The outbreak of Pine Shoot Beetle (PSB, Tomicus spp.) posed a significant threat to the health of Yunnan pine forests, necessitating the development of an efficient and accurate remote sensing monitoring method. The integration of unmanned aerial vehicle (UAV) imagery and deep learning algorithms shows great potential for monitoring forest-damaged trees. Previous studies have utilized various deep learning semantic segmentation models for identifying damaged trees in forested areas; however, these approaches were constrained by limited accuracy and misclassification issues, particularly in complex forest backgrounds. This study evaluated the performance of five semantic segmentation models in identifying PSB-damaged trees (UNet, UNet++, PAN, DeepLabV3+ and FPN). Experimental results showed that the FPN model outperformed the others in terms of segmentation precision (0.8341), F1 score (0.8352), IoU (0.7239), mIoU (0.7185) and validation accuracy (0.9687). Under the pure Yunnan pine background, the FPN model demonstrated the best segmentation performance, followed by mixed grassland-Yunnan pine backgrounds. Its performance was the poorest in mixed bare soil-Yunnan pine background. Notably, even under this challenging background, FPN still effectively identified diseased trees, with only a 1.7% reduction in precision compared to the pure Yunnan pine background (0.9892). The proposed method in this study contributed to the rapid and accurate monitoring of PSB-damaged trees, providing valuable technical support for the prevention and management of PSB.
This research proposes an artificial intelligence (AI)-powered digital twin framework for highway slope stability risk monitoring and prediction. For highway slope stability, a digital twin replicates the geological and structural conditions of highway slopes while continuously integrating real-time monitoring data to refine and enhance slope modeling. The framework employs instance segmentation and a random forest model to identify embankments and slopes with high landslide susceptibility scores. Additionally, artificial neural network (ANN) models are trained on historical drilling data to predict 3D subsurface soil type point clouds and groundwater depth maps. The USCS soil classification-based machine learning model achieved an accuracy score of 0.8, calculated by dividing the number of correct soil class predictions by the total number of predictions. The groundwater depth regression model achieved an RMSE of 2.32. These predicted values are integrated as input parameters for seepage and slope stability analyses, ultimately calculating the factor of safety (FoS) under predicted rainfall infiltration scenarios. The proposed methodology automates the identification of embankments and slopes using sub-meter resolution Light Detection and Ranging (LiDAR)-derived digital elevation models (DEMs) and generates critical soil properties and pore water pressure data for slope stability analysis. This enables the provision of early warnings for potential slope failures, facilitating timely interventions and risk mitigation.
Occlusions of granular particles in images significantly affect the accuracy of evaluating particle morphology for granular materials. In this study, a novel framework of SOLO-PCNet is proposed, which can automatically segment all the particles and predict the complete contours of the occluded particles in densely packed materials. Firstly, the instance segmentation model SOLOv2 is trained for the prediction of all the detectable particles. Then a self-supervised learning algorithm PCNetM is introduced for the inference of the complete contours of the occluded particles so that the prediction of SOLOv2 can be directly input to PCNet-M for the subsequent completion. Thereafter, the particle morphology characteristics including elongation, equivalent mean size, convexity, and circularity are automatically calculated. Then, the evaluation metrics of the segmentation model and morphology characteristics are validated, and the results exhibit the strong generalization ability of the segmentation and completion tasks. Finally, the uncertainty of the completed contours with morphology properties is explored for reliable analysis. This study successfully acquires the complete contours for each particle and provides the foundation for evaluating the mechanical properties of the packed granular materials from individual particles.
In this study, inexpensive dual precursors namely rice husk ash (RHA) and quarry dust (QD) were mixed with an alkali solution (NaOH solution) in different mix proportions. Next, the effect of the dual precursors-alkali solutions (geopolymers) on the strength parameters of the different soil-geopolymer mixes subjected to various curing days were assessed. The result obtained from the assessment showed that the dual precursors-alkali solutions improved the strength properties of all the different soil-geopolymer mixes. The optimum soil-geopolymer mix proportion that resulted in the best improvement of the expansive subgrade soil properties was observed at a combination of 15%RHA+15%QD+3M NaOH. Furthermore, micro-pore analyses (porosity and pore size distribution), which were implemented with scanning electron micrographs and an image segmentation technique, local adaptive thresholding algorithm, showed progressive reduction in the porosity of the natural expansive subgrade soil and that of the optimal soil-geopolymer mix proportion cured for 7 and 28 days.
Forests are essential to our planet's well-being, playing a vital role in climate regulation, biodiversity preservation, and soil protection, thus serving as a cornerstone of our global ecosystem. The threat posed by forest fires highlights the critical need for early detection systems, which are indispensable tools in safeguarding ecosystems, livelihoods, and communities from devastating destruction. In combating forest fires, a range of techniques is employed for efficient early detection. Notably, the combination of drones with artificial intelligence, particularly deep learning, holds significant promise in this regard. Image segmentation emerges as a versatile method, involving the partitioning of images into multiple segments to simplify representation, and it leverages deep learning for fire detection, continuous monitoring of high-risk areas, and precise damage assessment. This study provides a comprehensive examination of recent advancements in semantic segmentation based on deep learning, with a specific focus on Mask R-CNN (Mask Region Convolutional Neural Network) and YOLO (You Only Look Once) v5, v7, and v8 variants. The emphasis is placed on their relevance in forest fire monitoring, utilizing drones equipped with high-resolution cameras.