Search

Leaf blight, caused by bacterial and fungal infections, threatens plant health by inducing leaf discoloration and reducing crop yield. To address this, we propose a MobileNetV3-based classifier for the Jasmine plant, leveraging streamlined Convolutional Neural Networks (CNNs) to precisely identify disease stages. The design combines depthwise convolutional layers with max-pooling layers to extract superior features, focusing on vital low-level features that signal the condition. Via preprocessing methods that include data augmentation with a Conditional GAN and Particle Swarm Optimization for feature selection, the classifier achieves reliable performance. Assessment on carefully assembled datasets shows a remarkable 97% training accuracy, demonstrating its potency. Real-world trials across varying settings, including severe camera tilts and fluctuating illumination, verify the model’s durability, delivering test accuracies within the 94% to 96% range. The dataset’s bespoke structure for CNN-based classification ensures the reliability of the outcomes. Notably, the model’s compact classification, characterized by swift processing speed and minimal size, makes it a practical choice for immediate deployment. This all-encompassing strategy highlights the presented classifier’s value in managing leaf blight spot disease difficulties in cash crops.

Obtaining body mass in large-scale rearing and production environments is difficult, making non-contact assessment methods indispensable. This paper introduces an original pipeline for predicting pig body mass based on an adapted mask region convolutional neural network (mask R-CNN). The adaptation deploys ResNeSt as the backbone feature-extraction architecture, aiming to bolster the capacity to derive features from imagery. A feature pyramid network (FPN) was appended to this backbone to enable feature merging across multiple scales. Within the region proposal network (RPN), a channel attention mechanism (CAM) was installed alongside a spatial attention mechanism (SAM), enabling the dynamic combination of localized characteristics with their long-range interdependencies. This configuration captures holistic contextual knowledge, ultimately improving segmentation fidelity. The adapted architecture achieved precision (P) of 90.33%, recall (R) of 89.85%, and mean average precision (MAP) of 95.21%, thereby isolating pig areas in images with notable effectiveness. Five descriptive features drawn from the imagery were examined: back area, body length, body width, average depth, and eccentricity. Leveraging the pig depth images, five regression learning algorithms—ordinary least squares (OLS), AdaBoost, CatBoost, XGBoost, and random forest (RF)—were trained to forecast mass values. AdaBoost delivered the strongest predictive performance, evidenced by a coefficient of determination (R2) of 0.987, a mean absolute error (MAE) of 2.96 kg, a mean square error (MSE) of 12.87 kg², and a mean absolute percentage error (MAPE) of 8.45%. The experimental results confirm that machine learning models can reliably predict pig mass, providing a technical foundation for the informed management of pig production facilities.

The delineation of hazelnut plantations has the potential to assist land management and utilization strategies, thereby enabling the advancement of collaborative precision agriculture infrastructures. This investigation addresses the recognition of hazelnut cultivation by leveraging both optical and radar remote sensing. A comparative appraisal of various Machine Learning methodologies is conducted. The framework put forward harnesses multi-temporal information from the Sentinel-1 and Sentinel-2 archives, acquired over several years and processed in cloud computing environments. A data collection encompassing 62,982 annotated instances is furnished, with 16,561 cases assigned to the ‘hazelnut’ category and 46,421 to the ‘other’ category, gathered from 8 disparate geographic sectors within the province of Viterbo. Two distinct comparative assessments are performed: initially, a Nested 5-Fold Cross-Validation routine is used to fit, tune, and compare multiple Machine Learning classifiers confined to a single sector. In a follow-on trial, the classifiers were fitted on one sector and then validated against the remaining seven geographic sectors. The investigation confirms that AI-driven interpretation of Sentinel-1 and Sentinel-2 imagery is a viable solution for hazelnut cartography. The outcomes indicate that Random Forest delivers the most robust generalizability, achieving peak results in the second trial, with 96% accuracy and an F1 score of 91% for the ‘hazelnut’ group.

