Search

To explore the wound healing of knots in young Chinese fir stands, this study used 6-year-old Chinese fir plantations in the Weimin State-owned Forest Farm in Shaowu, Fujian Province, as research subjects. Experiments were designed using different pruning intensities, retention densities, knot directions, wound diameters, and paint protection. The healing rate of wounds was statistically analyzed over 20 months after pruning. Different pruning intensities were found to have significant effects on the healing rate of transverse wounds in young Chinese fir stands (P＜0. 05), with the order of transverse wound healing rates being 10 cm pruning intensity＞12 cm pruning intensity＞8 cm pruning intensity. The fastest wound healing rate was observed at a retention density of 1 800 tree· hm-2; where under a pruning intensity ＜8 cm, the knot healing rate of wounds increased with the increase in retention density of the young Chinese fir stands; a pruning intensity ＜10 cm resulted in the knot healing rate of wounds first decreasing and then increasing with the increase in retention density of the young Chinese fir stands. Knot wounds on the northeast side healed the fastest, while those on the southwest side healed the slowest, although different knot directions, different knot wound diameters, and paint protection had no significant effects on the healing rate of Chinese fir knot wounds (P＞0. 05) . In general, the healing rate of transverse wounds was always greater than that of longitudinal wounds under different treatments, indicating that transverse wounds would be completely healed sooner. [ABSTRACT FROM AUTHOR]

Automating pruning tasks entails overcoming several challenges, encompassing not only robotic manipulation but also environment perception and detection. To achieve efficient pruning, robotic systems must accurately identify the correct cutting points. A possible method to define these points is to choose the cutting location based on the number of nodes present on the targeted cane. For this purpose, in grapevine pruning, it is required to correctly identify the nodes present on the primary canes of the grapevines. In this paper, a novel method of node detection in grapevines is proposed with four distinct state-of-the-art versions of the YOLO detection model: YOLOv7, YOLOv8, YOLOv9 and YOLOv10. These models were trained on a public dataset with images containing artificial backgrounds and afterwards validated on different cultivars of grapevines from two distinct Portuguese viticulture regions with cluttered backgrounds. This allowed us to evaluate the robustness of the algorithms on the detection of nodes in diverse environments, compare the performance of the YOLO models used, as well as create a publicly available dataset of grapevines obtained in Portuguese vineyards for node detection. Overall, all used models were capable of achieving correct node detection in images of grapevines from the three distinct datasets. Considering the trade-off between accuracy and inference speed, the YOLOv7 model demonstrated to be the most robust in detecting nodes in 2D images of grapevines, achieving F1-Score values between 70% and 86.5% with inference times of around 89 ms for an input size of 1280 × 1280 px. Considering these results, this work contributes with an efficient approach for real-time node detection for further implementation on an autonomous robotic pruning system. [ABSTRACT FROM AUTHOR]

Avocado orchards (Persea americana) in Mexico are constantly being expanded to meet the increasing demand for the fruit in the national and international markets. The land-use change (LUC) caused by this expansion has numerous negative impacts, including greenhouse gas (GHG) emissions due to the loss of forest cover and the burning of pruning residues. To generate a comprehensive evaluation of this complex environmental issue, we calculate emissions from LUC and from residue burning between 1974 and 2017 at a local scale (1:20,000), and the energy potential of pruning residues was estimated as an alternative to revalue a waste product and mitigate the negative impacts of avocado cultivation. Our results show that land-use conversions emitted 390.5 GgCO2, of which 91% came from conversions to avocado orchards. Emissions of GHG from biomass burning amounted to an additional 20.68 GgCO2e released per year. Given that around 12,600 tons of dry avocado pruning residues are generated annually in the study region, their use for energy generation could replace 240 TJ/year of fossil fuels in rural industries and could mitigate around 31 GgCO2e per year. This study provides decision-makers with a concrete example of how to establish multiple-impact strategies at local scales. [ABSTRACT FROM AUTHOR]

Time Series Classification (TSC) is essential in fields like medicine, environmental science, and finance, enabling tasks such as disease diagnosis, anomaly detection, and stock price analysis. While machine learning models like Recurrent Neural Networks and InceptionTime are successful in numerous applications, they can face scalability issues due to computational requirements. Recently, ROCKET has emerged as an efficient alternative, achieving state-of-the-art performance and simplifying training by utilizing a large number of randomly generated features from the time series data. However, many of these features are redundant or non-informative, increasing computational load and compromising generalization. Here we introduce Sequential Feature Detachment (SFD) to identify and prune non-essential features in ROCKET-based models, such as ROCKET, MiniRocket, and MultiRocket. SFD estimates feature importance using model coefficients and can handle large feature sets without complex hyperparameter tuning. Testing on the UCR archive shows that SFD can produce models with better test accuracy using only 10% of the original features. We named these pruned models Detach-ROCKET. We also present an end-to-end procedure for determining an optimal balance between the number of features and model accuracy. On the largest binary UCR dataset, Detach-ROCKET improves test accuracy by 0.6% while reducing features by 98.9%. By enabling a significant reduction in model size without sacrificing accuracy, our methodology improves computational efficiency and contributes to model interpretability. We believe that Detach-ROCKET will be a valuable tool for researchers and practitioners working with time series data, who can find a user-friendly implementation of the model at https://github.com/gon-uri/detach_rocket. [ABSTRACT FROM AUTHOR]

Highlights The growing global demand for sustainable and environmentally friendly polymer materials is driving interest in cellulose‐based materials. In response to this need, lignocellulosic fillers (LF) were extracted from pruning waste of bing cherry tree (Prunus Avium L.) branches as an alternative source of filler materials. The extracted LF were characterized by Fourier transform infrared (FTIR) spectroscopy, X‐ray diffraction (XRD), thermogravimetric analysis (TGA) and scanning electron microscopy (SEM). Their chemical composition, density and particle size distribution (PSD) were also analyzed. In the second part of the study, biocomposites were prepared by incorporating fillers with particle sizes below 100 microns into an epoxy matrix at concentrations of 5%, 10% and 15% by weight. These biocomposites were then characterized by tensile test, three‐point bending test and SEM analyze to determine their mechanical and morphological properties. Among the biocomposites, the one with 5% wood filler showed the best properties with a tensile strength of 45 MPa, tensile modulus of 1883 MPa, flexural strength of 74 MPa and flexural modulus of 2559 MPa. The results demonstrate the effectiveness of lignocellulosic particles in improving polymer matrices and suggest their potential for use in non‐structural applications in the automotive and marine industries, such as interior panels. Cherry tree pruning waste has been characterized for the first time. The potential use of cherry tree pruning waste as a filler material in thermosets has been investigated for the first time. This study aims to contribute to the reduction of plastic consumption and the development of environmentally friendly products. [ABSTRACT FROM AUTHOR]

