Ground Deformation Forecasting and Modeling in Mining Areas Using Artificial Intelligence Techniques

This study aimed to develop and evaluate predictive models for injury prevention in athletes using advanced artificial intelligence (AI) techniques. The research focused on harnessing AI to identify injury risk factors, predict potential injuries, and ultimately enhance athlete safety and performance. We initiated the study by collecting extensive data from 500 athletes over a two-year period, amassing more than 1,000,000 data points. The dataset included diverse variables such as physiological metrics (heart rate, VO2 max), biomechanical data (joint angles, muscle activation patterns), training loads (frequency, intensity, and duration), historical injury records, and environmental conditions (temperature, humidity). Data preprocessing involved cleaning datasets to remove noise, addressing 5% of missing values through mean imputation, and normalizing data to a consistent range. We then employed a suite of AI techniques for predictive modeling. Random Forests and Gradient Boosting Machines (GBM) were used for their robustness in handling diverse data types and providing initial performance benchmarks. More complex models, including Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) with LSTM units, were applied for time-series data analysis to capture temporal patterns in training and injury data. We divided the data into training (60%), validation (20%), and test (20%) sets, and used 10-fold cross-validation to ensure model reliability. Our results demonstrated high efficacy in predicting injury risks. The Random Forest model achieved an accuracy of 0.89, while the GBM model attained an AUC-ROC score of 0.92, indicating excellent discriminatory power. CNNs and RNNs achieved mean squared errors of 0.015 and 0.012 on the validation set, respectively. In practical applications, including a partnership with the FC Barcelona Innovation Hub, our models contributed to a 20% reduction in injury rates and showed a correlation coefficient of 0.75 between predicted risks and actual injuries. Continuous learning mechanisms improved the models’ predictive accuracy by 5% over six months. Thus this study validated that AI-driven predictive modeling is a powerful tool for injury prevention in athletes. By integrating sophisticated AI techniques with comprehensive data analysis, we developed models that significantly enhanced injury prevention strategies. The practical applications and positive outcomes from the study underscore the transformative potential of AI in sports medicine, offering a new standard for creating safer and more effective athletic training environments.

The precise prediction of the remaining useful life (RUL) of lithium-ion batteries is of great significance for improving energy management efficiency and extending battery lifespan, and it is widely applied in the fields of new energy and electric vehicles. However, accurate RUL prediction still faces significant challenges. Although various methods based on deep learning have been proposed, the performance of their neural networks is strongly correlated with the hyperparameters. To overcome this limitation, this study proposes an innovative approach that combines the Alpha evolutionary (AE) algorithm with a deep learning model. Specifically, this hybrid deep learning architecture consists of convolutional neural network (CNN), time convolutional network (TCN), bidirectional long short-term memory (BiLSTM) and multi-scale attention mechanism, which extracts the spatial features, long-term temporal dependencies, and key degradation information of battery data, respectively. To optimize the model performance, the AE algorithm is introduced to automatically optimize the hyperparameters of the hybrid model, including the number and size of convolutional kernels in CNN, the dilation rate in TCN, the number of units in BiLSTM, and the parameters of the fusion layer in the attention mechanism. Experimental results demonstrate that our method significantly enhances prediction accuracy and model robustness compared to conventional deep learning techniques. This approach not only improves the accuracy and robustness of battery RUL prediction but also provides new ideas for solving the parameter tuning problem of neural networks.

