Smart agricultural technology最新文献

{"title":"An optimization framework for intelligent irrigation system installation in fragmented paddy fields","authors":"Runze Tian ,&nbsp;Kyoji Takaki ,&nbsp;Toshiaki Iida ,&nbsp;Keigo Noda ,&nbsp;Yohei Asada","doi":"10.1016/j.atech.2025.101740","DOIUrl":"10.1016/j.atech.2025.101740","url":null,"abstract":"<div><div>The fragmented structure of Japanese paddy fields increases labor requirements for patrol and irrigation management. While Intelligent Irrigation System (IIS) units can effectively reduce labor input, their benefits are influenced by the location of the installations. Consequently, determining optimal field combinations under varying installation conditions (varying numbers of IIS units and farmer datasets) has become a critical issue. This study proposes and validates a two-stage optimization framework for IIS unit installation that employs patrol-route distance reduction as the evaluation metric. In the first stage, Density-based Spatial Clustering of Applications with Noise (DBSCAN) with the Normalized Nearest-distance (NN-distance) method was applied to mitigate search space explosion under non-uniform densities. In the second stage, the 2-opt algorithm was used to optimize patrol routes and quantify labor reduction. Validation results showed that the framework compressed the candidate solution space and alleviated the computational complexity associated with the Non-deterministic Polynomial-time hard (NP-hard) nature of the problem. Furthermore, the NN-distance method maintained solution quality and outperformed the conventional k-distance approach by mitigating over-clustering and over-segmentation under non-uniform spatial distributions. Case analyses revealed that the benefits of IIS unit installation depend not only on the number of installed units but also strongly on the spatial distribution of fields. Overall, the proposed framework enhances the applicability of DBSCAN to non-uniform spatial data, provides guidance for differentiated installation strategies, and offers a reproducible methodological framework for deploying smart agricultural technologies in fragmented agricultural systems.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101740"},"PeriodicalIF":5.7,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925408","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Branch length recognition and pruning point localization method for greenhouse tomatoes based on improved YOLOv8","authors":"Kai Qiu ,&nbsp;Wenxian Li ,&nbsp;Guoying Shi ,&nbsp;Xianwei Hou ,&nbsp;Guanshan Zhang ,&nbsp;Yang Li ,&nbsp;Tianhua Li","doi":"10.1016/j.atech.2025.101739","DOIUrl":"10.1016/j.atech.2025.101739","url":null,"abstract":"<div><div>To achieve precise identification, length estimation, and pruning point localization for robotic tomato pruning, this study presents a method for branch length recognition and pruning point localization based on an improved YOLOv8 model. The proposed YOLOv8n-CE integrates the CBAM attention mechanism to enhance the model’s focus on critical branch features, and replaces the original loss with the EIOU loss to improve bounding box regression accuracy and convergence speed. Instance segmentation is subsequently performed within the detection boxes to obtain the main stem mask and extract the upper and lower endpoints of the stem. The pixel coordinates of these endpoints are transformed into three-dimensional camera coordinates to compute the stem length, and the actual branch length is derived by combining the pixel height ratio of the bounding box with the segmentation mask. The pruning point is localized along the central line of the main stem, 1.5 cm above the lower endpoint, ensuring reduced pathogen intrusion and faster wound healing. Experimental results demonstrate that the YOLOv8n-CE model improves mAP₅₀ by 1.7 percentage points compared with YOLOv8n. For branch length measurement, the method achieves an R² of 0.929, an MAE of 0.411 cm, and an RMSE of 0.494 cm. The pruning point localization success rate reaches 92%, with a mean absolute error of 0.247 cm. These results verify that the proposed approach meets the accuracy requirements for tomato branch measurement and pruning point localization, providing a reliable theoretical and technical foundation for robotic pruning applications.