Concrete is extensively used in the construction of infrastructure such as houses and bridges. However, the appearance of cracks in concrete structures over time can diminish their sealing and load-bearing capability, potentially leading to structural failures and disasters. The timely detection of cracks allows for repairs without the need to replace the entire structure, resulting in cost savings. Currently, manual inspection remains the predominant method for identifying concrete cracks. However, in today’s increasingly complex construction environments, subjective errors may arise due to human vision and perception. The purpose of this work is to investigate and design an autonomous convolutional neural network-based concrete detection system that can identify cracks automatically and use that information to calculate the crack proportion. The experiment’s findings show that the trained model can classify concrete cracks with an accuracy of 99.9%. Moreover, the clustering technique applied to crack images enables the clear identification of the percentage of cracks, which facilitates the development of concrete damage level detection over time.

Accurate forecasting of wind speed and direction is critical for the efficient integration of wind power into energy systems, ensuring reliable renewable energy production and grid stability. Traditional methods often struggle with capturing nonlinear interdependencies, quantifying uncertainties, and providing reliable long-term predictions, particularly in complex atmospheric conditions. To address these challenges, this study introduces multi-model Integration for dynamic forecasting (MIDF), an ensemble machine learning framework that combines the strengths of DeepAR and temporal fusion transformer (TFT) models through a two-step meta-learning process. MIDF leverages DeepAR’s probabilistic forecasting capabilities and TFT’s attention mechanisms to enhance accuracy, robustness, and interpretability. Using a custom meteorological dataset spanning January 2010 to May 2023, the model was evaluated against standalone alternatives across multiple metrics, including MSE, RMSE, and R². MIDF achieved superior performance, with MSE, RMSE, and R² values of 0.0035, 0.01913, and 0.89 for wind speed, and 0.00052, 0.02507, and 0.86 for wind direction, significantly reducing errors compared to existing methods. These results underscore the potential of ensemble learning in advancing wind forecasting accuracy, enabling more reliable renewable energy management, operational planning, and risk mitigation in meteorological applications.

This study presents an acoustic-based multitask technique for distinguishing construction equipment types and their associated activities using a customized ResNet-18 deep learning (DL) model. Construction equipment activities on sites require efficient recognition during real-time operations to ensure maximum productivity and safety. While other studies examine typical recognition, particularly visual or sensor-based methods, this study investigates the potential of equipment-generated audio signals as a feasible option for activity classification. Leveraging the acoustic signal and advanced DL techniques, we collected and preprocessed the audio signals from construction equipment and trained them on the customized pretrained ResNet-18 model to accurately recognize equipment and its corresponding activities. The proposed method shows superior accuracy and good performance in equipment and activity identification. Our findings demonstrate the capability and efficiency of the developed approach in accurately classifying equipment and activity types, achieving classification accuracies of up to 97%. Compared to other models, our method is more computationally efficient. The results suggest that the proposed method holds promise for real-time recognition of heavy equipment and associated activities. Its capability can reduce construction risks, enhance productivity, and improve decision-making at construction sites.

Automated recognition of celestial bodies from observational imagery is a cornerstone of autonomous space exploration. However, deploying deep learning models in space environments entails rigorous requirements not only for accuracy but also for reliability (calibration) and safety (anomaly rejection). Traditional Convolutional Neural Networks (CNNs) trained on small-scale astronomical datasets often suffer from overfitting and overconfidence on Out-of-Distribution (OOD) artifacts. In this work, we present a robust classification framework based on DINOv2, a Vision Transformer pre-trained via discriminative self-supervised learning. We curate a high-fidelity dataset of seven planetary classes sourced from NASA archives and propose a two-stage domain adaptation strategy to transfer large-scale foundation model features to this fine-grained task. Extensive experiments show that our method reaches 100% Top-1 accuracy on the canonical split, and remains highly stable under split variation, achieving 99.43% ± 0.85% Top-1 accuracy across R = 5 repeated stratified splits. More importantly, we address the critical issue of model trustworthiness. Through post hoc temperature scaling, our model achieves a state-of-the-art Expected Calibration Error (ECE) of 0.08%, representing a 36-fold improvement over ResNet50 (2.90%) and a 4.5-fold improvement over the EfficientNet-B3 baseline (0.36%). Furthermore, by integrating Energy-based OOD detection, the system effectively rejects non-planetary artifacts with an AUROC of 93.7%. Qualitative analysis using Grad-CAM reveals that self-supervised attention mechanisms naturally focus on intrinsic planetary features (e.g., surface textures and rings) while ignoring background noise, confirming the superior robustness of vision foundation models in astronomical vision tasks.

Accurate prediction of wind direction is a critical requirement for coastal safety management, renewable energy optimization, and weather-driven risk mitigation, particularly in highly dynamic atmospheric environments where statistical and deep learning models often struggle to capture nonlinear interactions and temporal dependencies. Existing approaches typically rely on raw or weakly processed meteorological inputs and treat directional information implicitly, which limits their ability to exploit the underlying physical structure of wind evolution. To address these challenges, this research designs a novel Physics Vector Driven (PVD) data pre-processing framework that explicitly encodes physically meaningful gradients and directional dynamics from multivariate meteorological observations, transforming raw measurements into sequence-aware vector representations suitable for deep time-series learning. Building on this foundation, a novel Directional Temporal Gradient Vector Network (DTGVectorNet) is proposed, which fuses a Directional Gradient Attention ResNet (DGResNet 1D CNN) for spatial-directional feature extraction with a Temporal Gradient LSTM (TGLSTM) designed to model the temporal evolution of wind vectors. The tight integration of Directional Gradient Attention (DGA) and Temporal Gradient (TG) memory enables the network to jointly learn instantaneous directional cues and their temporal propagation, significantly enhancing predictive fidelity. An experimental evaluation of the Busan wind datasets demonstrates that the proposed DTGVectorNet achieves a wind direction prediction accuracy of 99.12%, substantially outperforming conventional state-of-the-art baselines. These results confirm that physics-aware vector preprocessing combined with directional-temporal gradient fusion provides a powerful and generalizable paradigm for high-precision wind direction forecasting. To ensure reproducibility and facilitate further research, the complete dataset and implementation details of DTGVectorNet are publicly available through an open-access repository, Zenodo.

Urban green-space monitoring in dense cityscapes remains limited by accuracy–efficiency trade-offs and the absence of integrated, auditable area estimation. We introduce GreenViT, a Vision Transformer (ViT) based framework for precise segmentation and transparent quantification of urban green space. GreenViT couples a ViT-L/14 backbone with a lightweight single-path, progressive upsampling decoder (Green Head), preserving global context while recovering thin structures. Experiments were conducted on a manually annotated dataset of 20 high-resolution satellite images collected from Satellites.Pro, covering five land-cover classes (background, green space, building, road, and water). Using a 224 × 224 sliding window sampling scheme, the 20 images yield 62,650 training/validation patches. Under five-fold evaluation, it attains 0.9200 ± 0.0243 mIoU, 0.9580 ± 0.0135 Dice, and 0.9570 PA, and the calibrated estimator achieves 1.10% relative area error. Overall, GreenViT strikes a strong balance between accuracy and efficiency, making it particularly well-suited for thin or boundary-rich classes. It can be used to support planning evaluations, green-space statistics, urban renewal assessments, and ecological red-line verification, while providing reliable green-area metrics to support urban heat mitigation and pollution control efforts. This makes it highly suitable for decision-oriented long-term monitoring and management assessments.