This paper proposes the "ultimate compression" method as a solution to the expansive computation and high storage costs required by state-of-the-art neural network models in inference. Our approach uniquely combines tensor decomposition techniques with binary neural networks to create efficient deep neural network models optimized for edge inference. The process includes training floating-point models, applying tensor decomposition algorithms, binarizing the decomposed layers, and fine tuning the resulting models. We evaluated our approach in various state-of-the-art deep neural network architectures on multiple datasets, such as MNIST, CIFAR-10, CIFAR-100, and ImageNet. Our results demonstrate compression ratios of up to 169×, with only a small degradation in accuracy (1–2%) compared to binary models. We employed different optimizers for training and fine tuning, including Adam and AdamW, and used norm grad clipping to address the exploding gradient problem in decomposed binary models. A key contribution of this work is a novel layer sensitivity-based rank selection algorithm for tensor decomposition, which outperforms existing methods such as random selection and Variational Bayes Matrix Factorization (VBMF). We conducted comprehensive experiments using six different models and present a case study on crowd-counting applications, demonstrating the practical applicability of our method. The ultimate compression method outperforms binary neural networks and tensor decomposition when applied individually in terms of storage and computation costs. This positions it as one of the most effective options for deploying compact and efficient models in edge devices with limited computational resources and energy constraints. [ABSTRACT FROM AUTHOR]

Strip steel surface defect recognition research has important research significance in industrial production. Aiming at the problems of defect feature extraction, slow detection speed, and insufficient datasets, YOLOv5 is improved on the basis of YOLOv5, and the YOLO-LFPD (lightweight fine particle detection) model is proposed. By introducing the RepVGG (Re-param VGG) module, the robustness of the model is enhanced, and the expressive ability of the model is improved. FasterNet is used to replace the backbone network, which ensures accuracy and accelerates the inference speed, making the model more suitable for real-time monitoring. The use of pruning, a GA genetic algorithm with OTA loss function, further reduces the model size while better learning the strip steel defect feature information, thus improving the generalisation ability and accuracy of the model. The experimental results show that the introduction of the RepVGG module and the use of FasterNet can well improve the model performance, with a reduction of 48% in the number of parameters, a reduction of 13% in the number of GFLOPs, an inference time of 77% of the original, and an optimal accuracy compared with the network models in recent years. The experimental results on the NEU-DET dataset show that the accuracy of YOLO-LFPD is improved by 3% to 81.2%, which is better than other models, and provides new ideas and references for the lightweight strip steel surface defect detection scenarios and application deployment. [ABSTRACT FROM AUTHOR]

This paper explores the utilization of biomass resources derived from vineyard residues for producing densified solid biofuels in the Republic of Moldova, with the aim of quantitatively and qualitatively evaluating the residue from vine pruning, focusing on the feasibility of its use as raw material for the production of briquettes and pellets. The methodology includes the analysis of statistical data, as well as experimental investigations conducted at the Scientific Laboratory of Solid Biofuels of the Technical University of Moldova. Waste biomass samples were collected from various vineyards in the different districts of all three regions of the country, focusing on regions with significant plantations. Both quantitative and qualitative aspects of the biomass were assessed, considering the moisture content, calorific value, and ash content. It was found that about 1013 kg/ha of waste biomass is generated from the pruning of technical grape varieties with a net calorific value of 15.6 MJ/kg at a moisture content of 10 wt.% and about 1044 kg/ha with a calorific value of 16.4 MJ/kg from the table ones; both with an average ash content of 3 wt.%. The results indicated that vineyard pruning residues in the Republic of Moldova could provide a substantial biomass source, with an estimated total energy potential of approximately 370 TJ/y (80% located in the Southern region); they also highlighted the need for technological advancements and quality assurance procedures through which to ensure the efficiency and sustainability of biofuel production. The conclusions emphasize the numerous benefits of utilizing viticultural residue, both economically and ecologically, contributing to the sustainable development of the viticulture industry in the Republic of Moldova, as well as environmental protection. [ABSTRACT FROM AUTHOR]

Agriculture stands as the cornerstone of Africa's economy, supporting over 60% of the continent's labor force. Despite its significance, the quality assessment of agricultural products remains a challenging task, particularly at a large scale, consuming valuable time and resources. The African plum is an agricultural fruit that is widely consumed across West and Central Africa but remains underrepresented in AI research. In this paper, we collected a dataset of 2892 African plum samples from fields in Cameroon representing the first dataset of its kind for training AI models. The dataset contains images of plums annotated with quality grades. We then trained and evaluated various state-of-the-art object detection and image classification models, including YOLOv5, YOLOv8, YOLOv9, Fast R-CNN, Mask R-CNN, VGG-16, DenseNet-121, MobileNet, and ResNet, on this African plum dataset. Our experimentation resulted in mean average precision scores ranging from 88.2% to 89.9% and accuracies between 86% and 91% for the object detection models and the classification models, respectively. We then performed model pruning to reduce model sizes while preserving performance, achieving up to 93.6% mean average precision and 99.09% accuracy after pruning YOLOv5, YOLOv8 and ResNet by 10–30%. We deployed the high-performing YOLOv8 system in a web application, offering an accessible AI-based quality assessment tool tailored for African plums. To the best of our knowledge, this represents the first such solution for assessing this underrepresented fruit, empowering farmers with efficient tools. Our approach integrates agriculture and AI to fill a key gap. [ABSTRACT FROM AUTHOR]

In the scenario of industrial gas leakage, traditional object detection models face challenges such as high computational complexity, large parameter count, and slow detection speed, making it difficult to deploy them on terminal hardware devices with limited computational resources. Meanwhile, existing lightweight object detection models also encounter difficulties in balancing detection accuracy and real-time requirements. To address this issue, this paper proposes a lightweight object detection algorithm, P-YOLOv7-TFD, based on the YOLOv7-tiny model. Firstly, the YOLOv7-TFD algorithm is constructed, which utilizes FasterNet as the backbone network, reducing the computational complexity and improving feature representation capability through the application of rapid feature fusion and efficient upsampling modules, while maintaining accuracy and enhancing detection speed. Secondly, the network's head layer is reconstructed based on the diverse branch block module to enrich the diversity of feature space, thereby improving model performance without sacrificing inference speed. Finally, the YOLOv7-TFD model is pruned using the Network Slimming channel pruning algorithm to obtain the P-YOLOv7-TFD model, further reducing the model's parameter count. Experimental results show that, compared to the original YOLOv7-tiny model, the average precision of the P-YOLOv7-TFD object detection model decreases by only 1.6% on a self-built dataset, while the parameter count and computational load decrease by 75.7% and 71.8% respectively. The computation inference time is decreased by 88.3%, and the model weight size is only 4.4 MB, representing a reduction of 62.4%. [ABSTRACT FROM AUTHOR]

