Identification of Surface Deformation in InSAR Using Machine Learning

Brengman, Clayton; Barnhart, William D.

doi:10.1029/2020gc009204

Cited by 35 publications

(20 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Brengman et al used a SarNet CNN with transfer learning from a synthetic dataset to a real dataset. SarNet achieved an accuracy of 85.22% on a real interferogram dataset [26].…”

Section: Transfer Learningmentioning

confidence: 98%

“…The system developed is able to detect rapid large scale deformations, but cannot detect deformations that are slow or of a small scale nature [24]. Anantrasirichai et al [25] and Brengman et al [26] both used transfer learning in other ground deformation applications [25,26]. Brengman et al used a SarNet CNN with transfer learning from a synthetic dataset to a real dataset.…”

Section: Transfer Learningmentioning

confidence: 99%

See 1 more Smart Citation

Deep/Transfer Learning with Feature Space Ensemble Networks (FeatSpaceEnsNets) and Average Ensemble Networks (AvgEnsNets) for Change Detection Using DInSAR Sentinel-1 and Optical Sentinel-2 Satellite Data Fusion

Karim

Zyl

2021

Remote Sensing

View full text Add to dashboard Cite

Differential interferometric synthetic aperture radar (DInSAR), coherence, phase, and displacement are derived from processing SAR images to monitor geological phenomena and urban change. Previously, Sentinel-1 SAR data combined with Sentinel-2 optical imagery has improved classification accuracy in various domains. However, the fusing of Sentinel-1 DInSAR processed imagery with Sentinel-2 optical imagery has not been thoroughly investigated. Thus, we explored this fusion in urban change detection by creating a verified balanced binary classification dataset comprising 1440 blobs. Machine learning models using feature descriptors and non-deep learning classifiers, including a two-layer convolutional neural network (ConvNet2), were used as baselines. Transfer learning by feature extraction (TLFE) using various pre-trained models, deep learning from random initialization, and transfer learning by fine-tuning (TLFT) were all evaluated. We introduce a feature space ensemble family (FeatSpaceEnsNet), an average ensemble family (AvgEnsNet), and a hybrid ensemble family (HybridEnsNet) of TLFE neural networks. The FeatSpaceEnsNets combine TLFE features directly in the feature space using logistic regression. AvgEnsNets combine TLFEs at the decision level by aggregation. HybridEnsNets are a combination of FeatSpaceEnsNets and AvgEnsNets. Several FeatSpaceEnsNets, AvgEnsNets, and HybridEnsNets, comprising a heterogeneous mixture of different depth and architecture models, are defined and evaluated. We show that, in general, TLFE outperforms both TLFT and classic deep learning for the small dataset used and that larger ensembles of TLFE models do not always improve accuracy. The best performing ensemble is an AvgEnsNet (84.862%) comprised of a ResNet50, ResNeXt50, and EfficientNet B4. This was matched by a similarly composed FeatSpaceEnsNet with an F1 score of 0.001 and variance of 0.266 less. The best performing HybridEnsNet had an accuracy of 84.775%. All of the ensembles evaluated outperform the best performing single model, ResNet50 with TLFE (83.751%), except for AvgEnsNet 3, AvgEnsNet 6, and FeatSpaceEnsNet 5. Five of the seven similarly composed FeatSpaceEnsNets outperform the corresponding AvgEnsNet.

show abstract

“…Brengman et al used a SarNet CNN with transfer learning from a synthetic dataset to a real dataset. SarNet achieved an accuracy of 85.22% on a real interferogram dataset [26].…”

Section: Transfer Learningmentioning

confidence: 98%

Section: Transfer Learningmentioning

confidence: 99%

Deep/Transfer Learning with Feature Space Ensemble Networks (FeatSpaceEnsNets) and Average Ensemble Networks (AvgEnsNets) for Change Detection Using DInSAR Sentinel-1 and Optical Sentinel-2 Satellite Data Fusion

Karim

Zyl

2021

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…In general, CNN, which is a learning model that can extract features from data, is composed of five main layers, namely, 1) a convolution layer, 2) a normalization layer, 3) an activation layer, 4) a pooling layer, and 5) a linear layer, as shown in Figure 3 (Brengman and Barnhart, 2021). The convolutional layer is an important element in CNN and is combined with filters to analyze the input image.…”

Section: Convolutional Neural Networkmentioning

confidence: 99%

Surface deformation simulation for InSAR detection using a machine learning approach on the hantangang river volcanic field: A case study on the orisan mountain

Fadhillah

Hakim

Park

et al. 2022

Front. Environ. Sci.

