Search

Monitoring volcanic deformations allows us to track dynamic states of a volcano and to know where an eruptions could happen. Spaceborne Synthetic Aperture Radar (SAR) and SAR interferometry (InSAR) techniques created an opportunity to track volcanoes globally, even in inaccessible regions without ground measuring stations.This paper proposes a convolutional neural network (CNN) for detection of volcanic deformations in InSAR velocity maps. We had only a small amount of velocity maps over the region of central South American Andes, therefore the synthetic data are used to train the model from scratch. In the region of interest, the velocity maps contain the patterns of salt lakes and slope induced signal which confuse CNN models trained on synthetic data.In order to bridge the gap between the synthetic and real data, the hybrid synthetic-real data set is used for fine-tuning the model. The hybrid set consists of the real background signal data and synthetic volcanic data. Four fine-tuning sets which were created by different combinations of the original hybrid data, the filtered hybrid data, and simulated data have been used and compared with each other. Besides, we compared four fine-tuning approaches to determine where and how to fine-tune the model. Results show significant improvement in performance by majority of the approaches, and training the last or last two layers have given the best results. In addition, using the FT1 (containing only hybrid set), and FT4 (containing all sets) improved the area under the curve receiver operating characteristic (AUC ROC) from 55% to 86% and 88% respectively.

TecVolSA (Tectonics and Volcanoes in South America) is a project with a goal of developing intelligent Earth Observation (EO) data processing and exploitation for monitoring various geophysical processes in central south American Andes. Large amount of Sentinel-1 data over the period of about 5 years has been processed using mixed Permanent Scatterer and Distributed Scatterer (PS/DS) approaches. The received products are velocity maps with InSAR relative error in the order of 1 mm/yr on a large scale (>100km). The second milestone of the project was automatic extraction of information from the data. In this work, the focus is on detecting volcanic deformations. Since the real data prepared in such manner is limited, to train a deep learning model for detection of volcanic deformations, a synthetic training set is used. Models are trained from scratch and InceptionResNet v2 was selected for further experiments as it was found to give best performance among the tested models. The explainable AI (XAI) techniques were used to understand and analyze the confidence of the model and to understand how to improve it. The models trained on synthetic training set underperformed on real test set. Using GradCAM technique, it was identified that slope induced signal and salt lake deformations were mistakenly identified as volcanic deformations. These patterns are difficult to simulate and were not contained in synthetic training set. Bridging this distribution gap was performed using hybrid synthetic-real fine-tuning set, consisting of the real slope induced signal data and synthetic volcanic data. Additionally, false positive rate of the model is reduced using low-pass spatial filtering of the real test set, and finally by adjustments of the temporal baseline received from a sensitivity analysis. The model successfully detected all 10 deforming volcanoes in the region, ranging from 0.4 - 1.8 cm/yr in deformation.