Copyright of Systems Engineering & Electronics is the property of Journal of Systems Engineering & Electronics Editorial Department and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Object detection is a key technology to realize the blind inspection of industrial products. To improve the automation degree of building deep convolutional neural networks (CNNs) for object detection and further improve the detection accuracy, this paper proposes an improved neural architecture search method using exclusive-OR (XOR)-based channel feature fusion. First, an XOR-based channel fusion module is designed; it can fuse the feature mapping of different scales at the channel level in the case of multibranch access complementarily. Then, an improved cell pruning strategy is proposed to efficiently prune the connections between cells by setting the architecture parameters of the candidate operations to 0 s, which are in the alignment layers of the subsequent cells. The cell pruning strategy can directly search the multibranch CNN models and narrow the neural network architectures' gap between the search stage and the evaluation stage. The experimental results show that the proposed method takes approximately 0.75 GPU days to search the optimal neural network on a dataset including six classes for blind inspection of industrial products, and the mean average precision (mAP) is approximately 99.1% on a test dataset, which is higher than those of state-of-the-art methods, e.g., DenseNAS and CSPDarknet53. [ABSTRACT FROM AUTHOR]

'Xinomavro' (V. vinifera L.) is an important native red wine grape variety in Northern Greece, particularly in PDO (protected designation of origin) regions. Despite its significance, there is limited research on the effects of pruning type and severity on 'Xinomavro' vine physiology, yield, and berry quality across diverse environmental conditions. This study aimed to address this knowledge gap and provide growers with crucial information for optimizing vineyard management practices. The study was conducted over two consecutive years (2016 and 2017) in a vineyard in Thessaloniki, Northern Greece. Four treatments (B12: 12 buds on 6 spurs, B24: 24 buds on 12 spurs, M12: 12 buds on 2 canes, and M24: 24 buds on 4 canes) combining two bud load levels (12 or 24 count nodes) and two pruning types (short spurs or long canes) were applied to 'Xinomavro' vines in a complete block randomized design. The vine water status, gas exchange, canopy characteristics, yield components, and berry composition were measured. Bud load and pruning type significantly influenced vine canopy development, microclimate, and yield components. Short pruning with high bud load (B24) resulted in denser canopies and higher yields, whereas cane pruning (M12 and M24) led to more open canopies and improved berry quality indicators. Treatment effects on berry composition were inconsistent across years but showed a tendency for higher anthocyanin and total phenol content in cane-pruned vines. This study demonstrates that pruning type (short or long fruiting units) may have a greater impact on vine growth, yield, and berry composition than bud load alone in 'Xinomavro' vines. Cane pruning appears to be a more effective strategy for achieving vine balance and potentially improving grape quality under given experimental conditions. [ABSTRACT FROM AUTHOR]

The prevalence of apples as the most widely consumed fruit globally does not exempt them from storage-related issues, resulting in substantial harvest losses. A prominent concern is the development of rot due to various factors during storage. This research endeavors to examine the influence of agrotechnological methods on the longevity of apples and the incidence of rot throughout storage. Apple trees (Malus domestica Borkh. cv. Rubin) grafted on dwarfing rootstocks P60 were planted in 2010 in single rows with a spacing of 1.25 m between trees and 3.5 m between rows. Eight combinations of different growth control measures (manual, mechanical pruning, spraying, trunk cutting) were selected for the experiment. The implementation of mechanical pruning, in conjunction with trunk cutting and Ca-prohexadione spraying, as well as summer pruning, detrimentally impacted the shelf life of apples. Examination of the storage period revealed a loss of 33–40% of the crop due to rot. Conversely, manual pruning sustained a consistent level of phenolic compounds throughout the storage period. Other pruning methods resulted in a notable increase in phenolic compounds, ranging from 67% to a two-fold rise compared to the compounds present at harvest. However, the integration of mechanical pruning with subsequent manual pruning not only significantly augmented the yield of apples but also yielded a shelf life akin to that of manually pruned apples. Following the analysis of the results, it is advisable to conduct mechanical pruning of the apples intended for storage along with supplementary manual pruning. [ABSTRACT FROM AUTHOR]

This study focuses on existing pruning equipment; cutting blades show cutting resistance and lead to high energy consumption. Using finite element (FEA) numerical simulation technology, the branch stress wave propagation mechanism during pruning was studied. The cutting performance of the bionic blade was evaluated with cutting energy consumption as the test index and the branch diameter and branch angle as the test factors, respectively. The test results showed that the blades imitating the mouthparts of the three-pecten bull and the beak of the woodpecker performed well in pruning, and the energy consumption during cutting was reduced by 18.2% and 16.3% compared to traditional blades, making these blades significantly better. These two blades also effectively reduced the cutting resistance and branch splitting by optimizing the edge angle design and increasing the slip-cutting action. In contrast, the imitation shark's tooth blade increased cutting energy consumption by 14.4% due to the large amount of cutting resistance in the cutting process when cutting larger-diameter branches, making it unsuitable for application in the pruning field. Therefore, the blades imitating the mouthparts of the three pectins and the beak of the woodpecker have significant advantages in reducing the cutting resistance and improving the pruning quality. These findings provide an important theoretical reference for the development of energy-efficient pruning equipment. [ABSTRACT FROM AUTHOR]

Landscape maintenance is essential for ensuring agricultural productivity, promoting sustainable land use, and preserving soil and ecosystem health. Pruning is a labor-intensive task among landscaping applications that often involves repetitive pruning operations. To address these limitations, this paper presents the development of a dual-arm holonomic robot (called the KOALA robot) for precision plant pruning. The robot utilizes a cross-functionality sensor fusion approach, combining light detection and ranging (LiDAR) sensor and depth camera data for plant recognition and isolating the data points that require pruning. The You Only Look Once v8 (YOLOv8) object detection model powers the plant detection algorithm, achieving a 98.5% pruning plant detection rate and a 95% pruning accuracy using camera, depth sensor, and LiDAR data. The fused data allows the robot to identify the target boxwood plants, assess the density of the pruning area, and optimize the pruning path. The robot operates at a pruning speed of 10–50 cm/s and has a maximum robot travel speed of 0.5 m/s, with the ability to perform up to 4 h of pruning. The robot's base can lift 400 kg, ensuring stability and versatility for multiple applications. The findings demonstrate the robot's potential to significantly enhance efficiency, reduce labor requirements, and improve landscape maintenance precision compared to those of traditional manual methods. This paves the way for further advancements in automating repetitive tasks within landscaping applications. [ABSTRACT FROM AUTHOR]