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101739"},"PeriodicalIF":5.7,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Short-term forecasting and predictive control of rooftop greenhouse microclimate using multi-horizon machine learning models","authors":"Joaquim Cebolla-Alemany ,&nbsp;Yunyao Cheng ,&nbsp;Laia Pintó-Espín ,&nbsp;Michele Albano ,&nbsp;Marcel Macarulla ,&nbsp;Santiago Gassó-Domingo","doi":"10.1016/j.atech.2025.101733","DOIUrl":"10.1016/j.atech.2025.101733","url":null,"abstract":"<div><div>This study presents a data-driven forecasting and control framework tailored to rooftop smart greenhouses integrated into buildings for urban agriculture—a context rarely addressed in existing literature. By combining internal environmental data, external meteorological inputs, and actuator operation states, the framework enables short-term temperature forecasting with high temporal resolution (5, 10, and 15-minute horizons). Advanced feature engineering techniques—including lag variables, rolling statistics, and derived indicators—were applied to capture complex greenhouse dynamics. Seven regression-based machine learning models (namely, Decision Tree, Random Forest, Gradient Boosting, XGBoost, LightGBM, Support Vector Regression, and Multi-layer Perceptron) were trained and systematically compared using cross-validation and SHAP-based interpretability. The best-performing model for each horizon was selected and integrated into a threshold-based and a fuzzy logic-based predictive control system. Results from real-world rooftop greenhouse data show robust forecasting performance (R² &gt; 0.98, MAE between 0.280 and 0.311, and RMSE between 0.638 and 0.728 for the best performing model across all horizons) and demonstrate that the fuzzy controller achieved over 60% energy savings compared to traditional threshold-based strategies, while maintaining climate stability. This work highlights the feasibility of deploying data-driven MPC strategies in building-integrated greenhouse environments and their compatibility with digital twin ecosystems. It also identifies key challenges for generalization, including dataset size, system configuration, and geographical variability.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101733"},"PeriodicalIF":5.7,"publicationDate":"2025-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925490","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A grain flow detection method based on an improved YOLOv8n model and dual-line counting with a multi-criterion discrimination mechanism","authors":"Jie Shi,&nbsp;Xinrui Zhang,&nbsp;Yantao Zhao,&nbsp;Zhi Li,&nbsp;Linlin Yang,&nbsp;Wencai Yang","doi":"10.1016/j.atech.2025.101728","DOIUrl":"10.1016/j.atech.2025.101728","url":null,"abstract":"<div><div>To support non-contact grain yield monitoring in smart agriculture, this study presents a deep learning-based method for grain flow detection. The system includes two core components: an enhanced YOLOv8n detection model and a dual-line counting algorithm guided by multiple decision rules. In real harvesting conditions, challenges such as occlusion, fast grain movement, poor lighting, and small target sizes often lead to missed detections when using standard YOLO models. To improve performance, the YOLOv8n structure is modified in three ways. A minimum point distance IoU loss (MPDIoU) is applied to better distinguish overlapping grains. A spatial-channel shared attention (SCSA) module is added to help the model focus on weak or small grain targets. A lightweight Spatial and Channel reconstruction Convolution (SCConv) module is inserted into the network to reduce computation while maintaining accuracy. As a result, the updated model reaches 90.6 % in precision, 91.4 % in recall, and 94.5 % in mAP, improving the original version by 2.7, 1.7, and 2.5 points. In the counting stage, a dual-line method is used, combining rules such as movement direction checking, two-line crossing validation, abnormal shape removal, and IoU-based filtering. The full system remains stable even when grains move quickly, supporting reliable detection under field conditions.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101728"},"PeriodicalIF":5.7,"publicationDate":"2025-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925834","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Behaviour recognition of tail and ear biting in pigs using AI-based computer vision","authors":"Qinghua Guo ,&nbsp;Clémence A.E.M. Orsini ,&nbsp;Patrick P.J.H. Langenhuizen ,&nbsp;Yue Sun ,&nbsp;Shoujun Huo ,&nbsp;Lisette E. van der Zande ,&nbsp;Inonge Reimert ,&nbsp;J. Elizabeth Bolhuis ,&nbsp;Piter Bijma ,&nbsp;Peter H.N. de With","doi":"10.1016/j.atech.2025.101713","DOIUrl":"10.1016/j.atech.2025.101713","url":null,"abstract":"<div><div>Damaging behaviours in pigs, such as tail biting and ear