An attribution graph-based interpretable method for CNNs

Delineating ground deformation over the Tengiz oil field, Kazakhstan, using the Intermittent SBAS (ISBAS) DInSAR algorithm

Advances in artificial intelligence (AI) techniques that can be used for different purposes have enabled it to be used in many different industrial applications. These are mainly used for modeling, identification, optimization, prediction and control of complex systems under the influence of more than one parameter in industrial applications. With the increasing accuracy of AI techniques, it has also obtained a wide application area on friction stir welding (FSW), one of the production methods developed in recent years. In this study, commonly used AI techniques for FSW, results, accuracy and superiority of AI techniques are reviewed and evaluated. In addition, an overview of AI techniques for FSW in different material combinations is provided. Considering the articles examined; It is seen that welding speed, rotational speed, the plunge depth, spindle torque, shoulder design, base material, pin design/profile, tool type are used as input parameters and tensile strength, yield strength, elongation, hardness, wear rate, welding quality, residual stress, fatigue strength are used as output parameters. As can be seen from the studies, it made important contributions in deciding what input parameters should be in order to have the output parameter at the desired value. The most common used materials for FSW are Al, Ti, Mg, Brass, Cu and so on. When FSW studies using artificial intelligence techniques were examined, it was seen that 81% of the most used materials were AL alloys and 23% of them were made with dissimilar materials. The most commonly utilized AI techniques were said to be artificial neural networks (ANN), fuzzy logic, machine learning, meta-heuristic methods and hybrid systems. As a result of the examination, ANN was the most widely used method among these methods. However, in recent years, with the exploration of new hybrid methods it was seen that hybrid systems used with ANN have higher accuracy.

Simple SummaryDistinguishing between benign vs. malignant bone lesions is often difficult on imaging. Many bone lesions are infrequent or rarely seen, and often only specialist radiologists have sufficient expertise to provide an accurate diagnosis. In addition, some benign bone tumours may exhibit potentially aggressive features that mimic malignant bone tumours, making the diagnosis even more difficult. The rapid development of artificial intelligence (AI) techniques has led to remarkable progress in image-recognition tasks, including the classification and characterization of various tumours. This study will review the most recent evidence for AI techniques in differentiating bone lesions on various imaging modalities using a systematic approach. Potential clinical applications of AI techniques for clinical diagnosis and management of bone lesions will also be discussed.An accurate diagnosis of bone tumours on imaging is crucial for appropriate and successful treatment. The advent of Artificial intelligence (AI) and machine learning methods to characterize and assess bone tumours on various imaging modalities may assist in the diagnostic workflow. The purpose of this review article is to summarise the most recent evidence for AI techniques using imaging for differentiating benign from malignant lesions, the characterization of various malignant bone lesions, and their potential clinical application. A systematic search through electronic databases (PubMed, MEDLINE, Web of Science, and clinicaltrials.gov) was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A total of 34 articles were retrieved from the databases and the key findings were compiled and summarised. A total of 34 articles reported the use of AI techniques to distinguish between benign vs. malignant bone lesions, of which 12 (35.3%) focused on radiographs, 12 (35.3%) on MRI, 5 (14.7%) on CT and 5 (14.7%) on PET/CT. The overall reported accuracy, sensitivity, and specificity of AI in distinguishing between benign vs. malignant bone lesions ranges from 0.44–0.99, 0.63–1.00, and 0.73–0.96, respectively, with AUCs of 0.73–0.96. In conclusion, the use of AI to discriminate bone lesions on imaging has achieved a relatively good performance in various imaging modalities, with high sensitivity, specificity, and accuracy for distinguishing between benign vs. malignant lesions in several cohort studies. However, further research is necessary to test the clinical performance of these algorithms before they can be facilitated and integrated into routine clinical practice.

Deep fakes, synthetic media generated through artificial intelligence (AI) techniques, have become a growing concern due to their potential for misinformation, manipulation, and security threats. As the sophistication of deep fake technology continues to evolve, traditional methods of detection are becoming less effective, thus prompting the need for advanced AI-driven solutions. This article explores the various AI techniques employed in deep fake detection, such as deep learning, neural networks, and computer vision algorithms. A review of the latest advancements in AI technologies, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Generative Adversarial Networks (GANs), is presented to highlight their roles in identifying deep fakes. Additionally, we explore challenges related to data availability, model robustness, and ethical implications. The paper concludes by proposing a hybrid detection framework that integrates multiple AI techniques to enhance the accuracy and efficiency of deep fake identification.