Spiking neural networks (SNNs) have received increasing attention due to their high biological plausibility and energy efficiency. The binary spike-based information propagation enables efficient sparse computation in event-based and static computer vision applications. However, the weight precision and especially the membrane potential precision remain as high-precision values (e.g., 32 bits) in state-of-the-art SNN algorithms. Each neuron in an SNN stores the membrane potential over time and typically updates its value in every time step. Such frequent read/write operations of high-precision membrane potential incur storage andmemory access overhead in SNNs, which undermines the SNNs' compatibility with resource-constrained hardware. To resolve this inefficiency, prior works have explored the time step reduction and low-precision representation of membrane potential at a limited scale and reported significant accuracy drops. Furthermore, while recent advances in on-device AI present pruning and quantization optimization with different architectures and datasets, simultaneous pruning with quantization is highly under-explored in SNNs. In this work, we present SpQuant-SNN, a fully-quantized spiking neural network with ultra-low precision weights, membrane potential, and high spatial-channel sparsity, enabling the end-to-end low precision with significantly reduced operations on SNN. First, we propose an integer-only quantization scheme for the membrane potential with a stacked surrogate gradient function, a simple-yet-effective method that enables the smooth learning process of quantized SNN training. Second, we implement spatial-channel pruning with membrane potential prior, toward reducing the layer-wise computational complexity, and floating-point operations (FLOPs) in SNNs. Finally, to further improve the accuracy of low-precision and sparse SNN, we propose a self-adaptive learnable potential threshold for SNN training. Equipped with high biological adaptiveness, minimal computations, and memory utilization, SpQuant-SNN achieves state-of-the-art performance across multiple SNN models for both event-based and static image datasets, including both image classification and object detection tasks. The proposed SpQuant-SNN achieved up to 13× memory reduction and >4.7× FLOPs reduction with <1.8% accuracy degradation for both classification and object detection tasks, compared to the SOTA baseline. [ABSTRACT FROM AUTHOR]

Horticulture, a critical component of agriculture, encounters various challenges, including crop loss stemming from factors like pests, diseases, adverse weather conditions, and inefficient farming practices. The introduction of advanced technologies such as robotics and artificial intelligence (AI) holds great promise in mitigating crop losses and bolstering productivity in the field of horticulture. Robotic systems have been devised to automate labor‐intensive tasks involved in horticulture, such as harvesting, pruning, and weeding. Equipped with sensors, cameras, and intelligent algorithms, these robots are capable of identifying ripe fruits, detecting and removing weeds, and performing precise pruning operations. For example, Peixoto et al. in 2015 employed fuzzy systems to create a model for controlling soybean aphids, significantly improving the timing of predator release and enhancing integrated pest management (IPM). By reducing the reliance on human labor and enhancing operational efficiency, the integration of robotic solutions contribute to the minimization of crop losses and the augmentation of yields. In horticulture crop loss reduction, AI plays a vital role when coupled with machine learning algorithms. By analyzing extensive volumes of data encompassing weather patterns, soil conditions, and occurrences of pests and diseases, AI systems can provide farmers with real‐time insights and predictive models. This allows for proactive decision‐making regarding optimal timing for pesticide application, irrigation scheduling, and disease detection. Consequently, farmers can adopt preventive measures, minimizing losses and optimizing resource utilization. For instance, Ji et al. in 2007 developed an artificial neural network (ANN)‐based system for rice yield prediction in Fujian, China, improving accuracy over traditional models. Moreover, AI‐powered imaging techniques, such as computer vision, enable the early detection of diseases, pests, and nutrient deficiencies in plants. Early detection empowers farmers to take prompt action, averting the further spread of diseases and minimizing crop losses. Tobal and Mokthar in 2014 pioneered an AI‐assisted image processing method for weed identification, introducing an evolutionary ANN to optimize neural parameters using a genetic algorithm. However, the implementation of these technologies face challenges such as high initial costs, the need for technical expertise, and the integration of various data sources. Additionally, small‐scale farmers may find it difficult to adopt these technologies due to financial and infrastructural constraints. By harnessing the potential of robotics and AI, the horticulture sector can overcome challenges related to crop losses caused by pests, diseases, adverse weather conditions, and inefficient farming practices. These technological applications offer a pathway to enhanced productivity, reduced losses, and greater sustainability in horticulture. As we move forward, it is imperative to continue advancing and integrating these technologies, fostering innovation and collaboration between technology developers and the farming community. [ABSTRACT FROM AUTHOR]

This study investigates the influence of pruning on knot development and knot-free timber ratio in Chinese fir (Cunninghamia lanceolata). A 9-year-old Chinese fir plantation served as the research site, employing a randomized block design to establish pruning and non-pruning test treatments. Over nine years, the effects of pruning on knot development and knot-free timber ratio were examined. Standard wood methods were applied to sample nine Chinese fir trees subject to pruning and nine without pruning. Various techniques including sawing board, rotary cutting, and transverse and longitudinal section methods were employed to process the samples, with subsequent measurement of relevant indices. Analysis was conducted to ascertain differences in knot development and knot-free timber ratio between pruned and non-pruned trees. Results indicate a significant reduction in knot volume following pruning, with pruned Chinese fir demonstrating 48.70% and 34.44% lower knot volumes compared to non-pruned counterparts using the sawing board and rotary cutting methods, respectively. Furthermore, the impact of pruning on knot volume varied across different height sections, with the 0-2 m section experiencing the greatest reduction. Pruning also resulted in a notable increase in knot-free timber ratio within Chinese fir plantations, with pruned trees exhibiting 0.23% and 0.26% higher ratios compared to non-pruned trees using the sawing board and rotary cutting methods, respectively. Variations were observed in the determination of pruning effect indices using different methods, with the sawing board method emerging as a practical and effective approach for characterizing pruning effects in plantation settings. [ABSTRACT FROM AUTHOR]

Agriculture, which is spreading rapidly, is one of the major effectors on biodiversity - generally contributing to its decline. In the past few decades, most research efforts have focused on the impact of industrial agriculture on the environment and biodiversity. However, less attention has been paid on examining the impact of sustainable agricultural management practices on biodiversity. Challenges include the disruptive nature of some practices (e.g., agrochemical application) and the timing of others (e.g., tree pruning). Here, we highlight the value of passive acoustic monitoring in assessing the impact of agricultural management practices on biodiversity, using birds as indicators. We outline key considerations, including bird ecology and behavior, ARU sampling protocols, and data management. To demonstrate our approach, we present a case study from a coffee landscape in Costa Rica, where we analyzed the effects of pruning and pesticide application over two years. By focusing on selected focal species and using a subsample of the total hours recorded in combination with a mobile app for annotations, we found that pruning negatively impacted most species, while pesticide application adversely affected all species studied. Our methodology leverages technology to evaluate the impacts of agricultural management practices, offering insights to guide and assess sustainable agricultural strategies aimed at balancing biodiversity conservation with human well-being. [ABSTRACT FROM AUTHOR]

With the continuous development of green and high‐quality aquaculture technology, the process of industrialized aquaculture has been promoted. Automation, intelligence, and precision have become the future development trend of the aquaculture industry. Fish individual recognition can further distinguish fish individuals based on the determination of fish categories, providing basic support for fish disease analysis, bait feeding, and precision aquaculture. However, the high similarity of fish individuals and the complexity of the underwater environment presents great challenges to fish individual recognition. To address these problems, we propose a novel fish individual recognition method for precision farming that rethinks the knowledge distillation strategy and the chunking method in the vision transformer. The method uses the traditional convolutional neural network model as the teacher model, introducing the teacher token to guide the student model to learn the fish texture features. We propose stride patch embedding to expand the range of the receptive field, thus enhancing the local continuity of the image, and self‐attention‐pruning to discard unimportant tokens and reduce the model computation. The experimental results on the DlouFish dataset show that the proposed method in this paper improves accuracy by 3.25% compared to ECA Resnet152, with an accuracy of 93.19%, and also outperforms other vision transformer models. [ABSTRACT FROM AUTHOR]

