Search Results (406)

Lychee detection and maturity classification are crucial for yield estimation and harvesting. In densely packed lychee clusters with limited training samples, accurately determining ripeness is challenging. This paper proposes a new transformer model incorporating a Kolmogorov–Arnold Network (KAN), termed GhostResNet (GRN)–KAN–Transformer, for lychee detection and ripeness classification in dense on-tree fruit clusters. First, within the backbone, we introduce a stackable multi-layer GhostResNet module to reduce redundancy in feature extraction and improve efficiency. Next, during feature fusion, we add a large-scale layer to enhance sensitivity to small objects and to increase polling of the small-scale feature map during querying. We further propose a multi-layer cross-fusion attention (MCFA) module to achieve deeper hierarchical feature integration. Finally, in the decoding stage, we employ an improved KAN for the classification and localization heads to strengthen nonlinear mapping, enabling a better fitting to the complex distributions of categories and positions. Experiments on a public dataset demonstrate the effectiveness of GRN-KANformer. Compared with the baseline, GFLOPs and parameters of the model are reduced by 8.84% and 11.24%, respectively, while mean Average Precision (mAP) metrics mAP50 and mAP50–95 reach 94.7% and 58.4%, respectively. Thus, it lowers computational complexity while maintaining high accuracy. Comparative results against popular deep learning models, including YOLOv8n, YOLOv12n, CenterNet, and EfficientNet, further validate the superior performance of GRN-KANformer. Full article

PM_2.5 pollution is a significant environmental problem in global urbanization. However, traditional macro-scale studies are constrained by data resolution limitations, failing to accurately characterize the microscale built environment or thoroughly investigate its spatially heterogeneous effects on PM_2.5 concentrations. To address this gap, this study constructs a multidisciplinary framework of “Street View Imagery element extraction–spatial heterogeneity modeling–planning strategy optimization” with Xi’an as the case. Leveraging machine learning techniques, the study employs the ResNet50 deep learning model and the ADE20K dataset to precisely extract ten microscale built environment factors from tens of thousands of street view images. Combined with the High-resolution and High-quality Ground-level PM_2.5 Dataset for China, Ordinary Least Squares (OLS), Geographically Weighted Regression (GWR), and Multiscale Geographically Weighted Regression (MGWR) models were used to systematically reveal the impacts of the microscale built environment on PM_2.5 concentrations. Ten built environment factors were identified with varying spatial heterogeneity in their effects on the PM_2.5 concentrations, as follows: (1) factors with positive effects, in descending order of strength, include building, wall, fence, tree, sky, and grass; (2) factors with negative effects, in descending order of strength, include sidewalk, plant, and car; (3) compared with other factors, the road factor showed a relatively weaker effect. This research provides decision-making support for targeted urban planning and environmental protection, while offering valuable references for air pollution control in other cities. Full article

The integration of unmanned aerial vehicle-based laser scanning (ULS) and mobile laser scanning (MLS) enables the detection of forest three-dimensional structure in high-density canopy areas and has become an important tool for monitoring and managing forest ecosystems. However, MLS faces difficulties in positioning due to canopy occlusion, making integration challenging. Due to the variations in observation platforms, ULS and MLS point clouds exhibit significant structural discrepancies and limited overlapping areas, necessitating effective methods for feature extraction and correspondence establishment between these features to achieve high-precision registration and integration. Therefore, we propose a registration algorithm that introduces a Feature Deviation Metric to enable feature extraction and correspondence construction for forest point clouds in complex regional environments. The algorithm first extracts surface point clouds using the hidden point algorithm. Then, it applies the proposed dual-threshold method to cluster individual tree features in ULS, using cylindrical detection to construct a Feature Deviation Metric from the feature points and surface point clouds. Finally, an optimization algorithm is employed to match the optimal Feature Deviation Metric for registration. Experiments were conducted in 8 stratified mixed tropical rainforest plots with complex mixed-species canopies in Malaysia and 6 structurally simple, high-canopy-density pure forest plots in anorthern China. Our algorithm achieved an average RMSE of 0.17 m in eight tropical rainforest plots with an average canopy density of 0.93, and an RMSE of 0.05 m in six northern forest plots in China with an average canopy density of 0.75, demonstrating high registration capability. Additionally, we also conducted comparative and adaptability analyses, and the results indicate that the proposed model exhibits high accuracy, efficiency, and stability in high-canopy-density forest areas. Moreover, it shows promise for high-precision ULS-MLS registration in a wider range of forest types in the future. Full article

