Search

Ground Penetrating Radar (GPR) is a non-destructive geophysical technique utilizing electromagnetic pulses to detect subsurface material properties. The analysis of regions of interest (ROIs) in GPR images often entails the identification of hyperbolic reflection regions of underground targets through accurate segmentation, a crucial preprocessing step. Currently, this represents a research gap. In the hyperbolic reflection region, manual segmentation not only demands professional expertise but is also time-consuming and error-prone. Automatic segmentation can aid in accurately determining the location and depth of the reflection region, thereby enhancing data interpretation and analysis. This study presents a deep residual Convolutional Neural Network (Res-CNN) that integrates skip connections within an encoder-decoder framework for ROI-binarized hyperbolic segmentation. The proposed framework includes designed downsampling and upsampling modules that facilitate feature computation sharing between these two modules through skip connections within network blocks. In the evaluation of both simple and complex models, our method attained PSNR, SSIM, and FSIM values of 57.1894, 0.9933, and 0.9336, and 58.4759, 0.9958, and 0.9677, respectively. Compared to traditional segmentation methods, the proposed approach demonstrated clearer segmentation results, enabling intelligent and effective identification of the ROI region containing abnormal hyperbolic reflection waves in GPR images. [ABSTRACT FROM AUTHOR]

Data assimilation for unconfined seepage analysis has faced significant challenges due to hybrid causes, such as sparse measurements, heterogeneity of porous media, and computationally expensive forward models. To address these bottlenecks, this paper introduces a physics-informed neural network (PINN) model to resolve the data assimilation problem for seepage analysis of unsaturated earth–rockfill dams. This strategy offers a solution that decreases the reliance on numerical models and enables an accurate and efficient prediction of seepage parameters for complex models in the case of sparse observational data. For the first attempt in this study, the observed values are obtained by random sampling of numerical solutions, which are then contributed to the synchronous constraints in the loss function by informing both the seepage control equations and boundary conditions. To minimize the effects of sharp gradient shifts in seepage parameters within the research domain, a residual adaptive refinement (RAR) constraint is introduced to strategically allocate training points around positions with significant residuals in partial differential equations (PDEs), which could facilitate enhancing the prediction accuracy. The model's effectiveness and precision are evaluated by analyzing the proposed strategy against the numerical solutions. The results indicate that even with limited sparse data, the PINN model has great potential to predict seepage data and identify complex structures and anomalies inside the dam. By incorporating coupling constraints, the validity of our PINN model could lead to theoretically viable applications of hydrogeophysical inversion or multi-parameter seepage inversion. The results show that the proposed framework can predict the seepage parameters for the entire research domain with only a small amount of observation data. Furthermore, with a small amount of observation data, PINNs are able to obtain more accurate results than purely data-driven DNNs. [ABSTRACT FROM AUTHOR]

As a fundamental part of ground penetrating radar (GPR) data processing, reverse time migration (RTM) can correctly position reflection waves and focusing diffraction waves on the proper spatial position. Least-squares reverse-time migration (LSRTM) is widely used in the seismic field for its ability to suppress artifacts and generate high-resolution images in comparison to conventional RTM. However, in the particular case of GPR detection methods, LSRTM is extremely susceptible to aliasing artifacts caused by under-sampling. In pursuit of enhanced precision in underground structure characterization, this paper presents the development of a new LSRTM based on modified total variation (MTV) regularization to improve imaging resolution. Initially, the objective function of LSRTM is derived by combining the Born approximation in 2-D transversal magnetic mode. Next, the adjoint equations and their gradients are solved using the Lagrange multiplier method. The objective function is then constrained by MTV regularization to ensure the precision and convergence of the LSRTM, which delivers a refined edge with reconstruction details. In the numerical experiments, in comparison to the conventional LSRTM method, the LSRTM-MTV algorithm demonstrated a 30.4% increase in computational speed and a 21.1% reduction in mean squared error (MSE). The outperformance of the proposed method is verified in detail through the image resolution and amplitude preservation in the test of synthetic data and laboratory data. Future research efforts will center on applying the proposed method to models featuring dispersive or anisotropic media that closely mimic real-world conditions and extending the application to various imaging techniques involving objective function minimization. [ABSTRACT FROM AUTHOR]

Estimating the quantity of seepage through the foundation and body of a dam using proper health and safety monitoring is critical to the effective management of disaster risk in a reservoir downstream of the dam. In this study, a deep learning model was constructed to predict the extent of seepage through Pakistan's Tarbela dam, the world's second largest clay and rock dam. The dataset included hydro-climatological, geophysical, and engineering characteristics for peak-to-peak water inflows into the dam from 2014 to 2020. In addition, the data are time series, recurring neural networks (RNN), and long short-term memory (LSTM) as time series algorithms. The RNN–LSTM model has an average mean square error of 0.12, and a model performance of 0.9451, with minimal losses and high accuracy, resulting in the best-predicted dam seepage result. Damage was projected using a deep learning system that addressed the limitations of the model, the difficulties of calculating human activity schedules, and the need for a different set of input data to make good predictions. [ABSTRACT FROM AUTHOR]

The chaos oscillation particle swarm optimization (COPSO) algorithm is prone to binge trapped in the local optima when dealing with certain complex models in ground-penetrating radar (GPR) data inversion, because it inherently suffers from premature convergence, high computational costs, and extremely slow convergence times, especially in the middle and later periods of iterative inversion. Considering that the bilateral connections between different particle positions can improve both the algorithmic searching efficiency and the convergence performance, we first develop a fast single-trace-based approach to construct an initial model for 2-D PSO inversion and then propose a TV-regularization-based improved PSO (TVIPSO) algorithm that employs total variation (TV) regularization as a constraint technique to adaptively update the positions of particles. B by adding the new velocity variations and optimal step size matrices, the search range of the random particles in the solution space can be significantly reduced, meaning blindness in the search process can be avoided. By introducing constraint-oriented regularization to allow the optimization search to move out of the inaccurate region, the premature convergence and blurring problems can be mitigated to further guarantee the inversion accuracy and efficiency. We report on three inversion experiments involving multilayered, fluctuated terrain models and a typical complicated inner-interface model to demonstrate the performance of the proposed algorithm. The results of the fluctuated terrain model show that compared with the COPSO algorithm, the fitness error (MAE) of the TVIPSO algorithm is reduced from 2.3715 to 1.0921, while for the complicated inner-interface model the fitness error (MARE) of the TVIPSO algorithm is reduced from 1.9539 to 1.5674. [ABSTRACT FROM AUTHOR]