Papers by Mahmoud El Hussieny
Automated deformation detection and interpretation using InSAR data and a multi-task ViT model
International journal of applied earth observation and geoinformation, Apr 1, 2024
Many geological hazards are associated with ground deformations. Prompt and accurate detection an... more Many geological hazards are associated with ground deformations. Prompt and accurate detection and interpretation of ground deformation is therefore vital to geohazard mitigation. Multitemporal Interferometric Synthetic Aperture Radar (MT-InSAR) is an effective geodetic technique for monitoring ground deformation. However, accurate computation and interpretation of deformation using InSAR are often hindered by various errors and a lack of expert knowledge. We present a new advanced deep learning model based on a multi-task vision transformer (MT-ViT) to automatically detect, locate, and interpret deformation using single interferograms. To address the issue of limited training data in InSAR applications, the proposed model utilizes pretrained weights from optical images and transfers them to a simulated InSAR dataset. Then real interferograms are used to fine-tune the weights in the network. An overall loss function is designed, which considers the classification and localization losses in the model. The effectiveness of the proposed model is demonstrated using both simulated and real InSAR datasets that contain either coseismic or volcanic deformation. The experimental results from the model are also compared with the state-of-the-art convolutional neural network (CNN) based techniques. The results show significant improvement in both the accuracy of the results and the computational efficiency over the CNN-based approaches. The MT-ViT model achieved 99.4 % classification accuracy, 54.1 % mean intersection over union (IOU), and 0.9 km localization accuracy. A comprehensive evaluation of the hyperparameters in training the MT-ViT model was carried out, which will inform future research in this direction. The research results highlight the promising capabilities of MT-ViT in near real-time deformation monitoring and automated deformation interpretation.
Many geological hazards are associated with ground deformations. Prompt and accurate detection an... more Many geological hazards are associated with ground deformations. Prompt and accurate detection and interpretation of ground deformation is therefore vital to geohazard mitigation. Multitemporal Interferometric Synthetic Aperture Radar (MT-InSAR) is an effective geodetic technique for monitoring ground deformation. However, accurate computation and interpretation of deformation using InSAR are often hindered by various errors and a lack of expert knowledge. We present a new advanced deep learning model based on a multi-task vision transformer (MT-ViT) to automatically detect, locate, and interpret deformation using single interferograms. To address the issue of limited training data in InSAR applications, the proposed model utilizes pretrained weights from optical images and transfers them to a simulated InSAR dataset. Then real interferograms are used to fine-tune the weights in the network. An overall loss function is designed, which considers the classification and localization losses in the model. The effectiveness of the proposed model is demonstrated using both simulated and real InSAR datasets that contain either coseismic or volcanic deformation. The experimental results from the model are also compared with the state-of-the-art convolutional neural network (CNN) based techniques. The results show significant improvement in both the accuracy of the results and the computational efficiency over the CNN-based approaches. The MT-ViT model achieved 99.4 % classification accuracy, 54.1 % mean intersection over union (IOU), and 0.9 km localization accuracy. A comprehensive evaluation of the hyperparameters in training the MT-ViT model was carried out, which will inform future research in this direction. The research results highlight the promising capabilities of MT-ViT in near real-time deformation monitoring and automated deformation interpretation.
Land deformation monitoring by GNSS in the Nile Delta and the measurements analysis, 2021
Nile Delta, one of the most populated deltas around the world, is suffering from subsidence due t... more Nile Delta, one of the most populated deltas around the world, is suffering from subsidence due to the natural compaction of its sediment. Land subsidence has a great impact on the infrastructure, economic, and social. This study investigatesthe evaluation of the land subsidence in Nile Delta using GNSS measurements. Eight stations of GNSS points through period 2013-2015 distributed around Nile Delta and the Northern part of Egypt are utilized using differential GPS with nine IGS stations by GAMIT/GLOBK V. 10.61. The main objective is to monitor the spatiotemporal variations of land surface within the Nile Delta by time series analysis of the ellipsoidal height. The northern part of Egypt can be divided into three major parts; the western part shows subsidence rate of 2 mm/year downward; Nile Deltaof 2.5 to 10 mm/year downward from west to east and the eastern part shows an uplift.
Assessment of NRCAN PPP online service in determination of crustal velocity: case study Northern Egypt GNSS Network, 2021
Precise point positioning (PPP) has the ability to give precise positioning with high accuracy. T... more Precise point positioning (PPP) has the ability to give precise positioning with high accuracy. That may be an alternative to precise Deferential GNSS (DGNSS), in addition to being a low-cost alternative among all processing strategies of Global Navigation Satellite System (GNSS), especially when online services are used. To see the assessment study of PPP in the determination of crustal velocity, 66 days of GNSS data distributed in the period 2013-2015 were used. Composing a network of eight stations called the Northern Egyptian Permanent GNSS Network (N-EPGN) with nine IGS stations was processed. Scientific software GAMIT/GLOBK and Bernese used to calculate the final precise coordinate and the associated velocity of each station in the International Terrestrial Reference Frame (ITRF 2008). Each campaign consists of 3 days. These data were processed by PPP approach. The final precise coordinate and the associated velocity of each EPGN station estimated. The final results compare to those different methods of analysis and programs refer to high level of agreement between the coordinates and velocity which confirm that PPP approach can be applied for the investigation of crustal deformation.
Uploads
Papers by Mahmoud El Hussieny