Historically, the rotorcraft community has experienced a higher fatal accident rate than other aviation segments, including commercial and general aviation. To date, traditional methods applied to reduce incident rates have not proven hugely successful for the rotorcraft community. Recent advancements in artificial intelligence (AI) and the application of these technologies in different areas of our lives are both intriguing and encouraging. When developed appropriately for the aviation domain, AI techniques may provide an opportunity to help design systems that can address rotorcraft safety challenges. Our recent work demonstrated that AI algorithms could use video data from onboard cameras and correctly identify different flight parameters from cockpit gauges, e.g., indicated airspeed. These AI-based techniques provide a potentially cost-effective solution, especially for small helicopter operators, to record the flight state information and perform post-flight analyses. We also showed that carefully designed and trained AI systems can accurately predict rotorcraft attitude (i.e., pitch and yaw) from outside scenes (images or video data). Ordinary offthe-shelf video cameras were installed inside the rotorcraft cockpit to record the outside scene, including the horizon. The AI algorithm was able to correctly identify rotorcraft attitude at an accuracy in the range of 80%. In this work, we combined five different onboard camera viewpoints to improve attitude prediction accuracy to 94%. Our current approach, which is referred to as ensembled prediction, significantly increased the reliability in the predicted attitude (i.e., pitch and yaw). For example, in some camera views, the horizon may be obstructed or not visible. The proposed ensemble method can combine visual details recorded from other cameras and predict the attitude with high reliability. In our setup, the five onboard camera views included pilot windshield, co-pilot windshield, pilot Electronic Flight Instrument System (EFIS) display, co-pilot EFIS display, and the attitude indicator gauge. Using video data from each camera view, we trained a variety of convolutional neural networks (CNNs), which achieved prediction accuracy in the range of 79% to 90%. We subsequently ensembled the learned knowledge from all CNNs and achieved an ensembled accuracy of 93.3%. Our efforts could potentially provide a cost-effective means to supplement traditional Flight Data Recorders (FDR), a technology that to date has been challenging to incorporate into the fleets of most rotorcraft operators due to cost and resource constraints. Such cost-effective solutions can gradually increase the rotorcraft community's participation in various safety programs, enhancing safety and opening up helicopter flight data monitoring (HFDM) to historically underrepresented segments of the vertical flight community.

Satellite interferometry for high-precision detection of ground deformation at a carbon dioxide storage site

Deep neural networks have been used in various fields, but their internal behavior in how they understand images is not well known. In this study, we discuss two counterintuitive properties of convolutional neural networks (CNNs). First, we evaluated the size of the receptive field of CNNs with their classification accuracy. Previous studies have attempted to increase the size of the receptive field for performance gain. However, we observed that some CNNs with a smaller receptive field can achieve higher classification accuracy. In this regard, we claim that a larger receptive field does not guarantee improved classification accuracy. Second, using the effective receptive field, we examined the contribution of each pixel to the output of CNN. Intuitively, each pixel is expected to equally contribute to the final output, but we found that there exist pixels in a partially dead state with little contribution to the output. We reveal that the reason for dead pixels lies in even stride operations with odd-sized kernels in CNN and propose a kernel padding method to remove the dead pixels. We demonstrated the vulnerability of CNNs with dead pixels when we detect a noise or small box that is on dead pixels. Our findings on dead pixels should be understood and considered in practical applications of CNN.

