River ice breakup classification using dual- (HH&HV) or compact-polarization RADARSAT Constellation Mission data.
Saved in:
| Title: | River ice breakup classification using dual- (HH&HV) or compact-polarization RADARSAT Constellation Mission data. |
|---|---|
| Authors: | Geldsetzer, Torsten1 (AUTHOR) torsten.geldsetzer@gmail.com, Svacina, Nicolas1 (AUTHOR), Tolszczuk-Leclerc, Simon1 (AUTHOR), van der Sanden, Joost1 (AUTHOR) |
| Source: | Remote Sensing of Environment. Oct2024, Vol. 312, pN.PAG-N.PAG. 1p. |
| Subjects: | Ice on rivers, lakes, etc., Ice sheets, Synthetic aperture radar, Recursive partitioning, Snow cover |
| Abstract: | Ice jams and associated flooding during river ice breakup are seasonal hazards for many northern communities. Synthetic Aperture Radar (SAR) data enable timely monitoring of river ice conditions in inclement weather. River ice types and open water are discriminated during breakup using dual-polarized (HH&HV; DPH) and compact polarimetric (CP) C-band SAR imagery from the RADARSAT Constellation Mission (RCM). Study areas comprise nineteen locations on rivers in northern Ontario, the Northwest Territories, and Alberta, Canada. Rubble ice, sheet ice, and open water areas are sampled in RCM imagery using corroborating evidence from shoreline cameras and oblique airborne photography. Samples are obtained in the incidence angle range 19° to 48°, for open water with various wind speeds and flow conditions, and for rubble ice and sheet ice with varying surface roughness, snow cover, and wetness. Discrimination relies on a primary classification of rubble ice, sheet ice, and open water, and a secondary classification of ice roughness. The primary classification model development uses a recursive-partitioning machine-learning technique. Overall accuracies for the best DPH models are 83.8% to 89.6%, depending on image noise floor. DPH models use both HH and HV polarizations for low noise floor image modes, whereas only HH is used for higher noise floor image modes. The best CP model accuracy is 90.2%, using the RL, RR, and RVRH parameters. Secondary classification associates increasing ice roughness with increasing HH backscatter for DPH data, and with increasing RH backscatter for CP data. The angular dependencies of rubble ice or sheet ice are used to normalize backscatter, which is then divided into roughness categories. The DPH and CP classification models are named IceBC-DP and IceBC-CP, respectively. Qualitative analysis illustrates good overall ice type and open water classification. Confounding situations are mixed pixel issues, whitewater and wind roughening of water, moist snow microwave absorption, and superficial water on sheet ice. The development of robust methods for monitoring ice jam formation with RCM is an operational concern for departments within the Government of Canada and within Provincial and Territorial Governments. • Dual-polarized and compact-polarized SAR are effective for river ice monitoring. • Rubble ice, sheet ice, and open water are discriminated with 90% accuracy. • Cross-polarized backscatter aids co-polarized if the SAR noise floor is low. • Same-sense, opposite-sense, and co-polarization ratio parameters work in concert. • Model development with recursive partitioning produces robust SAR decision trees. [ABSTRACT FROM AUTHOR] |
| Copyright of Remote Sensing of Environment is the property of Elsevier B.V. and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
| FullText | Text: Availability: 0 |
|---|---|
| Header | DbId: egs DbLabel: Engineering Source An: 178941807 AccessLevel: 6 PubType: Academic Journal PubTypeId: academicJournal PreciseRelevancyScore: 0 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: River ice breakup classification using dual- (HH&HV) or compact-polarization RADARSAT Constellation Mission data. – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Geldsetzer%2C+Torsten%22">Geldsetzer, Torsten</searchLink><relatesTo>1</relatesTo> (AUTHOR)<i> torsten.geldsetzer@gmail.com</i><br /><searchLink fieldCode="AR" term="%22Svacina%2C+Nicolas%22">Svacina, Nicolas</searchLink><relatesTo>1</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Tolszczuk-Leclerc%2C+Simon%22">Tolszczuk-Leclerc, Simon</searchLink><relatesTo>1</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22van+der+Sanden%2C+Joost%22">van der Sanden, Joost</searchLink><relatesTo>1</relatesTo> (AUTHOR) – Name: TitleSource Label: Source Group: Src Data: <searchLink fieldCode="JN" term="%22Remote+Sensing+of+Environment%22">Remote Sensing of Environment</searchLink>. Oct2024, Vol. 312, pN.PAG-N.PAG. 1p. – Name: Subject Label: Subjects Group: Su Data: <searchLink fieldCode="DE" term="%22Ice+on+rivers%2C+lakes%2C+etc%2E%22">Ice on rivers, lakes, etc.