biting, compromise animal welfare and farm productivity. Continuous monitoring of these behaviours is essential to intervene before escalation, gain insights into underlying causes and develop breeding programs to select pigs with lower genetic propensity for such behaviours. However, manual observations are impractical at a large-scale. To address this challenge, we propose a video-based behaviour recognition model that facilitates the automated monitoring of individual pigs. Two state-of-the-art video-based methods are investigated: SlowFast and Improved Multiscale Vision Transformers (MViTv2) for recognizing tail and ear biting in pigs, by exploiting spatiotemporal domain features. Data are collected on a commercial pig farm. In total, 532 tail-biting events (63,815 frames) and 750 ear-biting events (78,132 frames) are annotated across seven pens of tail-docked pigs. Tail biting and ear biting are defined as nibbling, sucking, chewing, or biting the tail or the ear of a pen mate. The best-performing method is based on the MViTv2-S model, which enables efficient spatiotemporal modeling. The detection accuracies obtained for tail and ear biting are 72.22 % and 72.37 %, respectively. An important and novel aspect to our knowledge is that for the first time, behaviour detection is developed without a posture requirement on the biter or the victim. The conducted experiments demonstrate the feasibility of computer-vision-based models for the recognition of damaging behaviours on commercial pig farms. This study is a crucial step towards the development of an automated early-warning approach and breeding programs to reduce tail biting and ear biting.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101713"},"PeriodicalIF":5.7,"publicationDate":"2025-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790774","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Disturbance observer based adaptive sliding mode control for driving motor speed regulation in Maize Electric Fertilizer Applicator","authors":"Zhiqiang Li,&nbsp;Kun Luo,&nbsp;Liang Tao,&nbsp;Yan Zhou","doi":"10.1016/j.atech.2025.101721","DOIUrl":"10.1016/j.atech.2025.101721","url":null,"abstract":"<div><div>Electric fertilizer applicators significantly improve the uniformity, stability, and efficiency of summer maize fertilization. However, the complex soil environment in farmlands introduces uncertainties such as parameter variations, load disturbances, frictional resistance, and positioning errors, which degrade the control accuracy and robustness of the driving motor. To address these challenges, this study proposes a disturbance observer (DO)-based adaptive sliding mode control (ASMC) method. First, a control model for the soil-straw coupling system of the electric fertilizer applicator (EFA) was established, accounting for parameter variations and external load disturbances, thereby simplifying controller design. Second, a convergence rate mechanism was introduced to accelerate convergence time, ensuring the system reaches the sliding surface within a finite time, with the convergence rate being adjustable through parameter design. Additionally, a disturbance observer was designed to estimate both mismatched and matched disturbances, enabling feedforward compensation to improve tracking accuracy and reduce system chattering. Experimental results demonstrate that the proposed method achieves high control accuracy and robustness, ensuring rapid and stable state regulation for the EFA. This work provides new ideas for the design of smart agricultural machinery controllers and effectively promotes the control upgrade of agricultural electromechanical systems.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101721"},"PeriodicalIF":5.7,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790642","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Estimating wheat disease severity from high-resolution UAV multispectral imagery using deep learning","authors":"Maitiniyazi Maimaitijiang ,&nbsp;Dalitso Yabwalo ,&nbsp;Ubaid-ur-Rehman Janjua ,&nbsp;Mohammad Maruf Billah ,&nbsp;Bruce Millett ,&nbsp;Sunish K. Sehgal ,&nbsp;Shaukat Ali","doi":"10.1016/j.atech.2025.101729","DOIUrl":"10.1016/j.atech.2025.101729","url":null,"abstract":"<div><div>Bacterial Leaf Streak (BLS) and Fusarium Head Blight (FHB) are among the most damaging diseases of wheat (Triticum aestivum), with severe consequences for grain yield, quality, and ultimately food safety and security. Rapid and precise assessment of disease severity in the fields is crucial for effective field management, potential yield