RUF: Effective Sea Ice Floe Segmentation Using End-to-End RES-UNET-CRF with Dual Loss
Sea ice observations through satellite imaging have led to advancements in environmental research, ship navigation, and ice hazard forecasting in cold regions. Machine learning and, recently, deep learning techniques are being explored by various researchers to process vast amounts of Synthetic Aper...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-06-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/13/2460 |
| _version_ | 1797528968785559552 |
|---|---|
| author | Anmol Sharan Nagi Devinder Kumar Daniel Sola K. Andrea Scott |
| author_facet | Anmol Sharan Nagi Devinder Kumar Daniel Sola K. Andrea Scott |
| author_sort | Anmol Sharan Nagi |
| collection | DOAJ |
| description | Sea ice observations through satellite imaging have led to advancements in environmental research, ship navigation, and ice hazard forecasting in cold regions. Machine learning and, recently, deep learning techniques are being explored by various researchers to process vast amounts of Synthetic Aperture Radar (SAR) data for detecting potential hazards in navigational routes. Detection of hazards such as sea ice floes in Marginal Ice Zones (MIZs) is quite challenging as the floes are often embedded in a multiscale ice cover composed of ice filaments and eddies in addition to floes. This study proposes a segmentation model tailored for detecting ice floes in SAR images. The model exploits the advantages of both convolutional neural networks and convolutional conditional random field (Conv-CRF) in a combined manner. The residual UNET (RES-UNET) computes expressive features to generate coarse segmentation maps while the Conv-CRF exploits the spatial co-occurrence pairwise potentials along with the RES-UNET unary/segmentation maps to generate final predictions. The whole pipeline is trained end-to-end using a dual loss function. This dual loss function is composed of a weighted average of binary cross entropy and soft dice loss. The comparison of experimental results with the conventional segmentation networks such as UNET, DeepLabV3, and FCN-8 demonstrates the effectiveness of the proposed architecture. |
| first_indexed | 2024-03-10T10:06:46Z |
| format | Article |
| id | doaj.art-aa3b3765bccc4cd7a15f7c04f4c6aede |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T10:06:46Z |
| publishDate | 2021-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-aa3b3765bccc4cd7a15f7c04f4c6aede2023-11-22T01:30:32ZengMDPI AGRemote Sensing2072-42922021-06-011313246010.3390/rs13132460RUF: Effective Sea Ice Floe Segmentation Using End-to-End RES-UNET-CRF with Dual LossAnmol Sharan Nagi0Devinder Kumar1Daniel Sola2K. Andrea Scott3Department of Systems Design Engineering, University of Waterloo, Waterloo, ON N2L G31, CanadaSchool of Medicine, Stanford University, Stanford, CA 94305, USADepartment of Systems Design Engineering, University of Waterloo, Waterloo, ON N2L G31, CanadaDepartment of Systems Design Engineering, University of Waterloo, Waterloo, ON N2L G31, CanadaSea ice observations through satellite imaging have led to advancements in environmental research, ship navigation, and ice hazard forecasting in cold regions. Machine learning and, recently, deep learning techniques are being explored by various researchers to process vast amounts of Synthetic Aperture Radar (SAR) data for detecting potential hazards in navigational routes. Detection of hazards such as sea ice floes in Marginal Ice Zones (MIZs) is quite challenging as the floes are often embedded in a multiscale ice cover composed of ice filaments and eddies in addition to floes. This study proposes a segmentation model tailored for detecting ice floes in SAR images. The model exploits the advantages of both convolutional neural networks and convolutional conditional random field (Conv-CRF) in a combined manner. The residual UNET (RES-UNET) computes expressive features to generate coarse segmentation maps while the Conv-CRF exploits the spatial co-occurrence pairwise potentials along with the RES-UNET unary/segmentation maps to generate final predictions. The whole pipeline is trained end-to-end using a dual loss function. This dual loss function is composed of a weighted average of binary cross entropy and soft dice loss. The comparison of experimental results with the conventional segmentation networks such as UNET, DeepLabV3, and FCN-8 demonstrates the effectiveness of the proposed architecture.https://www.mdpi.com/2072-4292/13/13/2460sea iceice floeSARdeep learningconditional random fieldssegmentation |
| spellingShingle | Anmol Sharan Nagi Devinder Kumar Daniel Sola K. Andrea Scott RUF: Effective Sea Ice Floe Segmentation Using End-to-End RES-UNET-CRF with Dual Loss Remote Sensing sea ice ice floe SAR deep learning conditional random fields segmentation |
| title | RUF: Effective Sea Ice Floe Segmentation Using End-to-End RES-UNET-CRF with Dual Loss |
| title_full | RUF: Effective Sea Ice Floe Segmentation Using End-to-End RES-UNET-CRF with Dual Loss |
| title_fullStr | RUF: Effective Sea Ice Floe Segmentation Using End-to-End RES-UNET-CRF with Dual Loss |
| title_full_unstemmed | RUF: Effective Sea Ice Floe Segmentation Using End-to-End RES-UNET-CRF with Dual Loss |
| title_short | RUF: Effective Sea Ice Floe Segmentation Using End-to-End RES-UNET-CRF with Dual Loss |
| title_sort | ruf effective sea ice floe segmentation using end to end res unet crf with dual loss |
| topic | sea ice ice floe SAR deep learning conditional random fields segmentation |
| url | https://www.mdpi.com/2072-4292/13/13/2460 |
| work_keys_str_mv | AT anmolsharannagi rufeffectiveseaicefloesegmentationusingendtoendresunetcrfwithdualloss AT devinderkumar rufeffectiveseaicefloesegmentationusingendtoendresunetcrfwithdualloss AT danielsola rufeffectiveseaicefloesegmentationusingendtoendresunetcrfwithdualloss AT kandreascott rufeffectiveseaicefloesegmentationusingendtoendresunetcrfwithdualloss |