Abstract:Due to their superior performance, deep neural networks have been widely applied in fields such as image recognition and object detection. However, they contain a large number of parameters and require immense computational power, posing challenges for deployment on mobile edge devices that require low latency and low power consumption. To address this issue, a compression algorithm that replaces multiplication operations with bit-shifting and addition is proposed. This algorithm compresses neural network parameters to low bit-widths through pruning and quantization. This algorithm reduces the hardware deployment difficulty under limited multiplication resources, meets the requirements of low latency and low power consumption on mobile edge devices, and improves operational efficiency. Experiments conducted on classical neural networks with the ImageNet dataset revealed that when the neural network parameters were compressed to 4 bits, the accuracy remained essentially unchanged compared to the full-precision neural network. Furthermore, for ResNet18, ResNet50, and GoogleNet, the Top-1/Top-5 accuracies even improved by 0.38%/0.22%, 0.35%/0.21%, and 1.14%/0.57%, respectively. When testing the eighth convolutional layer of VGG16 deployed on Zynq7035, the results showed that the compressed network reduced the inference time by 51.1% and power consumption by 46.7%, while using 43% fewer DSP resources. [ABSTRACT FROM AUTHOR]

The high performance of deep learning networks relies on large datasets and powerful computational resources. However, collecting enough diseased training samples is a daunting challenge. In addition, existing few-shot learning models tend to suffer from large size, which makes their deployment on edge devices difficult. To address these issues, this study proposes a pruning-based lightweight few-shot learning (Prune-FSL) approach, which aims to utilize a very small number of labeled samples to identify unknown classes of crop diseases and achieve lightweighting of the model. First, the disease few-shot learning model was built through a metric-based meta-learning framework to address the problem of sample scarcity. Second, a slimming pruning method was used to trim the network channels by the γ coefficients of the BN layer to achieve efficient network compression. Finally, a meta-learning pruning strategy was designed to enhance the generalization ability of the model. The experimental results show that with 80% parameter reduction, the Prune-FSL method reduces the Macs computation from 3.52 G to 0.14 G, and the model achieved an accuracy of 77.97% and 90.70% in 5-way 1-shot and 5-way 5-shot, respectively. The performance of the pruned model was also compared with other representative lightweight models, yielding a result that outperforms those of five mainstream lightweight networks, such as Shufflenet. It also achieves 18-year model performance with one-fifth the number of parameters. In addition, this study demonstrated that pruning after sparse pre-training was superior to the strategy of pruning after meta-learning, and this advantage becomes more significant as the network parameters are reduced. In addition, the experiments also showed that the performance of the model decreases as the number of ways increases and increases as the number of shots increases. Overall, this study presents a few-shot learning method for crop disease recognition for edge devices. The method not only has a lower number of parameters and higher performance but also outperforms existing related studies. It provides a feasible technical route for future small-sample disease recognition under edge device conditions. [ABSTRACT FROM AUTHOR]

To address the issues of low detection accuracy and large model parameters in crop pest detection in natural scenes, this study improves the deep learning object detection model and proposes a lightweight and accurate method RTMDet++ for crop pest detection. First, the real-time object detection network RTMDet is utilized to design the pest detection model. Then, the backbone and neck structures are pruned to reduce the number of parameters and computation. Subsequently, a shortcut connection module is added to the classification and regression branches, respectively, to enhance its feature learning capability, thereby improving its accuracy. Experimental results show that, compared to the original model RTMDet, the improved model RTMDet++ reduces the number of parameters by 15.5%, the computation by 25.0%, and improves the mean average precision by 0.3% on the crop pest dataset IP102. The improved model RTMDet++ achieves a mAP of 94.1%, a precision of 92.5%, and a recall of 92.7% with 4.117M parameters and 3.130G computations, outperforming other object detection methods. The proposed model RTMDet++ achieves higher performance with fewer parameters and computations, which can be applied to crop pest detection in practice and aids in pest control research. [ABSTRACT FROM AUTHOR]

Musculoskeletal Disorders (MSDs) stand as a prominent cause of injuries in modern agriculture. Scientific research has highlighted a causal link between MSDs and awkward working postures. Several methods for the evaluation of working postures, and related risks, have been developed such as the Rapid Upper Limb Assessment (RULA). Nevertheless, these methods are generally applied with manual measurements on pictures or videos. As a consequence, their applicability could be scarce, and their effectiveness could be limited. The use of wearable sensors to collect kinetic data could facilitate the use of these methods for risk assessment. Nevertheless, the existing system may not be usable in the agricultural and vine sectors because of its cost, robustness and versatility to the various anthropometric characteristics of workers. The aim of this study was to develop a technology capable of collecting accurate data about uncomfortable postures and repetitive movements typical of vine workers. Specific objectives of the project were the development of a low-cost, robust, and wearable device, which could measure data about wrist angles and workers' hand positions during possible viticultural operations. Furthermore, the project was meant to test its use to evaluate incongruous postures and repetitive movements of workers' hand positions during pruning operations in vineyard. The developed sensor had 3-axis accelerometers and a gyroscope, and it could monitor the positions of the hand–wrist–forearm musculoskeletal system when moving. When such a sensor was applied to the study of a real case, such as the pruning of a vines, it permitted the evaluation of a simulated sequence of pruning and the quantification of the levels of risk induced by this type of agricultural activity. [ABSTRACT FROM AUTHOR]

Real-time seed detection on resource-constrained embedded devices is essential for the agriculture industry and crop yield. However, traditional seed variety detection methods either suffer from low accuracy or cannot directly run on embedded devices with desirable real-time performance. In this paper, we focus on the detection of rapeseed varieties and design a dual-dimensional (spatial and channel) pruning method to lighten the YOLOv7 (a popular object detection model based on deep learning). We design experiments to prove the effectiveness of the spatial dimension pruning strategy. And after evaluating three different channel pruning methods, we select the custom ratio layer-by-layer pruning, which offers the best performance for the model. The results show that using custom ratio layer-by-layer pruning can achieve the best model performance. Compared to the YOLOv7 model, this approach results in mAP increasing from 96.68% to 96.89%, the number of parameters reducing from 36.5 M to 9.19 M, and the inference time per image on the Raspberry Pi 4B reducing from 4.48 s to 1.18 s. Overall, our model is suitable for deployment on embedded devices and can perform real-time detection tasks accurately and efficiently in various application scenarios. [ABSTRACT FROM AUTHOR]