In recent years, the classification of land change has revolutionized the ability to monitor and understand dynamic changes occurring on the Earth’s surface. Artificial intelligence (AI) techniques must improve the performance and accuracy of land change detection by extracting spectral features from several Convolutional Neural Networks (CNNs) and integrating them. In this study, AI techniques were applied to classify the land use and land cover (LULC) of the Najran city map in Saudi Arabia based on 2020 Landsat 8 satellite imagery. This was achieved using several hybrid models combining CNN and random forest (RF) models, namely AlexNet-RF and GoogLeNet-RF, as well as the combined spectral features of AlexNet-GoogLeNet with RF. The results showed that LULC classification using a hybrid system was superior to CNN and proved that the proposed hybrid system of combined spectral features extracted from AlexNet-GoogLeNet with RF provided better results than using the hybrid system proposed by AlexNet with RF and GoogLeNet with RF. The proposed hybrid system of combined spectral features extracted from AlexNet-GoogLeNet with RF achieved an accuracy of 96.95%, a Kappa coefficient of 0.9638, sensitivity of 96.95%, AUC of 98.4%, and specificity of 99.83%. The proposed hybrid methods aim to enhance the classification accuracy and increase the robustness of the system, ensuring consistent performance across diverse earth-change scenarios. It substantially impacts various domains, including environmental monitoring, disaster management, and sustainable urban planning.

In Japan, precipitation has recently become more localized, concentrated, and intense. Extreme precipitation has caused a rapid land deformation, which results in ground water levels and loss of building stability. For the appropriate maintenance and renewal of infrastructure, including the design method of underground structures, the utilization of remote sensing technology which can always observe the ground deformation over a wide range with high accuracy is expected. This research presents the results of long-term continuous land deformation (subsidence or uplift) monitoring and using Global Navigation Satellite System (GNSS) and Consecutive Differential Interferometric Synthetic Aperture Radar (DInSAR) technique in Yokohama, Yokosuka and Miura cities, Kanagawa Prefecture, Japan. We adopted the Consecutive DInSAR method with the small spatial and short time baseline between satellites. The satellite data used thirty-three pairs of Sentinel-1 data in descending orbit from December 27, 2017 to November 17, 2020, and twenty-two Sentinel-1 data pairs in ascending orbit from June 26, 2019 to October 30, 2020. The validation of the SAR images used the daily coordinate values of the GNSS control point provided by the Geographical Survey Institute (GSI). As a result of the analysis, the estimated land displacement and GNSS data agreed well in Yokohama, Yokosuka, and Miura cities. The RMSE in descending orbit of Yokohama, Yokosuka and Miura cities were found to be 0.41 cm, 0.49 cm, and 0.49 cm, respectively and the RMSE in ascending orbit of Yokohama, Yokosuka and Miura cities were found to be 0.68 cm, 0.53 cm, and 0.58 cm, respectively, demonstrating its capability to analyze the land displacement monitoring over under different topographical conditions. In the future, this research result is expected to applicable in the field of disaster monitoring and environmental change observation which continuously require high accuracy observation over a wide area.

An earthquake of magnitude 6.4 occurred in Albania on November 26, 2019 at 03:54:12 (local time). The epicenterof the earthquake was about20 km from the coastal city of Durres. The purpose of this study is toestimate the ground deformation associated with thisearthquakeby Differential Interferometric Synthetic Aperture Radar (DInSAR) technique. A differential interferogram was formed using the two Sentinel-1 Single Look Complex (SLC) images. The flattened and filtered differential interferogram was unwrapped and converted to ground displacement. The results indicatethat in the area close tothe epicentrethe ground is uplifted at a maximum of 8 cm along the radar line-of-sight (LOS). Progressively smaller ground uplift isdetected in an area of about 10-15 km around the major uplift zone. The earthquake could be related with under thrustingof Adria microplate or fault processes within the under thrustedAdria microplate itself. Further monitoring by satellite imagery is needed to investigate the post-seismic ground deformation in the area.

