Baltic Sea Ice Concentration Estimation From C-Band Dual-Polarized SAR Imagery by Image Segmentation and Convolutional Neural Networks

Abstract

In this study application of convolutional neural networks (CNNs) preceded by synthetic aperture radar (SAR), image segmentation for sea ice concentration (SIC) estimation over the Baltic Sea from dual-polarized C-band SAR imagery is studied. Three algorithm variants were studied and trained using FMI ice chart SIC or a synthetic SIC dataset with different SIC values generated by combining pure open water and sea ice blocks by applying binary masks. The first two algorithm variants were trained using only open water and fully ice-covered patches, based on the FMI ice charts, and they had a similar CNN structure. These two algorithm variants differ only in deriving the segmentwise SIC from the CNN output. In the third algorithm, variant synthetic SIC data derived as a mixture of open water and fully ice-covered ice patches according to the ice charts were used in training. The estimation results were evaluated with respect to the FMI ice chart SIC for an independent test dataset. The results were very encouraging for operational purposes and significantly better than for our earlier SAR-based SIC estimation algorithms. The algorithm version trained with synthetic SIC data clearly outperformed the two other algorithm versions and our earlier SIC estimation results based on dual-polarized SAR, using independent FMI ice chart SIC as a reference.

Sea Ice Extraction via Remote Sensing Imagery: Algorithms, Datasets, Applications and Challenges

“…Experimental comparisons with existing machine-learning algorithms based on texture features and RF demonstrated improved accuracy and efficiency. CNNbased SIC estimation was shown to outperform earlier estimation algorithms in [92]. Additionally, Malmgren-Hansen et al [93] tested CNN under the scenario of disparate resolutions between Sentinel-1 SAR and AMSR 2 sensors and found that CNN was suitable for multi-sensor fusion with high robustness.…”

Pan-Arctic Sea Ice Concentration from SAR and Passive Microwave

“…The microwave signatures in C-band SAR imagery show patterns related to sea ice formations, but the discrimination between different sea ice conditions is challenged by ambiguities in backscatter intensities, noise phenomena, and wind-induced roughness on the ocean surface, etc. Such ambiguities can degrade the predictive performance of SAR-based sea ice retrieval algorithms (Stokholm et al, 2022;Khaleghian et al, 2021;Boulze et al, 2020;Karvonen, 2022). Approaches based on multisensor data fusion schemes that combine SAR imagery and PMW observations have been shown to yield better predictive performances on sea ice concentration than purely SAR-based approaches (Malmgren-Hansen et al, 2021;Karvonen, 2017).…”

IceGCN: An Interactive Sea Ice Classification Pipeline for SAR Imagery Based on Graph Convolutional Network

“…In recent years, deep learning has become popular in remote sensing from SAR imagery. Among different data-driven deep learning models, convolutional neural networks (CNNs) are widely adopted for sea ice classification [25][26][27][28][29][30][31] and sea ice concentration estimation [9,[32][33][34][35]. The ability to learn robust features automatically from a large volume of training data makes the CNN-based model a more preferable choice for sea ice classification compared with traditional machine learning.…”