This study investigated the distribution of pine wood nematodes (PWNs, Bursaphelenchus xylophilus) and their co-occurrence with B. mucronatus in recently dead pine trees across coastal and inland regions while monitoring the seasonal emergence patterns of Monochamus alternatus from 2021 to 2023. Nematodes were extracted from felled trees and beetle bodies using the Baermann funnel method. Aggregation pheromone traps were used to monitor vector activity and to assess temperature-dependent emergence. The results showed a negative correlation between PWN and B. mucronatus density (r = −0.73, p < 0.01), which prompted tests on interspecific interactions. M. alternatus emergence was positively associated with average temperature (r = 0.74–0.78), supporting the temperature-informed surveillance timing in this dataset. These findings highlight the role of climate-driven dynamics in shaping vector behavior and nematode population structures. This study supports the development of sustainable temperature-responsive management strategies for controlling pine wilt disease. These strategies provide a foundation for climate-resilient forest health and long-term ecosystem sustainability. Full article

The selection of tree species for afforestation in Mediterranean environments involves challenges related to adaptability, impact on soil properties, and overall environmental quality. Cupressus lusitanica has been recognized for its rapid growth, environmental resilience, and versatile applications, positioning it as a promising candidate for these regions. Although it has been used for afforestation in Northeast Portugal since the 1990s, no comprehensive studies have evaluated its performance under local conditions. To address this knowledge gap, this study assessed a 14-year-old C. lusitanica stand in Northeast Portugal. The wood’s anatomical, physical, chemical, and mechanical properties, as well as biomass production, were evaluated. The species showed superior radial growth and adaptability compared with other species under similar environmental conditions. Despite exhibiting lower fiber length (1.6 mm) and basic wood density (404 kg/m³), shrinkage values fell within the typical range for softwoods. Nevertheless, a marked tendency for warping was observed. The extractive content was relatively high (5.1%), with the ethanol-soluble fraction being predominant (3.6%). Mechanical tests revealed low values for both Modulus of Elasticity (MOE) (3592.5–3617.1 MPa) and Modulus of Rupture (MOR) (57.7–68.9 MPa), with both properties significantly influenced by knot presence. Given the results obtained, the species C. lusitanica, despite its low wood density and potential limitations in use, exhibits remarkable growth and adaptability, which confer a high potential for biomass production and carbon sequestration, as well as potential applications of its wood in reconstituted panels and fiber- or particle-based boards. Full article

The manuscript explores the valorization of forest logging residues, collected during forest management operations between summer 2023 and spring 2025 in mixed deciduous and coniferous forests, as a raw material for producing valuable bioactive products. These products offer a sustainable alternative to synthetic pesticides and fertilizers. Seven batches of biomass, comprising understory trees and branches from deciduous (mainly aspen, birch, and grey alder) and coniferous (mainly Scots pine) species, were collected during different seasons, crushed, and extracted using an ethanol–water solution. The yield of hydrophilic extracts containing proanthocyanidins (PACs) ranged from 18 to 25% per dry biomass. The highest PACs concentration (42% of extract dry mass) was found in small branches with a high bark content. The extracts and PACs at concentrations of 6.25–12.50 mg mL⁻¹ showed fungicidal activity against several pathogenic fungi, including Botrytis cinerea Pers., Mycosphaerella sp. Johanson, Heterobasidion annosum (Fr.) Bref., and Heterobasidion parviporum Niemelä & Korhonen. Residual biomass after extraction, enriched with sea buckthorn berry pomace and a siliceous complex, was characterized and evaluated for its impact on the growth of Scots pine seedlings and selected agricultural crops. Results from forest and agricultural field trials in 2023–2025 confirmed a positive effect of the fertilizer on crop yield and quality at a low application rate (40 kg ha⁻¹ per crop). Fertilizer increased the yield of radish, dill, potatoes, and wheat by up to 44% (highest for potatoes and dill) compared to the reference, confirming its agronomic value. Full article