This research presents a comprehensive assessment of the latest YOLOv8 architecture’s suitability for object detection applications, specifically adapted for smoke and wildfire recognition, with particular attention to agricultural and environmental safeguarding. Every available YOLOv8 variant underwent initial fine-tuning on a field-specific dataset that covers diverse conditions essential for robust agricultural oversight. Performance indicators led to selecting the ‘large’ model (YOLOv8l) for deeper hyperparameter tuning. This model was subjected to meticulous hyperparameter refinement via the One Factor At a Time (OFAT) strategy, zeroing in on pivotal variables such as learning rate, batch size, weight decay, epochs, and optimizer. Knowledge gained from the OFAT phase was used to define search boundaries for the subsequent Random Search (RS). The ultimate model produced through RS displayed considerable advancement over the baseline fine-tuned version, boosting overall precision by 1.39%, recall by 1.48%, F1-score by 1.44%, [email protected] by 0.70%, and [email protected]:0.95 by 5.09%. The improved model’s performance was confirmed across a wide range of real-world photographs depicting different agricultural landscapes, affirming its robustness in smoke and fire recognition. These outcomes underscore the model’s trustworthiness and practical value in contexts vital to agricultural protection and environmental surveillance. This investigation, marking a notable leap forward in machine learning-based fire and smoke identification, lays a firm foundation for subsequent inquiries and innovations aimed at protecting agricultural areas and natural ecosystems.

The Nong Han Chaloem Phrakiat Lotus Park functions as a visitor destination and an educational hub centered on lotus plants. Nevertheless, its role as a training ground suffers from a lack of engagement and learning drive, a consequence of the outdated way lotus plant data is displayed. This investigation deployed the smart learning paradigm in the locale to amplify curiosity and motivation for education. To build a mobile application capable of displaying details for each type, convolutional neural networks (CNNs) were used to categorize lotus plant species. The boundaries of the work were defined by sorting 11 lotus plant species using a bespoke CNN model that relied on distinct methods (augmentation, dropout, and L2) and hyperparameters (dropout and the number of epochs). The goal was to secure a high-efficacy CNN model with a lower total parameter count than using three separate pre-trained CNN models (Inception V3, VGG16, and VGG19) as reference points. Model effectiveness was measured using accuracy, F1-score, precision, and recall. The outcomes showed that the CNN model integrating augmentation, dropout, and L2 regularization at a dropout rate of 0.4 and an epoch count of 30 achieved the highest test accuracy of 0.9954. The proposed superior model proved more precise than the pre-trained CNN benchmarks, with a pronounced improvement over Inception V3. In addition, the overall parameter count was cut down by a factor of about 1.80–2.19. These revelations showed that the proposed model, with its modest parameter count, achieved an acceptable level of classification precision.

Litchi, a delicate and juicy fruit with a short shelf life, has a harvest period of fewer than 14 days each year. The emergence of intelligent agriculture has spurred the integration of vision-based, automated systems for litchi collection. Nevertheless, standard techniques frequently rely on laboratory-based, high-capacity computing machinery, which raises concerns about bulk, power consumption, and real-world operability in litchi groves. To surmount these hurdles, we put forward a real-time litchi identification strategy tailored for complex settings, leveraging handheld, energy-efficient edge computing devices. To begin with, imagery from litchi orchards was compiled to strengthen data generalizability. After that, a single-stage detector, YOLOx, anchored in a convolutional neural network (CNN), was developed to pinpoint litchi fruits in captured images accurately. To ease installation on handheld, energy-miserly edge devices, channel pruning, coupled with layer pruning, was used to shrink the final model, reducing both its bulk and parameter count. In addition, knowledge distillation was applied to further polish the network. Test outcomes revealed the suggested technique achieved 97.1% compression, resulting in a tight litchi detection model occupying just 6.9 MB, while maintaining a 99 frames per second (FPS) frame rate. The technique proved 1.8 times swifter than the untreated model. Hence, our method slots effortlessly into compact, low-computation automatic harvesting apparatus, ensuring instantaneous and precise litchi spotting right inside orchard landscapes.

Remarkable progress has been made in applying deep learning models to distinguish between weed and crop species within agricultural settings. Still, a significant gap persists in the body of work on how models are trained and evaluated when environmental conditions diverge. The dominant tendency involves either dividing a single dataset distribution into training, validation, and test splits or first pooling datasets originating from varied conditions or distributions and then partitioning them. What our study brings forward is an effort to close this gap: we incorporate multiple broadleaf weed categories spanning dissimilar distributions, probe what happens when convolutional neural networks are trained exclusively on condition- or distribution-specific data, and then gauge their behaviour in entirely foreign environments across three experimental setups. By surveying an array of network designs alongside distinct learning modes (fine-tuning versus feature extraction), testing multiple architectures, adopting varied training protocols, and consolidating data, we arrived at straightforward guiding principles that ensure models retain precision and accuracy even under starkly contrasting deployment conditions. Experiment 1, confined to a single uniform environment, produced accuracies of 80%–100% across models and strategies; fine-tuning yielded higher accuracies (94%–99.9%) than feature extraction (80%–92.96%). Experiment 2 showed a steep drop-off in performance: whenever the test partition came from sources other than the training and validation sets, accuracy ranged from 25% to 60%, most often around 40%. Experiment 3, centred on merging datasets and distributions, delivered encouraging numbers — ResNet in fine-tuning mode attained a high of 99.6%, whereas InceptionV3 in feature extraction mode dropped to a low of 69.9%. These core insights underscore that fusing data from mismatched distributions, paired with fine-tuning on modern backbones like ResNet and MobileNet, markedly improves capability, unlike the comparatively modest outputs of leaner architectures such as AlexNet. Our results suggest that accepting data heterogeneity and flexible training regimens are fundamental to sharpening weed classification pipelines. The study provides a workable approach for handling diverse datasets that reflect genuine agricultural variability.