Energy use analysis and greenhouse gas (GHG) emissions are among the most important aspects regarding the sustainability performance of a farming system. The aim of this study was to assess the environmental impact, in terms of energy consumption and GHG emissions in thirty vineyards located on Crete, Greece, and deliver a digital, decision support tool (DST). A simplified life cycle approach was used to collect data from the vineyards up to farm gate, located in the top wine-producing Cretan municipalities, regarding farming practices, inputs, and yield for a 2-year period. Sum energy and non-renewable energy intensity and efficiency were calculated. GHG emissions were estimated in terms of CO2 equivalents, following IPCC methodology, while the emissions intensity is also reported. Fossil fuels consumed by machinery for weed management, transportation, soil management, pest control, and synthetic fertilizers were the practices found to be accountable for the higher energy consumption. Synthetic fertilizers and fossil fuel consumption were the main sources of GHG emissions, followed by burning of the pruning residues. Omitting burning pruning residues, reducing tillage intensity, and replacing mechanical weed management are highlighted as the main practices that can improve the sustainability of viticulture on Crete. Making use of the collected data, a crop-specific DST, named "ECO2VINE", for calculating a vineyard's energy use and GHG emissions was developed, validated, and made publicly available. [ABSTRACT FROM AUTHOR]

Increasing the planting density of mango orchards appears promising for obtaining higher yields, particularly during the first productive years. However, the challenge is to maintain a good balance between vegetative growth and fruit production in the longer term. The objective of this study was to decipher the effects of planting density, training system and cultivar on young mango trees' growth and production. The experiment, conducted in North Queensland, consisted of five combinations of planting density and training system applied to the cultivars Keitt, Calypso and NMBP-1243. The planting densities were low (208 tree ha−1), medium (416 tree ha−1) and high (1250 tree ha−1). The closed vase conventional training system was applied at each density. Single leader and espalier on trellis training systems were applied at medium and high densities, respectively. The tree canopy dimensions were measured every 6 months from planting, and tree production was recorded from the third to the fifth years after planting. Vegetative growth and fruit production were the results of complex interactions between planting density, training system, cultivar and/or time. The expected increase in orchard yield with higher planting density was observed from the first productive year, despite lower individual tree production at high planting density. Lower vegetative growth and fruit production at high planting density were probably caused by competition between trees. NMBP-1243 and Keitt showed more rapid vegetative growth. Keitt was the most productive cultivar during the first three productive years. The detailed results of this study provide avenues to further explore the behaviour of mango trees at high planting densities. [ABSTRACT FROM AUTHOR]

To address the challenges of high model complexity and low accuracy in detecting small targets in insulator defect detection using UAV aerial imagery, we propose a lightweight algorithm, PAL-YOLOv8. Firstly, the baseline model, YOLOv8n, is enhanced by incorporating the PKI Block from PKINet to improve the C2f module, effectively reducing the model complexity and enhancing feature extraction capabilities. Secondly, Adown from YOLOv9 is employed in the backbone and neck for downsampling, which retains more feature information while reducing the feature map size, thus improving the detection accuracy. Additionally, Focaler-SIoU is used as the bounding-box regression loss function to improve model performance by focusing on different regression samples. Finally, pruning is applied to the improved model to further reduce its size. The experimental results show that PAL-YOLOv8 achieves an mAP50 of 95.0%, which represents increases of 5.5% and 2.6% over YOLOv8n and YOLOv9t, respectively. Furthermore, GFLOPs is only 3.9, the model size is just 2.7 MB, and the parameter count is only 1.24 × 106. [ABSTRACT FROM AUTHOR]

Swin Transformer, as a layered visual transformer with shifted windows, has attracted extensive attention in the field of computer vision due to its exceptional modeling capabilities. However, its high computational complexity limits its applicability on devices with constrained computational resources. To address this issue, a pruning compression method was proposed, integrating weight sharing and distillation. Initially, weight sharing was implemented across layers, and transformation layers were added to introduce weight transformation, thereby enhancing diversity. Subsequently, a parameter dependency mapping graph for the transformation blocks was constructed and analyzed, and a grouping matrix F was built to record the dependency relationships among all parameters and identify parameters for simultaneous pruning. Finally, distillation was then employed to restore the model's performance. Experiments conducted on the ImageNet-Tiny-200 public dataset demonstrate that, with a reduction of 32% in model computational complexity, the proposed method only results in approximately a 3% performance degradation at minimum. It provides a solution for deploying high-performance artificial intelligence models in environments with limited computational resources. [ABSTRACT FROM AUTHOR]

The study examined the effect of plant spacing on yield productivity and assessed yield after pruning for jatropha. The experimental design was split-plot in a randomized complete block, including two main-plots of planting space comprised of 1 × 1.5 m (S1) and 2 × 1.5 m (S2), and 14 sub-plots were jatropha varieties of 11 interspecific hybrids crossed between jatropha and peregrina, and three jatropha varieties with four replications. The data collection was done continuously for two years and found that planting space was not significantly different in traits. In year one, S1 showed a higher value of fruit yield ha− 1 and seed yield ha− 1 than S2, while in year two, S2 displayed a higher mean, fruit weight plant− 1, and seed weight plant− 1 than S1. The inter-specific hybrid jatrophas, KUJL70 and KUJL106 gave more seed yield ha− 1 and oil content than other jatropha breeding lines. [ABSTRACT FROM AUTHOR]

Copyright of Arab Journal of Plant Protection is the property of Arab Society for Plant Protection and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

To address the challenges of efficient hardware design for error-tolerant applications, several techniques of applied approximate computing have been proposed. Pruning algorithms aim to approximate circuits with reduced design requirements at the cost of an acceptable degradation of their quality of result. In this letter, we present the effects of resynthesis, an iterative application of logic synthesis along with pruning algorithms, into a state-of-the-art approximate design flow, AxLS. Resynthesis strategy improves the approximation, achieving up to 70% area-power savings for the same error in the output, and reducing the number of iterations, and hence the time required to explore the design space in up to $30\times $ , to obtain an approximated design. [ABSTRACT FROM AUTHOR]

Despite all the recent development and success of deep neural networks, deployment of a deep model onto the resource-constrained devices still remains challenging. However, model pruning can resolve this issue for Convolutional Neural Networks (CNNs), since it is one of the most popular approaches to reducing computational complexities. Therefore, this article presents a pruning model for convolutional neural networks. The proposed method classifies and arranges similar filters into the same cluster where the similarity is calculated using a three-dimensional normalized cross-correlation. Moreover, these steps can be completed entirely based on the filter values while not requiring a set of test images as well as the acquisition of any filter activation. In the research, the performances of the proposed model pruning method have been evaluated, where it is observed that the proposed approach is computationally light and requires significantly less time and resources compared to ML and activation-based approaches. In the experiments, using the VGG16 model on the Cifar10 dataset, the proposed approach results in the pruned model(s) which are comparable in performance with models found using activation-based methods and expensive ML-based methods. Similar results are found when pruning a custom CNN on the MNIST and Fashion MNIST datasets as well.