Pistacia terebinthus L. which has been traditionally used in diet and medicine, remains underexplored in Greece, particularly regarding its chemical composition and antioxidant activity. The current study aims to comparatively evaluate the chemical profile of cold-pressed terebinth fruit oils, obtained from wild trees growing in the Greek Island of Chios (North East Aegean Sea), harvested during three years (2019, 2020 and 2021). The oils’ lipid profile was dominated by oleic acid (C18:1 cis-9) (42–45%) followed by palmitic acid (C16:0) (24–30%) and linoleic acid (C18:2 cis-9,12) (19–22%). Their phenolic acid content, expressed as anacardic acids—known for their bioactive properties—was quantified via q-¹H-NMR and found to be markedly high (1.91–2.98 mmol/kg oil). Total phenolic content (TPC) of the fruit extract showed interesting high value (185.92 ± 2.61 mg GAE/g) accompanied by strong antioxidant activity (DPPH, exhibiting > 80% inhibition at a concentration of 100 µg/mL) which was positively correlated with TPC. Additionally, the fruits demonstrated a rich nutritional profile, particularly in crude fibers (38.9%) and essential minerals (K, Mg, and Zn), along with low sodium content, suggesting notable dietary benefits. The cold-pressed oil exhibited high lipid content and low specific extinction coefficients (K₂₃₂, K₂₇₀), indicating minimal oxidation and confirming the oil’s freshness. These findings highlight the potential of P. terebinthus fruit oil as a high-value functional raw material with nutritional and antioxidant properties. Comparable to olive oil in lipid quality, Greek turpentine fruit and oil could play a promising role towards further applications in the food, cosmetic and pharmaceutical sectors. Full article

This paper addresses the problem of recognising and classifying hand movements to control an upper-limb exoskeleton. To solve this problem, a multisensory system based on the fusion of data from electromyography (EMG) sensors, inertial measurement units (IMUs), and virtual reality (VR) trackers is proposed, which provides highly accurate detection of users’ movements. Signal preprocessing (noise filtering, segmentation, normalisation) and feature extraction were performed to generate input data for regression and classification models. Various machine learning algorithms are used to recognise motor activity, ranging from classical algorithms (logistic regression, k-nearest neighbors, decision trees) and ensemble methods (random forest, AdaBoost, eXtreme Gradient Boosting, stacking, voting) to deep neural networks, including convolutional neural networks (CNNs), gated recurrent units (GRUs), and transformers. The algorithm for integrating machine learning models into the exoskeleton control system is considered. In experiments aimed at abandoning proprietary tracking systems (VR trackers), absolute position regression was performed using data from IMU sensors with 14 regression algorithms: The random forest ensemble provided the best accuracy (mean absolute error = 0.0022 metres). The task of classifying activity categories out of nine types is considered below. Ablation analysis showed that IMU and VR trackers produce a sufficient informative minimum, while adding EMG also introduces noise, which degrades the performance of simpler models but is successfully compensated for by deep networks. In the classification task using all signals, the maximum result (99.2%) was obtained on Transformer; the fully connected neural network generated slightly worse results (98.4%). When using only IMU data, fully connected neural network, Transformer, and CNN–GRU networks provide 100% accuracy. Experimental results confirm the effectiveness of the proposed architectures for motor activity classification, as well as the use of a multi-sensor approach that allows one to compensate for the limitations of individual types of sensors. The obtained results make it possible to continue research in this direction towards the creation of control systems for upper exoskeletons, including those used in rehabilitation and virtual simulation systems. Full article