The tractor functions as a fundamental power source for farming activities. Still, the available power is frequently not fully exploited because of poor matching between the tractor and the attached implement under actual field conditions. To improve overall system productivity, it is essential to gather information on performance-related variables characterising the tractor-implement unit. In this work, we build a real-time Instrumented Tractor Performance Monitoring System (ITPMS) that leverages the Internet of Things (IoT). The hardware foundation comprises a Raspberry Pi, a GPS sensor, a proximity sensor, a rotary potentiometer, and a three-point hitch dynamometer. The rotary potentiometer records tillage depth, whereas the three-point hitch dynamometer captures draft force values. Proximity sensors are mounted on a two-wheel-drive (2WD) tractor to measure forward speed and drive-wheel slip. A dedicated web server is set up within a Google® Firebase® project, allowing storage of data streams from all sensors via the Raspberry Pi. Alongside this, a web interface and a mobile application are constructed to present the sensor-generated data in real time. Field experiments were conducted to assess and observe the performance variables of the tractor-implement unit employing the built ITPMS. The findings reveal that the system can reliably monitor the performance variables required for the tractor-implement unit. Furthermore, the system’s ability to transmit updated information to the IoT server without delay is verified. Altogether, designing and deploying this Raspberry Pi–centred IoT framework provides a trustworthy and effective approach for real-time performance monitoring of instrumented tractors.

Quantifying fruit mass and volume manually is a laborious, repetitive task that can hinder production scheduling. This investigation developed a computer vision framework to predict the volume and mass of baobab fruits from single-view photographs captured with affordable, widely accessible cameras, including smartphones. The baobab fruits were sourced from two experimental sites situated within the savanna ecological zone. Following image capture, the fruits were detected and segmented with an accuracy exceeding 97%. The segmented outputs were converted to binary format, from which two-dimensional (2D) attributes—namely the segmented area, centroid, bounding box, equivalent diameter, and major diameter—were derived. Volume estimations were generated from the 2D attributes by employing random forest, linear, polynomial, and radial support vector machine regressors. Every model achieved strong goodness-of-fit, yet the random forest regressor performed best, with an R2 of 99.8%. The linkage between mass and volume conformed to a quadratic expression (mass = 38.23 + 0.25 × volume + 4.49e−05 × volume2) and achieved an R2 of 92%. Correlations were substantiated through graphical plots and statistical evaluations, with credible intervals for the point estimates derived from the posterior distributions of their respective samples. This finding spotlights the potential to deploy artificial intelligence techniques in a less restrictive environment to support ecological studies and agricultural oversight. Commercial firms that manufacture baobab powder and seed oil ought to adopt these models to streamline production planning. Acquiring deeper insight into how climatic gradients shape the morphological properties of baobab fruits would help refine the model.

Presently, the agricultural domain faces major hurdles in efficiently monitoring livestock health. Established observation practices tend to be labor-heavy, slow, and ill-suited to the evolving requirements of contemporary farming. The rapid proliferation of IoT devices in this field has brought scalability issues and increased processing demands to the forefront, creating an urgent need for streamlined, scalable architectures. This work proposes a cloud-centric framework to elevate the standard of livestock wellness tracking. Engineered to monitor key physiological metrics such as motility, thermal state, and cardiac rhythm, the setup leverages AWS for resilient data orchestration, alongside Python for analytical computation and live-stream analytics. The framework embraces Narrow Band IoT (NB-IoT) mechanisms, fine-tuned for minimal bandwidth usage and extended transmission range, which are ideal for isolated agricultural terrains. Its scalable design enables seamless administration of a dynamic number of IoT devices, a vital feature for accommodating changes in livestock density and operational scale. Early-stage testing conducted to gauge system capability has confirmed its resilience and practical value, reflecting a cohesive integration of AWS IoT Cloud capabilities with field-level IoT devices. Beyond this, the inquiry examines the deployment of predictive algorithmic models to support anticipatory healthcare strategies in animal populations. By forecasting impending health anomalies before visible symptoms, the mechanism stands to advance animal well-being and farm output substantially. The confluence of cloud computing power and IoT infrastructure not only meets the intensifying digital demands of modern agriculture but also pioneers a new paradigm for sustainable ecological farming methods. The revelations from this investigation may carry far-reaching weight for the trajectory of animal stock oversight, potentially catalyzing broad-based acceptance of digitally driven wellness surveillance frameworks across the farming sector. Such a shift would aid in streamlining husbandry-based health management worldwide, thereby enhancing yields and long-term viability in the agrarian economy.

