Identifying Fault-Related Deformation in Large InSAR Datasets Using Transfer Learning
- 3 February 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 1-12
Abstract
The rapid expansion of SAR archives (e.g., Sentinel-1) offers new opportunities to detect fault-related deformation signals, such as weak coseismic-like patterns and interseismic shallow fault creep. However, their low signal-to-noise ratio and coherence variability pose challenges for large-scale screening. To address this, we develop a transfer learning–based convolutional neural network for automatic recognition of fault-related deformation in single wrapped interferograms. Built on SarNet and fine-tuned with globally curated Sentinel-1 data, the model achieves 92.25% validation accuracy (AUC 0.9689) with a low false-alarm rate. Cross-mission tests on ALOS-2 (L-band) and ENVISAT (C-band) confirm strong portability (accuracy 87.24%, precision 95.54%, recall 78.13%, F1 85.96%, AUC 0.9686), though recall decreases due to sensor-specific domain shifts. Grad-CAM visualizations indicate that the classifier relies on meaningful interferometric fringes rather than noise artifacts, supporting physical interpretability. These results show that transfer learning enables robust and scalable automated recognition of fault-related deformation, thereby supporting systematic, large-scale mining of multi-sensor InSAR archives.Keywords
Funding Information
- Basic Scientific Funding of the Institute of Geology, China Earthquake Administration (IGCEA2433)
- National Natural Science Foundation of China (U2139202)
- State Key Laboratory of Earthquake Dynamics and Forecasting (LED2023A06)
This publication has 69 references indexed in Scilit:
- Newly discovered shallow creep along the Gozha Co fault in northwestern Tibet: Spatial extent, rate and temporal evolutionEarth and Planetary Science Letters, 2023
- Optimal ratio for data splittingStatistical Analysis and Data Mining: An ASA Data Science Journal, 2022
- Overfitting, Model Tuning, and Evaluation of Prediction Performance2022
- Deep Learning: A Comprehensive Overview on Techniques, Taxonomy, Applications and Research DirectionsSN Computer Science, 2021
- Geodetic imaging of shallow creep along the Xianshuihe fault and its frictional propertiesEarth and Planetary Science Letters, 2021
- A Survey on Deep Transfer Learning2018
- The geophysics, geology and mechanics of slow fault slipEarth and Planetary Science Letters, 2018
- Interseismic strain accumulation across the central North Anatolian Fault from iteratively unwrapped InSAR measurementsJournal of Geophysical Research: Solid Earth, 2016
- Localized and distributed creep along the southern San Andreas FaultJournal of Geophysical Research: Solid Earth, 2014
- Large-Scale Machine Learning with Stochastic Gradient Descent2010