loss evaluation, and high-throughput phenotyping. This research examined the utility of UAV-based multispectral imagery in combination with both traditional machine learning and modern deep learning approaches to estimate wheat disease severity under field conditions. Data collection was carried out at two wheat experimental fields in South Dakota, USA, where Unmanned Aerial Vehicle (UAV) multispectral imagery was acquired in parallel with plot-level measurements of BLS and FHB severity. Spectral and textural metrics extracted from the UAV imagery served as inputs for machine/deep learning-based regression analyses. Regression models evaluated in this work comprised traditional machine learning methods Partial Least Squares Regression (PLSR), Random Forest Regression (RFR), and three deep learning architectures: Deep Neural Networks (DNN), Convolutional Neural Networks (CNN), and multi-head self-attention (MHSA)-enhanced CNN (Att-CNN). In addition, a deep transfer learning framework was tested by transferring an Att-CNN model trained on BLS to FHB severity estimation. The results showed that deep learning methods, particularly CNN-based architectures, consistently outperformed conventional machine learning approaches. Incorporation of a MHSA mechanism into the CNN architecture further enhanced performance, especially for BLS severity estimation. Att-CNN achieved the best results for both diseases, with R² = 0.83 and RRMSE = 30.55 % for BLS, and R² = 0.70 and RRMSE = 37.05 % for FHB. While estimation of FHB severity remained more challenging, transfer learning from BLS substantially improved prediction accuracy, raising R² from 0.70 to 0.79 and reducing RRMSE from 37.05 % to 31.31 %. The study highlights the considerable potential of UAV multispectral imagery, though with notable limitations, for monitoring crop diseases. This work also demonstrates the added value of attention-based deep learning and transfer learning techniques in addressing complex applications in agricultural remote sensing.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101729"},"PeriodicalIF":5.7,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925835","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effect of spraying drone flight parameters on vineyard canopy coverage and droplet deposition","authors":"Vasilis Psiroukis ,&nbsp;Aikaterini Kasimati ,&nbsp;Konstantinos Nychas ,&nbsp;Konstantinos Dagres ,&nbsp;George Papadopoulos ,&nbsp;Evangelos Anastasiou ,&nbsp;Spyros Fountas","doi":"10.1016/j.atech.2025.101730","DOIUrl":"10.1016/j.atech.2025.101730","url":null,"abstract":"<div><div>Unmanned Aerial Vehicles (UAVs) are increasingly promoted as alternatives to conventional plant protection product (PPP) spraying in vineyards, yet limited evidence exists on how flight parameters and application configurations influence spray deposition under real vineyard conditions, especially within Europe. This study evaluated the performance of a DJI Agras T16 spraying drone across eight operational configurations combining two flight altitudes (2.0 and 2.5 m AGL), two flight speeds (1.0 and 1.5 m/s), and two aircraft positioning strategies (over-row and inter-row). Spray deposition, canopy coverage, and ground losses were quantified using water-sensitive papers (WSPs) positioned at multiple canopy heights and ground locations, while meteorological conditions were monitored, following methodologies adapted from the ISO 22,866/22,522 protocols. Results showed strong interactions among altitude, speed, flight path, and wind parameters, with over-row treatments concentrating deposition in upper canopy layers and inter-row treatments producing more homogeneous profiles but higher sensitivity to wind direction. Lower altitude flights (2.0 m) combined with slower speed (1 m/s) substantially increased ground deposition regardless of UAV positioning, whereas higher altitude (2.5 m) and speed values (1.5 m/s) reduced spray losses to the ground. The increase in pump output associated with higher speed under a constant application rate is expected to produce finer droplets, which can enhance penetration but may also elevate in-field drift risk. These findings demonstrate that UAV spraying performance depends on integrated optimisation of operational settings and environmental conditions. The results provide practical guidance for improving drone-based vineyard spraying and highlight the need for updated EU regulatory frameworks tailored to UAV application characteristics.