TSSA: Two-Step Semi-Supervised Annotation for Radargrams on the Greenland Ice Sheet

Thumbnail Image

Files

TSSA_Two-Step_Semi-Supervised_Annotation_for_Radargrams_on_the_Greenland_Ice_Sheet.pdf (1.18 MB)

Links to Files

https://ieeexplore.ieee.org/document/10282880

Permanent Link

https://doi.org/10.1109/IGARSS52108.2023.10282880
 http://hdl.handle.net/11603/35807

Collections

UMBC Information Systems Department
UMBC College of Engineering and Information Technology Dean's Office
UMBC Faculty Collection
UMBC Student Collection

Author/Creator

Author/Creator ORCID

https://orcid.org/0000-0003-4043-2193
 https://orcid.org/0000-0002-1821-6438
 https://orcid.org/0000-0003-0130-6135

Date

2023-10-20

Type of Work

conference papers and proceedings
Text

Department

Program

Citation of Original Publication

Jebeli, Atefeh, Bayu Adhi Tama, Vandana P. Janeja, Nicholas Holschuh, Claire Jensen, Mathieu Morlighem, Joseph A. MacGregor, and Mark A. Fahnestock. “TSSA: Two-Step Semi-Supervised Annotation for Radargrams on the Greenland Ice Sheet.” In IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium, 56–59, 2023. https://doi.org/10.1109/IGARSS52108.2023.10282880.

Rights

This work was written as part of one of the author's official duties as an Employee of the United States Government and is therefore a work of the United States Government. In accordance with 17 U.S.C. 105, no copyright protection is available for such works under U.S. Law.

Subjects

ice sheet
Training
Computational modeling
supervised learning
Deep learning
unsupervised learning
Annotations
Data augmentation
Transfer learning
Transformers
deep neural network
ice penetrating radar

Abstract

Ice-penetrating radar surveys have been conducted across the Greenland Ice Sheet since the 1960s, producing radargrams that measure ice thickness and detect the ice sheet’s radiostratigraphy. However, these radargrams are relatively under-explored and not yet fully annotated, mapped, or interpreted glaciologically. We aim to move towards automatic radargram annotation using deep learning-based methods. To provide a training set for these methods, we develop a two-step semi-supervised annotation (TSSA) approach that uses an existing unsupervised layer annotation (ARESELP) method and a deep learning-based segmentation approach (U-Net) to detect surface, and bottom reflectors (representing the bedrock) layers in radargrams. Here we focus on two evaluations of our approach: 1. Surface and bottom annotations; and 2. Data augmentation and transfer learning techniques for improving the performance of deep learning methods. Our study is a foundation for improving the efficacy of AI-based methods for auto-annotation of radargrams, where the training set is generated seamlessly through unsupervised learning.