Automatic Segmentation of Ice Shelves with Deep Learning
Radar sounders (RSs) provide information on the subsurface of the cryosphere through the use of electromagnetic (EM) signals by producing radargrams. Radargrams are used to detect and analyze relevant targets in the subsurface of icy regions. Up to now, studies of the subsurface structure of the cry...
| Published in: | 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS |
|---|---|
| Main Authors: | , , |
| Format: | Conference Object |
| Language: | English |
| Published: |
IEEE
2021
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/11582/328568 https://doi.org/10.1109/IGARSS47720.2021.9553610 https://ieeexplore.ieee.org/document/9553610 |
| Summary: | Radar sounders (RSs) provide information on the subsurface of the cryosphere through the use of electromagnetic (EM) signals by producing radargrams. Radargrams are used to detect and analyze relevant targets in the subsurface of icy regions. Up to now, studies of the subsurface structure of the cryosphere with radargrams have been conducted manually or applying semiautomatic techniques. However, these techniques present efficiency and adaptability disadvantages. To overcome these issues, we propose automatic analysis techniques for radargrams of icy regions based on deep learning (DL). Experimental analysis is conducted for the automatic segmentation of areas of interest in radargrams of ice shelves of coastal areas acquired by the radar sounder MCoRDS2. |
|---|