Purpose: selection of ways for grape bush forming for the conditions of the Crimean Peninsula based on the analysis of their use in similar natural and climatic conditions. Discussion. Obtaining high yields of table grape varieties of appropriate quality is possible when meeting all requirements of this crop: optimal nutrition and irrigation regime, high level of agrotechnical measures, selection of planting patterns and methods for a grape bush forming. The overview of scientific and research papers of domestic and foreign scientists on the selection of the optimal number of buds, shoots, fruit vines on a grape bush, ways of forming a grape bush taking into account the climatic conditions of the region of grape cultivation, its varietal characteristics and cultivation technology to obtain a stable high yield of proper quality is provided. The growth and development of grape leaf blades and the occurrence of various physiological and biochemical processes in them were also considered. The volume of annual growth of fruit shoots, the coefficient of fruitfulness of shoots were studied and the various systems of bush management for growing grapes in different weather conditions were considered. Conclusions. The analysis of the results of studies conducted in different years in different agrobiological conditions showed that for grape growing in the conditions of the Crimean Peninsula, the optimal planting scheme is 3.0 × 2.0 m with pruning the length of fruit vines to four buds and a shoot load of 50 thousand pcs./ha. The issues of selecting irrigation modes and methods, which are important factors in obtaining high and stable yields, have been studied little. Therefore, conducting research devoted to studying the effect of irrigation with various methods of forming a grape bush under the conditions of the Crimean Peninsula is relevant.

Summary. Grapevine yellows bois noir (BN) and the grapevine trunk disease esca complex (EC) cause serious yield losses in European vineyards and are often widespread in the same vineyard. In a Chardonnay vineyard in north-eastern Italy, evolution of the two diseases from 2007 to 2020 was compared and their possible interaction was investigated. Evolution of symptomatic grapevines over the 16 years was very different between the two diseases, with a substantial linear increase for BN and an exponential increase for EC. The BN increase from one year to another was associated with the abundance of Hyalesthes obsoletus, the BN-phytoplasma vector, whereas the exponential increase in EC was likely due to the amount of inoculum and the increased size of pruning cuts over time. The courses of the two diseases were also very different, with a much greater occurrence of dead grapevines from EC than from BN. Some grapevines showed symptoms of both diseases, but the probability was less that a grapevine symptomatic for BN or EC showed symptoms of the other disease. Examinations of the spatial distribution of the two diseases showed dissociation between them. Data indicated that mechanisms of induced defense were involved in the lower probability that a grapevine affected by one showed symptoms of the other. [ABSTRACT FROM AUTHOR]

High-complexity network models are challenging to execute on agricultural robots with limited computing capabilities in a large-scale pineapple planting environment in real time. Traditional module replacement often struggles to reduce model complexity while maintaining stable network accuracy effectively. This paper investigates a pineapple detection framework with a YOLOv7-tiny model improved via pruning and a lightweight backbone sub-network (the RGDP-YOLOv7-tiny model). The ReXNet network is designed to significantly reduce the number of parameters in the YOLOv7-tiny backbone network layer during the group-level pruning process. Meanwhile, to enhance the efficacy of the lightweight network, a GSConv network has been developed and integrated into the neck network, to further diminish the number of parameters. In addition, the detection network incorporates a decoupled head network aimed at separating the tasks of classification and localization, which can enhance the model's convergence speed. The experimental results indicate that the network before pruning optimization achieved an improvement of 3.0% and 2.2%, in terms of mean average precision and F1 score, respectively. After pruning optimization, the RGDP-YOLOv7-tiny network was compressed to just 2.27 M in parameter count, 4.5 × 10 9 in computational complexity, and 5.0MB in model size, which were 37.8%, 34.1%, and 40.7% of the original YOLOv7-tiny network, respectively. Concurrently, the mean average precision and F1 score reached 87.9% and 87.4%, respectively, with increases of 0.8% and 1.3%. Ultimately, the model's generalization performance was validated through heatmap visualization experiments. Overall, the proposed pineapple object detection framework can effectively enhance detection accuracy. In a large-scale fruit cultivation environment, especially under the constraints of hardware limitations and limited computational power in the real-time detection processes of agricultural robots, it facilitates the practical application of artificial intelligence algorithms in agricultural engineering. [ABSTRACT FROM AUTHOR]

In the Vase system, the most common training system for peach-growing countries for more than a century, light distribution to the canopy is uneven, and access to the canopy for pruning, thinning, and harvest labor is difficult. It is important to identify alternative systems to the Vase system considering the cultivar and growing environment to facilitate labor and enhance productivity and quality. In Türkiye, one of the important centers of peach growing worldwide, detailed research has yet to be published on the applicability of training systems alternative to the widely used Vase system. Therefore, this study aimed to evaluate the effect of different training systems (Vase, Catalan Vase, Quad-V, Tri-V) on growth, yield, fruit quality, and labor costs of peach cultivars (ExtremeVR 314, ExtremeVR 436, ExtremeVR 568). The experiment was conducted from 2017 to 2022. Although the distance between rows in all training systems is 5 m, the distance between trees on the row is determined as 4 m in Vase, 3 m in Catalan Vase, 2.5 m in Quad-V, and 2 m in Tri-V. In the experiment, vegetative development parameters, such as canopy volume, trunk sectional area, and the amount of winter pruning weights, differed according to the training system. In the final year, the Vase system, which produces the most pruning weight, generates 48.0% more pruning weight compared with the Tri-V system, which produces the least. Concerning yield per tree and hectare, trained to the Vase system yielded higher fruit per tree regardless of cultivar, while the Quad-V and Tri-V systems yielded more fruit per hectare. The training system and cultivar affected the fruit size; the largest fruits were obtained from the ExtremeVR 568 cultivar trained according to the Vase system. The most time needed for winter pruning was obtained from the Vase (79.4 min/tree) system, and the Tri-V (57.4 min/tree) and Quad-V (60.3 min/tree) systems required the least time. The Catalan Vase (31.1 min/tree) system required the least time for summer pruning. The most fruit harvest in an hour was obtained from the trees trained according to the Tri-V (164.5 kg/h) and Quad-V (132.02 kg/h) systems. These results suggest that Quad-V and Catalan Vase systems performed well and could be alternatives to the Vase system. [ABSTRACT FROM AUTHOR]

Health and quality of the root system are imperative to ensure the successful establishment of a citrus tree after transplant from the nursery into the field. Containerized citrus production in enclosed nurseries restricts root growth and can result in root circling and intertwining. This may hinder root expansion and result in root girdling after transplant, negatively affecting tree establishment and growth. The root structure of a transplanted citrus tree can also be affected by the container type used in the nursery. Containers with root-pruning properties like chemical pruning or air pruning reduce root circling and may produce superior root systems compared with regular, nonpruning containers. The aim of this study was to evaluate the effects of different nursery containers on root physiological and morphological traits and plant performance over 15 months of growth in the nursery. Three container types, chemical pruning containers, air-pruning containers, and standard nursery containers, were compared. The chemical pruning containers were standard citrus nursery containers with a mixture of copy hydroxide [Cu(OH)2] and copper carbonate (CuCO3) [10% copper (Cu)] applied to the inner wall. Pruning occurs upon contact of the roots with the Cu on the wall of the containers. The air-pruning containers were customsized Air-Pots in which pruning occurs on holes in the wall of the containers upon contact of the roots with the air. Two rootstocks, US-812 and US-942 (Citrus reticulata × Poncirus trifoliata), were included for comparison in the nongrafted stage and 12 months after grafting with 'Valencia' orange (Citrus sinensis). Chemical root pruning positively influenced tree height, shoot mass, leaf area, rootstock trunk diameter, and the nonfibrous root biomass. No differences among container types were observed for the fibrous root biomass, but chemical pruning produced more roots that were finer with a higher specific root length and a higher respiration rate. In contrast, air pruning produced more roots that were thicker compared with the other containers. Most of the leaf nutrients were lower in trees grown in the chemical pruning containers compared with the standard containers, except for Cu and zinc (Zn), which were highest in the former. Trees growing in air-pruning containers were not significantly different in growth from trees growing in standard containers. [ABSTRACT FROM AUTHOR]