View full text Add to dashboard Cite

Recent developments in remote sensing research have resulted in a large amount of variability in the data provided by researchers. Synthetic aperture radar (SAR) is a tool used to measure surface deformation and assess changes in the Earth’s surface. Here, we consider the usefulness of Interferometric Synthetic Aperture Radar (InSAR) in assessing past volcanic activity as a key to learning the characteristics of the deformation around a volcano. The Hantangang River volcanic field (HRVF) is a geoheritage site in the Korean Peninsula that has interesting geological characteristics. This volcanic field has formed along 110 km of the paleochannel of the Hantangang River. Since the eruptions occurred from 0.15 to 0.51 Ma, the source is limited, which has raised interest in the assessment of volcanic landforms. The recent integration of machine learning and InSAR processing has shown promising results for many purposes, such as classifying, modeling, and detecting surface deformation. To examine the future impact based on information from the past, we utilized a synthetic interferogram with the Okada model and transferred it to a machine learning algorithm. The synthetic interferogram was formed based on Sentinel-1 C-band satellite data to simulate the deformation phases. The orbital errors, the topographical data errors, and the atmospheric effect were also simulated and added to the synthetic interferogram to enrich the learning input. A convolutional neural network (CNN) trained with the unwrapped simulated interferogram data and its performance was evaluated. Our proposed method exhibits the capability to detect volcanic activity’s deformation patterns with synthetic interferogram data. The results show that an overall accuracy of more than 80% was achieved using the CNN algorithms on the validation dataset. This study is the first to use machine learning approaches for detecting prehistorical volcanic deformation and demonstrates potential techniques for developing an approach based on satellite imagery. In addition, this study has introduced the possibility of developing a rapid detection of surface deformation using InSAR data based on a machine learning approach.

show abstract

“…DL has also been applied in geology, such as in the study of surface deformation caused by volcanic activity. For example, Anantrasirichai et al [21,22] used a convolutional neural network (CNN) trained with dichotomous (i.e., "background" and "volcanic") markers from InSAR observations to quickly identify ground deformations caused by volcanic activity; Valade et al [23] used a CNN to learn interferogram features to detect intense deformations in real interferograms and demonstrated the CNN's accuracy through a series of recent volcanic eruptions; Matthew et al [24] developed a "two-headed model" that can locate and classify volcanic deformations in a single interferogram; and Clayton et al [25] used CNN networks to classify surface deformations in synthetic interferograms of volcanoes and used CAMs (class activation maps) to show the location of surface deformations. Regarding landslide identification based on InSAR technology, KJ et al [26] used the AlexNet model to classify the interferometric stripes of landslide motion and then predict the landslide boundaries.…”

Section: Introductionmentioning

confidence: 99%

Study of the Automatic Recognition of Landslides by Using InSAR Images and the Improved Mask R-CNN Model in the Eastern Tibet Plateau

Liu

Yao

et al. 2022

Remote Sensing

View full text Add to dashboard Cite

The development of landslide hazards is spatially scattered, temporally random, and poorly characterized. Given the advantages of the large spatial scale and high sensitivity of InSAR observations, InSAR is becoming one of the main techniques for active landslide identification. The difficult problem is how to quickly extract landslide information from extensive InSAR image data. Since the instance segmentation model (Mask R-CNN) in deep learning can provide highly robust target recognition, we select the landslide-prone eastern edge of the Tibetan Plateau as a specific test area. Introducing and optimizing this model achieves high-speed and accurate recognition of InSAR observations. First, the InSAR patch landslide instance segmentation dataset (SLD) is established by developing a common object in context (COCO) annotation format conversion code based on InSAR observations. The Mask R-CNN+++ is found by adding three functions of the ResNext module to increase the fineness of the network segmentation results and enhance the noise resistance of the model, the DCB (deformable convolutional block) to improve the feature extraction ability of the network for geometric morphological changes of landslide patches, and an attention mechanism to selectively enhance usefully and suppress features less valuable to the native Mask R-CNN network. The model achieves 92.94% accuracy on the test set, and the active landslide recognition speed based on this model under ordinary computer hardware conditions is 72.3 km2/s. The overall characteristics of the results of this study show that the optimized model effectively enhances the perceptibility of image morphological changes, thereby resulting in smoother recognition boundaries and further improvement of the generalization ability of segmentation detection. This result is expected to serve to identify and monitor active landslides in complex surface conditions on a large spatial scale. Moreover, active landslides of different geometric features, motion patterns, and intensities are expected to be further segmented.

show abstract

Identification of Surface Deformation in InSAR Using Machine Learning

Cited by 35 publications

References 40 publications

Deep/Transfer Learning with Feature Space Ensemble Networks (FeatSpaceEnsNets) and Average Ensemble Networks (AvgEnsNets) for Change Detection Using DInSAR Sentinel-1 and Optical Sentinel-2 Satellite Data Fusion

Deep/Transfer Learning with Feature Space Ensemble Networks (FeatSpaceEnsNets) and Average Ensemble Networks (AvgEnsNets) for Change Detection Using DInSAR Sentinel-1 and Optical Sentinel-2 Satellite Data Fusion

Surface deformation simulation for InSAR detection using a machine learning approach on the hantangang river volcanic field: A case study on the orisan mountain

Study of the Automatic Recognition of Landslides by Using InSAR Images and the Improved Mask R-CNN Model in the Eastern Tibet Plateau

Contact Info

Product

Resources

About