multi-sensor application for flooded areas mapping

Parallel selective sampling method for imbalanced and large data classification

Contact Info

Product

Resources

About