Evaluating the restoration quality of university outdoor spaces is often constrained by subjective surveys and manual assessment, limiting scalability and objectivity. This study addresses this gap by applying explainable machine learning to predict restorative quality from campus imagery, enabling large-scale, data-driven evaluation and capturing complex nonlinear relationships that traditional methods may overlook. Using Fujian Agriculture and Forestry University as a case study, this study extracted road network data, generated 297 coordinates at 50-m intervals, and collected 1197 images. Surveys were conducted to obtain restorative quality scores. The Mask2Former model was used to extract landscape features, and decision tree algorithms (RF, XGBoost, GBR) were selected based on MAE, MSE, and EVS metrics. The combination of optimal algorithms and SHAP was employed to predict restoration quality and identify key features. This research also used a multivariate linear regression model to identify features with significant statistical impact but lower features importance ranking. Finally, the study also analyzed heterogeneity in scores for three restoration indicators and five campus zones using k-means clustering. Empirical results show that natural elements like vegetation and water positively affect psychological perception, while structural components like walls and fences have negative or nonlinear effects. On this basis, this study proposes spatial optimization strategies for different campus areas, offering a foundation for creating high-quality outdoor environments with restorative and social functions. Full article

To solve the low efficiency of traditional sheet metal measurement, this paper proposes a digital inspection method for sheet metal parts based on 3D point clouds. The 3D point cloud data of sheet metal parts are collected using a 3D laser scanner, and the topological relationship is established by using a K-dimensional tree (KD tree). The pass-through filtering method is adopted to denoise the point cloud data. To preserve the fine features of the parts, an improved voxel grid method is proposed for the downsampling of the point cloud data. Feature points are extracted via the intrinsic shape signatures (ISS) algorithm and described using the fast point feature histograms (FPFH) algorithm. After rough registration with the sample consensus initial alignment (SAC-IA) algorithm, an initial position is provided for fine registration. The improved iterative closest point (ICP) algorithm, used for fine registration, can enhance the registration accuracy and efficiency. The greedy projection triangulation algorithm optimized by moving least squares (MLS) smoothing ensures surface smoothness and geometric accuracy. The reconstructed 3D model is projected onto a 2D plane, and the actual dimensions of the parts are calculated based on the pixel values of the sheet metal parts and the conversion scale. Experimental results show that the measurement error of this inspection system for three sheet metal workpieces ranges from 0.1416 mm to 0.2684 mm, meeting the accuracy requirement of ±0.3 mm. This method provides a reliable digital inspection solution for sheet metal parts. Full article

Computer vision is of wide interest to perform the phenotyping of horticultural crops such as apple trees at high throughput. In orchards specially constructed for variety testing or breeding programs, computer vision tools should be able to extract phenotypical information form each tree separately. We focus on segmenting individual apple trees as the main task in this context. Segmenting individual apple trees in dense orchard rows is challenging because of the complexity of outdoor illumination and intertwined branches. Traditional methods rely on supervised learning, which requires a large amount of annotated data. In this study, we explore an alternative approach using prompt engineering with the Segment Anything Model and its variants in a zero-shot setting. Specifically, we first detect the trunk and then position a prompt (five points in a diamond shape) located above the detected trunk to feed to the Segment Anything Model. We evaluate our method on the apple REFPOP, a new large-scale European apple tree dataset and on another publicly available dataset. On these datasets, our trunk detector, which utilizes a trained YOLOv11 model, achieves a good detection rate of $97\%$ based on the prompt located above the detected trunk, achieving a Dice score of 70% without training on the REFPOP dataset and 84% without training on the publicly available dataset.We demonstrate that our method equals or even outperforms purely supervised segmentation approaches or non-prompted foundation models. These results underscore the potential of foundational models guided by well-designed prompts as scalable and annotation-efficient solutions for plant segmentation in complex agricultural environments. Full article

Adversarial machine learning exploits the vulnerabilities of artificial intelligence (AI) models by inducing malicious distortion in input data. Starting with the effect of adversarial methods on well-known MNIST and CIFAR-10 open datasets, this paper investigates the ability of Uniform Manifold Approximation and Projection (UMAP) in providing useful representations of both legitimate and malicious images and analyzes the attacks’ behavior under various conditions. By enabling the extraction of decision rules and the ranking of important features from classifiers such as decision trees, eXplainable AI (XAI) achieves zero false positives and negatives in detection through very simple if-then rules over UMAP variables. Several examples are reported in order to highlight attacks behaviour. The data availability statement details all code and data which is publicly available to offer support to reproducibility. Full article

