Airborne radar sounders (RSs) are active sensors that acquire subsurface data for Earth observation. RS data (radargrams) provide information on buried geology by imaging subsurface dielectric discontinuities. Recently, several automatic RS target identification techniques have been proposed, with convolutional neural network (CNN)-based methods being the most promising. However, they require numerous labeled data that are hard to retrieve in the subsurface environment targeted by RS. Furthermore, they are not designed to effectively deal with problems showing unbalanced classes, such as RS segmentation. We introduce newer cryosphere subsurface targets in the inland and coastal areas that can have a very low probability. To deal with the higher complexity and variability than previous works, we propose a transfer learning framework for RS data to mitigate the need for a large amount of labeled data and handle extremely unbalanced target classes. Herewith, we propose two transfer learning-based mechanisms for radargram segmentation. The first uses a lightweight architecture whose pretraining is supervised with a large labeled dataset from other domains. The second mechanism uses a deep architecture pretrained in the RS domain, considering the pretest task of radargram reconstruction. The architectures are modified to deal with the characteristics of RS data and the radargram segmentation task. Finally, both methods are fine-tuned with a few labeled radargrams to learn radargram features useful for segmentation. We reveal experimental results on radargrams acquired in Antarctica by MCoRDS-1 and MCoRDS-3. The results demonstrate the effectiveness of transfer learning for radargram segmentation.
A weakly supervised transfer learning approach for radar sounder data segmentation / Hoyo, Garcia; Donini, E.; Bovolo, F.. - In: IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING. - ISSN 1558-0644. - 61:(2023), pp. 450101801-450101818. [10.1109/TGRS.2023.3252939]
A weakly supervised transfer learning approach for radar sounder data segmentation
Hoyo Garcia;Donini E.;Bovolo F.
2023-01-01
Abstract
Airborne radar sounders (RSs) are active sensors that acquire subsurface data for Earth observation. RS data (radargrams) provide information on buried geology by imaging subsurface dielectric discontinuities. Recently, several automatic RS target identification techniques have been proposed, with convolutional neural network (CNN)-based methods being the most promising. However, they require numerous labeled data that are hard to retrieve in the subsurface environment targeted by RS. Furthermore, they are not designed to effectively deal with problems showing unbalanced classes, such as RS segmentation. We introduce newer cryosphere subsurface targets in the inland and coastal areas that can have a very low probability. To deal with the higher complexity and variability than previous works, we propose a transfer learning framework for RS data to mitigate the need for a large amount of labeled data and handle extremely unbalanced target classes. Herewith, we propose two transfer learning-based mechanisms for radargram segmentation. The first uses a lightweight architecture whose pretraining is supervised with a large labeled dataset from other domains. The second mechanism uses a deep architecture pretrained in the RS domain, considering the pretest task of radargram reconstruction. The architectures are modified to deal with the characteristics of RS data and the radargram segmentation task. Finally, both methods are fine-tuned with a few labeled radargrams to learn radargram features useful for segmentation. We reveal experimental results on radargrams acquired in Antarctica by MCoRDS-1 and MCoRDS-3. The results demonstrate the effectiveness of transfer learning for radargram segmentation.File | Dimensione | Formato | |
---|---|---|---|
8_PDFsam_Wallace_virusDroseu_VirEvolA_Weakly_Supervised_Transfer_Learning_Approach_for_Radar_Sounder_Data_Segmentation.pdf
Solo gestori archivio
Tipologia:
Versione editoriale (Publisher’s layout)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
405.33 kB
Formato
Adobe PDF
|
405.33 kB | Adobe PDF | Visualizza/Apri |
1_PDFsam_Wallace_virusDroseu_VirEvolA_Weakly_Supervised_Transfer_Learning_Approach_for_Radar_Sounder_Data_Segmentation.pdf
Solo gestori archivio
Tipologia:
Versione editoriale (Publisher’s layout)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
10.13 MB
Formato
Adobe PDF
|
10.13 MB | Adobe PDF | Visualizza/Apri |
9_PDFsam_Wallace_virusDroseu_VirEvolA_Weakly_Supervised_Transfer_Learning_Approach_for_Radar_Sounder_Data_Segmentation_compressed1.pdf
Solo gestori archivio
Tipologia:
Versione editoriale (Publisher’s layout)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
162.48 kB
Formato
Adobe PDF
|
162.48 kB | Adobe PDF | Visualizza/Apri |
18_PDFsam_Wallace_virusDroseu_VirEvolA_Weakly_Supervised_Transfer_Learning_Approach_for_Radar_Sounder_Data_Segmentation.pdf
Solo gestori archivio
Tipologia:
Versione editoriale (Publisher’s layout)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
3.23 MB
Formato
Adobe PDF
|
3.23 MB | Adobe PDF | Visualizza/Apri |
15_PDFsam_Wallace_virusDroseu_VirEvolA_Weakly_Supervised_Transfer_Learning_Approach_for_Radar_Sounder_Data_Segmentation.pdf
Solo gestori archivio
Tipologia:
Versione editoriale (Publisher’s layout)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
7.51 MB
Formato
Adobe PDF
|
7.51 MB | Adobe PDF | Visualizza/Apri |
10_PDFsam_Wallace_virusDroseu_VirEvolA_Weakly_Supervised_Transfer_Learning_Approach_for_Radar_Sounder_Data_Segmentation.pdf
Solo gestori archivio
Tipologia:
Versione editoriale (Publisher’s layout)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
7.58 MB
Formato
Adobe PDF
|
7.58 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione