Search Results (148)

The disposal of orchard garbage (including pruning branches, fallen leaves, and non-biodegradable materials such as pesticide containers and plastic film) poses major difficulties for horticultural production and soil sustainability. Unlike general agricultural garbage, orchard garbage often contains both biodegradable organic matter and hazardous pollutants, which complicates efficient recycling. Traditional manual sorting methods are labour-intensive and inefficient in large-scale operations. To this end, we propose a lightweight YOLOv7-based detection model tailored for the orchard environment. By replacing the CSPDarknet53 backbone with MobileNetV3 and GhostNet, an average accuracy (mAP) of 84.4% is achieved, while the computational load of the original model is only 16%. Meanwhile, a supervised comparative learning strategy further strengthens feature discrimination between horticulturally relevant categories and can distinguish compost pruning residues from toxic materials. Experiments on a dataset containing 16 orchard-specific garbage types (e.g., pineapple shells, plastic mulch, and fertiliser bags) show that the model has high classification accuracy, especially for materials commonly found in tropical orchards. The lightweight nature of the algorithm allows for real-time deployment on edge devices such as drones or robotic platforms, and future integration with robotic arms for automated collection and sorting. By converting garbage into a compostable resource and separating contaminants, the technology is aligned with the country’s garbage segregation initiatives and global sustainability goals, providing a scalable pathway to reconcile ecological preservation and horticultural efficiency. Full article

In order to investigate the stage plant architecture and productivity characteristics of different cucumber varieties and pruning methods and to construct a comprehensive productivity evaluation system based on plant architecture parameters, this study used JY35 and JS206 as experiment materials and conducted a dual factor control experiment with variety and pruning methods (single-stem pruning (SP) and natural growth (NG)) to systematically analyze the key phenotypic characteristics and productivity indicators of cucumbers at different developmental stages based on variance analysis and principal component evaluation. The results indicated the following: (1) Varieties and pruning methods have a significant impact on the plant architecture characteristics and productivity indicators. (2) The dominant plant architecture characteristics and productivity indicators of JY35 include dry and fresh weights of the tendril, main stem, total stem, leaves, petioles, flowers, overground parts, and overall plant, as well as dry and fresh weight distribution index of the tendril, total stem, leaves, petioles, flowers, overground parts, and overall plant, main stem fresh weight distribution index, water content of roots, tendrils, main stem, leaves, petioles, and flowers, volume of total stem, main stem, and petioles, plant height, total leaf area per plant, leaf area index, and specific leaf area. The remaining plant architecture characteristics and productivity indicators are dominated by the plant architecture of JS206. (3) The dominant plant architecture characteristics and productivity indicators of the SP method include dry and fresh weight distribution index of roots, fruit carpopodiums, main stems, and total stems, water content of petioles, stems, and leaves, and root-to-shoot ratio. The remaining plant architecture characteristics and productivity indicators are dominated by the NG method. This study quantified the dynamic correlation between cucumber plant architecture and productivity characteristics, and the research results can provide a morphological basis for facility cucumber variety breeding and theoretical support for optimizing pruning cultivation mode and achieving efficient utilization of light and heat resources. Full article

The current image registration models have problems such as low feature point matching accuracy, high memory consumption, and significant computational complexity in heterogeneous image registration, especially in complex environments. In this context, significant differences in lighting and leaf occlusion in orchards can result in inaccurate feature extraction during heterogeneous image registration. To address these issues, this study proposes an AD-ResSug model for heterogeneous image registration. First, a VGG16 network was included as the encoder in the feature point encoder system, and the positional encoding was embedded into the network. This enabled us to better understand the spatial relationships between feature points. The addition of residual structures to the feature point encoder aimed to solve the gradient diffusion problem and enhance the flexibility and scalability of the architecture. Then, we used the Sinkhorn AutoDiff algorithm to iteratively optimize and solve the optimal transmission problem, achieving optimal matching between feature points. Finally, we carried out network pruning and compression operations to minimize parameters and computation cost while maintaining the model’s performance. This new AD-ResSug model uses evaluation indicators such as peak signal-to-noise ratio and root mean square error as well as registration efficiency. The proposed method achieved robust and efficient registration performance, verified through experimental results and quantitative comparisons of processing color with ToF images captured using heterogeneous cameras in natural apple orchards. Full article

In recent decades, substantial evidence has highlighted the integral roles of neuroglia, particularly astrocytes, microglia, oligodendrocytes, and ependymal cells, in the regulation of synaptic transmission, metabolic support, and immune mechanisms within the central nervous system. In addition to their structural role, these cells actively modulate neurotransmitter homeostasis and influence neuronal plasticity, thereby affecting cognition, mood, and behavior. This review discusses how neuroglial alterations contribute to the pathophysiology of five common psychiatric disorders: major depression, bipolar disorder, anxiety disorders, attention-deficit/hyperactivity disorder (ADHD), and schizophrenia. We synthesized preclinical and clinical findings illustrating that glial dysfunction, including impaired myelination and aberrant neuroinflammatory responses, often parallels disease onset and severity. Moreover, we outline how disruptions in astrocytic glutamate uptake, microglia-mediated synaptic pruning, and blood–brain barrier integrity may underlie the neurobiological heterogeneity observed in these disorders. The therapeutic implications range from anti-inflammatory agents to investigational compounds that aim to stabilize glial function or promote remyelination. However, challenges due to interindividual variability, insufficient biomarkers, and the multifactorial nature of psychiatric illnesses remain. Advances in neuroimaging, liquid biopsy, and more precise molecular techniques may facilitate targeted interventions by stratifying patient subgroups with distinct glial phenotypes. Continued research is essential to translate these insights into clinically efficacious and safe treatments. Full article

Biogenic volatile organic compounds (BVOCs) significantly impact air quality and climate. Mechanical injury is a common stressor affecting plants in both natural and urban environments, and it has potentially large influences on BVOC emissions. However, the interspecific variability in wounding-induced BVOC emissions remains poorly understood, particularly for subtropical trees and shrubs. In this study, we investigated the effects of controlled mechanical injury on isoprenoid and aromatic compound emissions in a taxonomically diverse set of 45 subtropical broad-leaved woody species, 26 species without and in 19 species with BVOC storage structures (oil glands, resin ducts and glandular trichomes for volatile compound storage). Emissions of light-weight non-stored isoprene and monoterpenes and aromatic compounds in non-storage species showed moderate and variable emission increases after mechanical injury, likely reflecting the wounding impacts on leaf physiology. In storage species, mechanical injury triggered a substantial release of monoterpenes and aromatic compounds due to the rupture of storage structures. Across species, the proportion of monoterpenes in total emissions increased from 40.9% to 85.4% after mechanical injury, with 32.2% of this increase attributed to newly released compounds not detected in emissions from intact leaves. Sesquiterpene emissions, in contrast, were generally low and decreased after mechanical injury. Furthermore, wounding responses varied among plant functional groups, with evergreen species and those adapted to high temperatures and shade exhibiting stronger damage-induced BVOC emissions than deciduous species and those adapted to dry or cold environments. These findings suggest that mechanical disturbances such as pruning can significantly enhance BVOC emissions in subtropical urban forests and should be considered when modeling BVOC fluxes in both natural and managed ecosystems. Further research is needed to elucidate the relationship between storage structure characteristics and BVOC emissions, as well as their broader ecological and atmospheric implications. Full article

Vineyards are significant demanders of fertilisers, pesticides, soil tillage and water. This study assessed the environmental profile of an organic grape production system with La Rioja qualified designation of origin using a cradle-to-gate life cycle assessment (LCA). The ReCiPe method was applied to assess the environmental impacts, while the Available WAter REmaining method was used to estimate the water scarcity. Additionally, the biodiversity loss, a global issue exacerbated by agricultural practices, was evaluated along with an ecosystem service indicator, pollination, to provide a more comprehensive analysis. This study employed two functional units: one kilogram of grapes and one hectare of land. The results revealed that the environmental impacts on global warming were more than ten times lower than those reported in most studies reviewed in the literature, primarily due to the effects of direct land use changes associated with pruning waste management. The total emissions in this category were 99.51 kg CO₂ eq per hectare or 15.31 g CO₂ eq per kilogram of grapes. Agrochemical-related emissions were identified as the environmental hotspot. The water scarcity was estimated at 48.4 litres per kilogram of grapes, mainly attributed to agrochemical dispersion. The biodiversity loss was largely driven by land transformation, with plants being the most impacted taxon. However, a high abundance of pollinators was observed in spring, contributing to improved grape quality and natural pest control. These findings could help highlight the environmental benefits of organic viticulture and the good practices implemented in this pilot. Full article

Forestry and wood-processing by-products, such as pine bark, offer promising opportunities for sustainable resource utilization within a circular economy. This study aimed to assess the environmental impact of an aqueous extraction process for polyphenolic compounds from various pine residues, including bark, cones, and pruning, using life cycle assessment (LCA). The analysis revealed that ground and sieved pine bark powder had the lowest environmental impact, attributed to its simpler extraction process without chemical modifications and reduced energy consumption compared to other pine-derived products. Electricity and natural gas were identified as the primary drivers of environmental impacts across all categories. Sensitivity analyses demonstrated that increasing the tannin concentration in pine-derived products and integrating renewable energy sources could further improve environmental performance. These findings highlight the potential of utilizing underutilized pine residues as sustainable feedstock for producing valuable polyphenolic extracts with a relatively low environmental footprint. The insights gained from this LCA study provide a comprehensive foundation for advancing sustainable extraction technologies. They emphasize the critical role of energy efficiency, tannin concentration, and renewable energy integration in minimizing environmental impacts. Furthermore, these findings offer actionable guidance for optimizing resource recovery from forestry by-products, enhancing their viability as eco-friendly alternatives to conventional tannin sources. Full article

Gesture is a natural and intuitive means of interpersonal communication. Sign language recognition has become a hot topic in scientific research, holding significant importance and research value in fields such as deep learning, human–computer interaction, and pattern recognition. The sign language recognition process needs to ensure real-time performance and ease of deployment. Based on these two requirements, this paper proposes an improved YOLOv5s-based sign language recognition algorithm. Firstly, the lightweight concept from ShuffleNetV2 was applied to achieve lightweight characteristics and improve the model’s deployability. The specific improvements are as follows: The algorithm achieved model size reduction by removing the Focus layer, using the ShuffleNetv2 algorithm, and then channel pruning YOLOv5 at the head of the neck layer. All the convolutional layers and the cross-stage partial bottleneck layer with three convolutional layers in the backbone network were replaced with ShuffleBlock, the spatial pyramid pooling layer and a subsequent cross-stage partial bottleneck layer structure with three convolutional layers were removed, and the cross-stage partial bottleneck layer module with three convolutional layers in the detection header section was replaced with a depth-separable convolutional module. Experimental results show that the parameters of the improved YOLOv5 algorithm decreased from 7.2 M to 0.72 M, and the inference speed decreased from 3.3 ms to 1.1 ms. Full article

Large Language Models (LLMs), like GPT and BERT, have significantly advanced Natural Language Processing (NLP), enabling high performance on complex tasks. However, their size and computational needs make LLMs unsuitable for deployment on resource-constrained devices, where efficiency, speed, and low power consumption are critical. Tiny Language Models (TLMs), also known as BabyLMs, offer compact alternatives by using advanced compression and optimization techniques to function effectively on devices such as smartphones, Internet of Things (IoT) systems, and embedded platforms. This paper provides a comprehensive survey of TLM architectures and methodologies, including key techniques such as knowledge distillation, quantization, and pruning. Additionally, it explores potential and emerging applications of TLMs in automation and control, covering areas such as edge computing, IoT, industrial automation, and healthcare. The survey discusses challenges unique to TLMs, such as trade-offs between model size and accuracy, limited generalization, and ethical considerations in deployment. Future research directions are also proposed, focusing on hybrid compression techniques, application-specific adaptations, and context-aware TLMs optimized for hardware-specific constraints. This paper aims to serve as a foundational resource for advancing TLMs capabilities across diverse real-world applications. Full article

This study addresses the challenges of detecting inferior fruits in table grapes in natural environments, focusing on subtle appearance differences, occlusions, and fruit overlaps. We propose an enhanced green grape fruit disease detection model named RF-YOLOv7. The model is trained on a dataset comprising images of small fruits, sunburn, excess grapes, fruit fractures, and poor-quality grape bunches. RF-YOLOv7 builds upon the YOLOv7 architecture by integrating four Contextual Transformer (CoT) modules to improve target-detection accuracy, employing the Wise-IoU (WIoU) loss function to enhance generalization and overall performance, and introducing the Bi-Former attention mechanism for dynamic query awareness sparsity. The experimental results demonstrate that RF-YOLOv7 achieves a detection accuracy of 83.5%, recall rate of 76.4%, mean average precision (mAP) of 80.1%, and detection speed of 58.8 ms. Compared to the original YOLOv7, RF-YOLOv7 exhibits a 3.5% increase in mAP, with only an 8.3 ms increase in detection time. This study lays a solid foundation for the development of automatic detection equipment for intelligent grape pruning. Full article

The Internet of Things (IoT) ecosystem is rapidly expanding. It is driven by continuous innovation but accompanied by increasingly sophisticated cybersecurity threats. Protecting IoT devices from these emerging vulnerabilities has become a critical priority. This study addresses the limitations of existing IoT threat detection methods, which often struggle with the dynamic nature of IoT environments and the growing complexity of cyberattacks. To overcome these challenges, a novel hybrid architecture combining Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and Deep Neural Networks (DNN) is proposed for accurate and efficient IoT threat detection. The model’s performance is evaluated using the IoT-23 and Edge-IIoTset datasets, which encompass over ten distinct attack types. The proposed framework achieves a remarkable 99% accuracy on both datasets, outperforming existing state-of-the-art IoT cybersecurity solutions. Advanced optimization techniques, including model pruning and quantization, are applied to enhance deployment efficiency in resource-constrained IoT environments. The results highlight the model’s robustness and its adaptability to diverse IoT scenarios, which address key limitations of prior approaches. This research provides a robust and efficient solution for IoT threat detection, establishing a foundation for advancing IoT security and addressing the evolving landscape of cyber threats while driving future innovations in the field. Full article

Frequent consumption of processed meat has been classified as carcinogenic to humans by the International Agency for Research on Cancer (Group 1), while red meat has been classified as probably carcinogenic (Group 2A). Mutagenic and carcinogenic compounds formed by heating in protein-rich food include, among others, heterocyclic aromatic amines (HAAs). Modifying the heat treatment of meat and using natural additives with antioxidant properties can lead to a reduction in their formation. The aim of this study was to determine polar HAAs (imidazoquinolines, IQ and MeIQ; imidazoquinoxalines, 8-MeIQx and 4,8-DiMeIQx; and phenylimidazopyridine, PhIP) in pork loin prepared without additives and with three types of dried fruit (apricots, cranberries, and prunes), baked in a roasting bag. HAAs were isolated from meat samples by solid-phase extraction. Quantitative analysis was performed by high-performance liquid chromatography with fluorescence detection (FLD) and a diode array detector (DAD). Only two HAAs, 8-MeIQx and PhIP, were detected in extracts isolated from meat samples. The total content of these compounds in meat roasted without additives was 5.9 ng/g. Using a dried fruit stuffing content of 200 g/kg of meat reduced these concentrations in dishes prepared with prunes, apricots, and cranberries by 42%, 47%, and 77%, respectively. Full article

The European hazelnut is a temperate nut crop, often managed as a multi-stemmed shrub due to its natural aptitude to produce several suckers at the base of the stump, depending on variety. Traditional hazelnut-growing regions such as Italy, Turkey, and Spain typically adopt this architecture, while other hazelnut-producing countries such as the USA, France, and Chile increasingly use single-trunk systems to facilitate orchard management. Multi-stemmed plants allow gradual renewal through sucker selection but may lead to excessively dense canopies, reducing the effectiveness of pest and disease control, increasing biennial bearing, and lowering nut yields in vigorous cultivars. In order to drive the ongoing enlargement of hazelnut cultivation, attempts in designing high-density (HD), and more occasionally super-high-density (SHD), hazelnut orchards are on-going, although these are poorly explored in terms of suitable plant training systems, such that, sometimes, multi-stemmed plant shapes are used; otherwise, single-trunk solutions are adopted. In order to explore new hazelnut planting and training solutions focused on sustainable intensification, a trial was established in 2019 in central Italy to evaluate the eligibility of three training systems (treatment A: regular four-stemmed shrub; treatment B: single-trunk sapling; treatment C: traditional multi-stemmed shrub), applied on unpruned three-year-old plants of the hazelnut cultivar Nocchione, planted in the HD approach (740 plant ha⁻¹). Over five growing seasons (2019–2023), measurements included pruned wood removed, yield, vigor, yield efficiency, nut and kernel traits, and incidence of the main commercial defects. In general, treatment A outperformed other plant-shaping systems, maintaining high yield levels particularly in the two last growing seasons, and showing a mean kernel/nut ratio of 37.7 and a low incidence of defects. Treatment B achieved the highest yield efficiency in 2023 but had lower overall yields. Treatment A demonstrated the most balanced performance, combining high nut quality and stable production, making it the most promising plant training system for HD hazelnut orchards with planting densities above 700 plants per hectare. Future research will assess the long-term adaptability of this plant training system under varying environmental and management conditions. Full article

Background: Camellia oleifera Abel (C. oleifera) is widely cultivated and serves as an important source of edible oil. Yet, during oil production, pruned branches generate significant waste and contribute to environmental pollution. Objectives: In this work, we obtain a natural polysaccharide from the branches of C. oleifera and optimize its extraction using Box–Behnken design (BBD), which is a statistical method commonly used in response surface methodology. Additionally, we study its properties, such as monosaccharide composition, structural features, antioxidant, and anti-inflammatory abilities. Results: BBD was employed to optimize polysaccharide extraction (solid-liquid ratio = 1:40; 90 °C; 130 min) for a higher yield. After purification, the major monosaccharides of branches of C. oleifera’s polysaccharide (CBP) were disclosed as glucose and galactose. Subsequent structural features of CBP were measured. The antioxidant and anti-inflammatory abilities were measured. The highly scavenging rates of the 2,2-diphenyl-1-picrylhydrazyl and hydroxyl radicals, with the chelating capacity of Fe²⁺, indicate potent antioxidant activity of CBP. Conclusions: In general, CBP demonstrated significant anti-inflammatory activity with down-regulating the expression of IL-6 and IL-1β in the LPS-induced macrophage RAW264.7 model. This bioactive polysaccharide adds value to waste branches by offering a novel approach to waste recycling and the development of C. oleifera. Full article

Large Language Models (LLMs) like BERT have gained significant prominence due to their remarkable performance in various natural language processing tasks. However, they come with substantial computational and memory costs. Additionally, they are essentially black-box models, being challenging to explain and interpret. In this article, Persistent BERT Compression and Explainability (PBCE) is proposed, a Green AI methodology to prune BERT models using persistent homology, aiming to measure the importance of each neuron by studying the topological characteristics of their outputs. As a result, PBCE can compress BERT significantly by reducing the number of parameters (47% of the original parameters for BERT Base, 42% for BERT Large). The proposed methodology has been evaluated on the standard GLUE Benchmark, comparing the results with state-of-the-art techniques achieving outstanding results. Consequently, PBCE can simplify the BERT model by providing explainability to its neurons and reducing the model’s size, making it more suitable for deployment on resource-constrained devices. Full article