This study introduces a novel pruning approach based on explainable graph convolutional networks, strategically amalgamating pruning and quantisation, aimed to tackle the complexities associated with existing machine learning and deep learning models for stress detection using ultra-short heart rate variability analysis. These complexities often impede the implementation ability of such models on resource-limited devices. The proposed method exhibits exceptional performance, demonstrating high accuracy (97.75%) and efficiency (97.66%) on the WESAD dataset, along with an impressive accuracy (94.48%) and efficiency (94.39%) on the SWELL dataset. Importantly, the runtime complexity saw a significant reduction, down by 63.4% and 69.34% compared to the original model. The proposed method's notable advantage lies in its ability to retain nearly all of the initial model's performance with negligible loss, even when the pruning levels are below 60%. This innovative approach, thus, offers a promising solution for effective stress detection, specifically designed to operate smoothly on devices with limited resources. [ABSTRACT FROM AUTHOR]

Nutrient management in coffee-based agroforestry systems plays a critical role in soil nitrogen (N) cycling, but has not been well documented. The objective of this study was to evaluate the effect of coffee canopy management and fertilization on soil N dynamics. This study used a randomized complete block design (2 × 3 × 2) with four replications. There were three factors: 1) coffee canopy management (T1: Pruned, T2: Unpruned), 2) fertilizer type (O: Organic, I: Inorganic; M: 50% Organic + 50% Inorganic), and 3) fertilizer dose (D1: low, D2: medium, D3: high). Soil N dynamic indicators (i.e., total N, ammonium (NH4+), nitrate (NO3−), net N-NH4+, net N-NO3−, soil microbial biomass N) were measured at two soil sampling depths (0–20 cm and 20–40 cm). Results showed that pruning increased soil total N and microbial biomass N (MBN) by 10–56% relative to unpruned coffee trees. In contrast, the unpruned coffee canopy had 15–345% higher NH4+, NO3−, net N-NH4+, net N-NO3−, and microbial biomass N concentration than pruned coffee. Mixed fertilizer application increased NO3− and net N-NH4+ accumulation by 5–15% relative to inorganic and organic fertilizers. In addition, medium to high dose fertilization led to a 19–86% higher net N-NO3− concentration and microbial biomass N as compared to low dose fertilization. The treatment of no pruning and mixed fertilizer at low to medium doses was the optimal management strategy to maintain soil available N, while pruning combined with organic fertilizer has the potential to improve soil total N and MBN. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The hedgerow growing system is prevalent in new olive orchards worldwide due to its fully mechanized harvesting. Several works have been published to compare cultivars planted in this system, focusing on productivity and oil composition. However, little research has been conducted on the long-term evaluation of cultivars' growth habits when trained in hedgerow systems and on how it affects their interannual productivity. In this work, we report the canopy growth habit, productivity, and their correlation for the 'Arbequina', 'Arbosana', 'Koroneiki', 'Lecciana', 'Oliana', and 'Sikitita' cultivars grown in a hedgerow system in Extremadura, central-western Spain, for 9 years. 'Koroneiki', 'Arbequina', and 'Lecciana' were the cultivars with the highest canopy growth, both in young and adult trees, and the ones with the highest pruning needs from 5 to 10 years after planting. The yield behavior in each of the years evaluated was stable in all cultivars except 'Lecciana'. This alternate bearing was associated with the distribution of total yearly produced biomass between fruits and vegetative growth. 'Oliana', 'Arbosana', and 'Sikitita' were the cultivars with the highest proportion of fruit of the total biomass, and 'Lecciana' showed the lowest. This study indicates that cultivars with higher fruit proportions of total biomass might have better suitability for long-term growing in hedgerow formation, fewer pruning needs, and more stable productivity across the years. In this sense, in the climatic conditions considered here, 'Arbosana', 'Sikitita', and 'Oliana' could be the most suitable cultivars for this growing system. [ABSTRACT FROM AUTHOR]

The COVID-19 pandemic has had a significant impact on students' academic performance. The effects of the pandemic have varied among students, but some general trends have emerged. One of the primary challenges for students during the pandemic has been the disruption of their study habits. Students getting used to online learning routines might find it even more challenging to perform well in face to face learning. Therefore, assessing various potential risk factors associated with students low performance and its prediction is important for early intervention. As students' performance data encompass diverse behaviors, standard machine learning methods find it hard to get useful insights for beneficial practical decision making and early interventions. Therefore, this research explores regularized ensemble learning methods for effectively analyzing students' performance data and reaching valid conclusions. To this end, three pruning strategies are implemented for the random forest method. These methods are based on out-of-bag sampling, sub-sampling and sub-bagging. The pruning strategies discard trees that are adversely affected by the unusual patterns in the students data forming forests of accurate and diverse trees. The methods are illustrated on an example data collected from university students currently studying on campus in a face-to-face modality, who studied during the COVID-19 pandemic through online learning. The suggested methods outperform all the other methods considered in this paper for predicting students at the risk of academic failure. Moreover, various factors such as class attendance, students interaction, internet connectivity, pre-requisite course(s) during the restrictions, etc., are identified as the most significant features. [ABSTRACT FROM AUTHOR]

The objective of this study was to evaluate the production and quality of 'Delite' blueberry fruit under different pruning times, during the 2021/2022 and 2022/2023 production cycles, in the Pelotas region, RS, Brazil. The experimental design was in randomized blocks, in a 2 x 3 factorial scheme, the factors being: summer pruning (with summer pruning x without summer pruning) and winter pruning times (early x conventional x late), with four replications and one plant per replication. The following parameters were evaluated: soluble solids content, pH, titratable acidity, number of fruits per plant, average fruit weight, fruit diameter, fruit length, production per plant, yield and color. 'Delite' blueberry plants that were not pruned in the summer showed the highest production per plant, regardless of the time of winter pruning. In addition, the early and late winter pruning periods provide greater fruit weight for 'Delite' blueberries under the conditions of the experiment and different pruning methods do not influence the soluble solids, pH and ohue content of the fruit. It can therefore be concluded that 'Delite' blueberry plants that were not pruned in the summer and pruned between June and August in the winter showed higher fruit production and quality. [ABSTRACT FROM AUTHOR]