This paper presents application of satellite interferometric methods (persistent scatterer interferometric synthetic aperture radar (PSInSAR™) and differential interferometric synthetic aperture radar (DInSAR)) for observation of ground deformation in the Upper Silesian Coal Basin (USCB) in Southern Poland. The presented results were obtained during the DORIS project (EC FP 7, Grant Agreement n. 242212, www.doris-project.eu). Several InSAR datasets for this area were analysed. Most of them were processed by Tele-Rilevamento Europa - T.R.E. s.r.l. Italy. Datasets came from different SAR satellites (ERS 1 and 2, Envisat, ALOS- PALSAR and TerraSAR-X) and cover three different SAR bands (L, C and X). They were processed using both InSAR techniques: DInSAR, where deformations are presented as interferometric fringes on the raster image, and PSInSAR, where motion is indentified on irregular set of persistent scatterer (PS) points. Archival data from the C-band European Space Agency satellites ERS and ENVISAT provided information about ground movement since 1992 until 2010 in two separate datasets (1992–2000 and 2003–2010). Two coal mines were selected as examples of ground motion within inactive mining areas: Sosnowiec and Saturn, where mining ceased in 1995 and 1997, respectively. Despite well pumping after closure of the mines, groundwater rose several dozen meters, returning to its natural horizon. Small surface uplift clearly indicated on satellite interferometric data is related to high permeability of the hydrogeological subregion and insufficient water withdrawal from abandoned mines. The older 1992–2000 PSInSAR dataset indicates values of ground motion ranging from –40.0 to 0.0 mm. The newer 2003–2010 dataset shows values ranging from –2.0 to +7.0 mm. This means that during this period of time subsidence was less and uplift greater in comparison to the older dataset. This is even more evident in the time series of randomly selected PS points from both coal mines. The presence of bentonite deposits in the Saturn coal mine can also have influence on the ground surface uplift. Analysis of interferometric L and X-band data in Upper Silesia has enabled observation and monitoring of the underground mining front for several months. It was indicated by the example from the Halemba–Wirek coal mine. Analysis of the TerraSAR–X dataset processed by SqueeSAR algorithms proved to be the most effective for this purpose. X-band persistent scatterer interferometry (PSI) time series can help to indentify small seemingly negligible movements and are successfully supplemented by fringes when displacement becomes significant. Differential interferograms from the L-band dataset detect similar displacement values but, thanks to longer wavelength, are characterized by better coherence, especially in the middle of the subsidence trough. Results on ground deformation proved that ground motion above abandoned mines continues long after their closure. Therefore, existing regulations stating that abandoned mines are considered fully safe five years after mine closure should change. Moreover, it should be emphasised that construction in these areas should be avoided due to existing potential risks.

The formation of exit burrs during the drilling of ductile metals such as aluminum is critical in precision manufacturing and manufacturing automation. Because drilling burrs are difficult to remove, methods to predict various burr types and/or implement burr minimization schemes that consider the various attributes of burrs must be devised. In this study, not only drilling process conditions, including feed, cutting speed, and drill diameter, but also an artificial neural network is implemented to predict the formation of burrs during the drilling of aluminum-7075, which is widely used in the aerospace and automobile industries. Based on the drilling conditions, the main exit burr characteristics, such as burr size and type, were classified experimentally. Three different types of exit burrs (uniform burrs, uniform burrs with caps, and transient burrs) were observed from the aluminum 7075 workpieces. The classification results were further analyzed using burr control charts and an empirical equation, which enables the understanding of the overall influence of the drilling conditions over the burr types. Moreover, acoustic emission (AE) sensor monitoring scheme was utilized to sample the sensitive time-series signals during drilling burr formation. Subsequently, burr types were predicted using artificial intelligence techniques, namely machining learning and deep learning. First, backpropagation (BP) neural networks were constructed using the drilling conditions and AE signals as input vectors. For a comparative prediction, a convolution neural network (CNN) was implemented to obtain spectrogram image inputs from the sampled AE data. The proposed scheme is useful in predicting drilling burr types by employing a sensitive sensor monitoring setup and advanced artificial intelligence techniques, where both prediction results are well matched with experimental results. In addition, the CNN model shows effectiveness for a commonly practiced manufacturing process as it predicts the burr types with better accuracy than the BP network model (0.9375 over 0.8571).