Moringa peregrina (Forssk.) Fiori is a tropical tree in southern Iran known as the most important natural coagulant in the world. Today, plant tissue culture is a new method that has a very high potential to produce valuable medicinal compounds on a commercial level. Advances in in vitro cultivation methods have increased the usefulness of plants as renewable resources. In this study, in addition to the phytochemical analysis of the extract of M. peregrina using HPLC, the interaction effect of different concentrations of aqueous extract of M. peregrina (0, 1, 1.5, and 3 mg/L) in two types of MS and ½ MS basal culture media over three weeks on the in vitro growth of Staurogyne repens (Nees) Kuntze was studied. The amounts of quercetin, gallic acid, caffeic acid, and myricetin in the aqueous extract of M. peregrina were 64.9, 374.8, 42, and 4.6 mg/g, respectively. The results showed that using M. peregrina leaf aqueous extract had a positive effect on the length of the branches, the percentage of green leaves, rooting, and the fresh and dry weight of S. repens samples. The highest increase in growth indices was observed in the MS culture medium supplemented with 3 mg/L of M. peregrina leaf aqueous extract after three weeks of cultivation. Of course, this effect was significantly greater in the MS medium and at higher concentrations compared to the ½ MS medium. Three weeks after cultivation at a concentration of 3 mg/L of the extract, the length of the S. repens branches was 5.3 and 1.8 cm in the two basic MS and ½ MS culture media, and the percentage of green leaves was 14 and 4 percent, respectively. Also, rooting was measured at 9.6 and 3.6 percent, fresh weight at 6 and 1.4 g, and dry weight at 1.1 and 0.03 g, respectively. Therefore, adding M. peregrina leaf aqueous extract as a stimulant significantly increased the in vitro growth of S. repens. Full article

Sugary phloem exudates are produced by many plant species and play key roles in carbon storage, defense, and ecological interactions. Among eucalypts, one such exudate, manna, is an important carbohydrate source for birds, mammals, and insects. Despite its ecological relevance, little is known about the composition and intra-specific variability of manna. Here, we investigated patterns of manna production in Eucalyptus viminalis, a widespread foundation tree species in southeastern Australia. We developed a repeatable ex situ method to extract and analyze manna, allowing us to characterize its sugar composition and examine variation within and between trees. Across years, manna contained six sugars, with sucrose and raffinose dominant. We found substantial variation in both the quality (sucrose/raffinose ratio) and quantity (mg) of manna produced. Both declined with increasing tree size (DBH), while quality increased with branch circumference. Seasonal and annual variation in manna was also evident, with quality increasing under drier conditions (positive correlation with aridity). Our findings demonstrate substantial intra-specific variation in phloem exudates (manna), shaped by temporal and tree-level factors. These patterns offer a foundation for future research into the ecological and physiological drivers of exudate variation and resource availability in foundation species like E. viminalis. Full article

3D object detection is a fundamental task in autonomous driving. In recent years, voxel-based methods have demonstrated significant advantages in reducing computational complexity and memory consumption when processing large-scale point cloud data. A representative method, Voxel-RCNN, introduces Region of Interest (RoI) pooling on voxel features, successfully bridging the gap between voxel and point cloud representations for enhanced 3D object detection. However, its robustness deteriorates when detecting distant objects or in the presence of noisy points (e.g., traffic signs and trees). To address this limitation, we propose an enhanced approach named Self-Attention Voxel-RCNN (SA-VoxelRCNN). Our method integrates two complementary attention mechanisms into the feature extraction phase. First, a full self-attention (FSA) module improves global context modeling across all voxel features. Second, a deformable self-attention (DSA) module enables adaptive sampling of representative feature subsets at strategically selected positions. After extracting contextual features through attention mechanisms, these features are fused with spatial features from the base algorithm to form enhanced feature representations, which are subsequently input into the region proposal network (RPN) to generate high-quality 3D bounding boxes. Experimental results on the KITTI test set demonstrate that SA-VoxelRCNN achieves consistent improvements in challenging scenarios, with gains of 2.49 and 1.87 percentage points at Moderate and Hard difficulty levels, respectively, while maintaining real-time performance at 22.3 FPS. This approach effectively balances local geometric details with global contextual information, providing a robust detection solution for autonomous driving applications. Full article