Scholarly record
SEA ICE CLASSIFICATION BASED ON NEURAL NETWORKS METHOD USING SENTINEL-1 DATA
Abstract
A supervised classification algorithm has been developed for automated sea ice retrieval in some regions of the Arctic seas. Sentinel-1 Extra Wide images were utilized as input for a Neural Network (NN)-based algorithm. Synthetic aperture radar (SAR) data has several features significantly affected on the microwave backscatter from sea ice. In this paper, an approach to improve data quality is proposed. Since different sea ice types can have similar backscattering coefficients, the extracted SAR texture features in addition to the backscattering coefficients were used. The NN has been trained with backpropagation learning method. Based on the analysis of classification errors and processing time the optimal topology of the NN was found. We consider the distribution of ice of different ages and focus on a particular aspect of the old ice edge retrieval. Results demonstrate the overall potential of dual-polarized SAR data for standalone using for automated sea ice types delineation as well as old ice boundary detection.
Publication Impact Profile
- Citations
- Scopus - Citation Indexes: 5
- Captures
- Mendeley - Readers: 2
Publication details
References0
Structured references will appear here after the reference import pass. The count is preserved now so the scholarly record is not incomplete.
Citing literature
Number of times cited according to Crossref: 4
View or Download full articleAccess options
SWS access login
Login as SWS Scientific CommitteeLogin as SWS Scientific PartnerLogin as SWS AuthorAuthors and approved SWS contributors will read and export their own linked papers after identity matching by SWS profile, email and SGEM GlobalID.
For librarian assistance: [email protected]
Purchase Instant Access
- Article can be downloaded after successful payment.
- Article may be used according to SWS library access terms.
- Article cannot be redistributed.