</searchLink><br /><searchLink fieldCode="DE" term="%22Ice+sheets%22">Ice sheets</searchLink><br /><searchLink fieldCode="DE" term="%22Synthetic+aperture+radar%22">Synthetic aperture radar</searchLink><br /><searchLink fieldCode="DE" term="%22Recursive+partitioning%22">Recursive partitioning</searchLink><br /><searchLink fieldCode="DE" term="%22Snow+cover%22">Snow cover</searchLink> – Name: Abstract Label: Abstract Group: Ab Data: Ice jams and associated flooding during river ice breakup are seasonal hazards for many northern communities. Synthetic Aperture Radar (SAR) data enable timely monitoring of river ice conditions in inclement weather. River ice types and open water are discriminated during breakup using dual-polarized (HH&HV; DPH) and compact polarimetric (CP) C-band SAR imagery from the RADARSAT Constellation Mission (RCM). Study areas comprise nineteen locations on rivers in northern Ontario, the Northwest Territories, and Alberta, Canada. Rubble ice, sheet ice, and open water areas are sampled in RCM imagery using corroborating evidence from shoreline cameras and oblique airborne photography. Samples are obtained in the incidence angle range 19° to 48°, for open water with various wind speeds and flow conditions, and for rubble ice and sheet ice with varying surface roughness, snow cover, and wetness. Discrimination relies on a primary classification of rubble ice, sheet ice, and open water, and a secondary classification of ice roughness. The primary classification model development uses a recursive-partitioning machine-learning technique. Overall accuracies for the best DPH models are 83.8% to 89.6%, depending on image noise floor. DPH models use both HH and HV polarizations for low noise floor image modes, whereas only HH is used for higher noise floor image modes. The best CP model accuracy is 90.2%, using the RL, RR, and RVRH parameters. Secondary classification associates increasing ice roughness with increasing HH backscatter for DPH data, and with increasing RH backscatter for CP data. The angular dependencies of rubble ice or sheet ice are used to normalize backscatter, which is then divided into roughness categories. The DPH and CP classification models are named IceBC-DP and IceBC-CP, respectively. Qualitative analysis illustrates good overall ice type and open water classification. Confounding situations are mixed pixel issues, whitewater and wind roughening of water, moist snow microwave absorption, and superficial water on sheet ice. The development of robust methods for monitoring ice jam formation with RCM is an operational concern for departments within the Government of Canada and within Provincial and Territorial Governments. • Dual-polarized and compact-polarized SAR are effective for river ice monitoring. • Rubble ice, sheet ice, and open water are discriminated with 90% accuracy. • Cross-polarized backscatter aids co-polarized if the SAR noise floor is low. • Same-sense, opposite-sense, and co-polarization ratio parameters work in concert. • Model development with recursive partitioning produces robust SAR decision trees. [ABSTRACT FROM AUTHOR] – Name: AbstractSuppliedCopyright Label: Group: Ab Data: <i>Copyright of Remote Sensing of Environment is the property of Elsevier B.V. and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.) |
| PLink | https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=178941807 |
| RecordInfo | BibRecord: BibEntity: Identifiers: – Type: doi Value: 10.1016/j.rse.2024.114313 Languages: – Code: eng Text: English PhysicalDescription: Pagination: PageCount: 1 StartPage: N.PAG Subjects: – SubjectFull: Ice on rivers, lakes, etc. Type: general – SubjectFull: Ice sheets Type: general – SubjectFull: Synthetic aperture radar Type: general – SubjectFull: Recursive partitioning Type: general – SubjectFull: Snow cover Type: general Titles: – TitleFull: River ice breakup classification using dual- (HH&HV) or compact-polarization RADARSAT Constellation Mission data. Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Geldsetzer, Torsten – PersonEntity: Name: NameFull: Svacina, Nicolas – PersonEntity: Name: NameFull: Tolszczuk-Leclerc, Simon – PersonEntity: Name: NameFull: van der Sanden, Joost IsPartOfRelationships: – BibEntity: Dates: – D: 01 M: 10 Text: Oct2024 Type: published Y: 2024 Identifiers: – Type: issn-print Value: 00344257 Numbering: – Type: volume Value: 312 Titles: – TitleFull: Remote Sensing of Environment Type: main |
| ResultId | 1 |