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101730"},"PeriodicalIF":5.7,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790771","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep ensemble learning for weed risk mapping: Hybrid RF–CatBoost and CNN–XGBoost algorithms for predicting Chenopodium album distribution in rapeseed fields","authors":"Emran Dastres ,&nbsp;Mohsen Edalat","doi":"10.1016/j.atech.2025.101731","DOIUrl":"10.1016/j.atech.2025.101731","url":null,"abstract":"<div><div>The escalating spread of the competitive weed <em>Chenopodium album</em> poses a substantial threat to strategic oilseed production, necessitating advanced, spatially explicit risk-assessment tools. This study develops and compares two hybrid machine-learning architectures, the RF–CatBoost stacked ensemble and the CNN–XGBoost deep ensemble, to model habitat suitability of <em>C. album</em> across the heterogeneous rapeseed-growing landscapes of Fars Province, Iran. Using 18 optimized environmental covariates combined with extensive field observations, both models delivered strong predictive performance (AUC ≥ 0.82), outperforming conventional classifiers. The RF–CatBoost hybrid demonstrated the highest accuracy (AUC = 0.84 ± 0.03; TSS = 0.68 ± 0.04; Kappa = 0.63 ± 0.05) and the most stable spatial behavior (σ = 0.069; CV = 8.1%), whereas the CNN–XGBoost model showed moderately lower accuracy (AUC = 0.82 ± 0.04; TSS = 0.65 ± 0.05) with higher spatial uncertainty (σ = 0.078). Variable-importance analyses indicated that habitat suitability is primarily governed by edaphic factors, especially Clay content, Organic Matter, and Phosphorus, alongside anthropogenic dispersal pathways including proximity to roads and rivers. The resulting Habitat Suitability Maps (HSMs) reveal substantial areas of High and Very High suitability, accounting for 22.9% of the province under the RF–CatBoost model, thus offering a robust evidence base for Precision Weed Management (PWM). Overall, the findings demonstrate the effectiveness of deep ensemble learning in producing reliable ecological risk assessments and highlight priority zones where targeted management can enhance the sustainability and economic efficiency of rapeseed production in semi-arid regions.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101731"},"PeriodicalIF":5.7,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"OysterMushNet: A deep learning-based mobile application for real-time mushroom disease detection and agronomist support","authors":"Deepesh Prakash Guragain ,&nbsp;Bijaya Shrestha ,&nbsp;Iswor Bajracharya ,&nbsp;Neelesh Sharma ,&nbsp;Dipesh Ghimire ,&nbsp;Showkat Ahmad Bhat","doi":"10.1016/j.atech.2025.101735","DOIUrl":"10.1016/j.atech.2025.101735","url":null,"abstract":"<div><div>Oyster mushroom cultivation is gaining popularity in Nepal because of its high nutritional value, low investment requirements, and favorable climatic conditions. However, frequent disease outbreaks significantly impact yield and farmer income. This study presents OysterMushNet, an end-to-end mobile application tailored for real-time disease classification, remedy recommendations, and agronomist support. A self-collected dataset of 8884 images from multiple farms was used to train and validate an ensemble deep-learning model that combined EfficientNetB0, DenseNet201, and ResNet50V2. The ensemble model achieved 98.8 % accuracy and a mean F1-score of 0.98, outperforming individual models. To enhance real-world usability, the model was optimized for mobile device deployment using quantization techniques. This technique significantly reduces the model memory footprint as well as the inference time. The application features bilingual support (English/Nepali) and real-time farmer-agronomist communication. Field testing with 50 farmers and 10 agronomists over six months demonstrated the application's high usability, with 94 % of users finding bi-directional communication essential and 90 % validating remedy recommendations. Additionally, farmers reported a 20–30 % reduction in crop loss due to early disease detection and expert intervention. These findings highlight the potential of OysterMushNet as a scalable AI-driven solution for precision agriculture in regions with limited access to agronomists. Future works aim to expand the dataset, improve model interpretability, and integrate additional features, such as market price tracking and video upload, to enhance the application's usability and overall impact.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101735"},"PeriodicalIF":5.7,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925496","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}