As global food requirements intensify and farming moves decisively toward sustainable and efficient models, establishing precise, forward-looking strategies to protect livestock health is increasingly critical. Within the egg and hatchery domain, hyperspectral imaging (HSI) has established itself as a sophisticated, non-invasive modality for swift, reliable egg quality evaluation, including the detection of embryonic death in chicks. Yet, steep costs and operational barriers relative to conventional RGB approaches present considerable barriers to HSI’s wide-scale uptake. Circumventing these constraints and enabling the broader promise of HSI, the reconstruction of hyperspectral data from ordinary RGB captures is a compelling answer. The current work focuses on producing hyperspectral imagery from RGB inputs to achieve non-destructive, early-stage identification of chick embryo mortality. An initial comparison was made among several image reconstruction frameworks—HRNET, MST++, Restormer, and EDSR—for generating hyperspectral representations of eggs at an early incubation phase. Following this, the reconstructed spectral information was used as input for XGBoost and Random Forest models to separate eggs with living embryos from those with deceased embryos. HRNET outperformed the alternatives in reconstruction quality, recording an MRAE of 0.0955, an RMSE of 0.0159, and a PSNR of 36.79 dB. This work promotes the concept that integrating imaging platforms with intelligent sensing and analytical processing can raise automation levels, strengthen biosecurity measures, and streamline resource use in pursuit of sustainable agriculture 4.0.

The surveillance of poultry movement patterns is an essential diagnostic tool for welfare evaluation, yet manual monitoring routines impose considerable temporal demands and physical strain, particularly in high-density flocks. Inadequately trained personnel may additionally fail to recognize subtle shifts in behavioral intensity, thereby overlooking critical windows for prophylactic intervention and remedial action. The present work merges conventional camera-based observation with a sophisticated automated recognition architecture to classify diverse broiler conduct within commercial housing. Deployment of a two-stage deep learning pipeline occurs as follows: initial detection localizes individual birds, followed by identification of five principal anatomical landmarks (cranial region, ventral region, bilateral limbs, and caudal appendage). From these points, a skeletal framework anchored in the ventral region is generated, yielding six angular measurements via trigonometric computations. A long short-term memory network subsequently processes these angular sequences, estimating the likelihood across behavioral categories encompassing “Standing”, “Walking”, “Resting”, “Eating”, “Preening”, and “Flapping”, with the highest-probability classification retained. To quantify the temporal distribution between stationary and dynamic states, a cascaded fuzzy-logic inference architecture was implemented, providing a resilient assessment of overall activity magnitude. Validation conducted within a mixed-gender rearing facility demonstrated that the proposed methodology attained performance floors of 85.2% for broiler categorization, 79.2% for anatomical landmark localization, and 50.8% for behavioral classification accuracy. Activity-level determinations generated by the system were consistent with evaluations by veteran poultry husbandry specialists.

Acquiring data across many locations and enabling fast decision-making are strengths of hyperspectral measurements. Nevertheless, a range of data processing approaches (Savisky-Golay [SG], first derivative [FD], and normalization) and modeling frameworks (partial least squares regression [PLS], weighted k-nearest neighbor [KKNN], support vector machine [SVM], and random forest [RF]) can be applied to establish the strongest link between physical observations and hyperspectral signatures. In this study, FD emerged as the top processing technique, while SVM was the leading model for estimating mean cotton (Gossypium spp., Malvaceae) height and node counts. Conversely, the integration of FD with RF achieved the highest accuracy in predicting cotton leaf area index, canopy cover, and chlorophyll levels throughout the growing period. Furthermore, outputs from models built with both SVM and RF corresponded closely to the pseudo-CHIME satellite wavebands, a process in which field-collected hyperspectral data were adjusted to the spectral intervals of an anticipated spaceborne hyperspectral sensor. The knowledge and outcomes supplied here will equip growers and other players in the cotton sector to execute swift, well-informed actions that could foster higher productivity and sustainable intensification.

A large segment of India’s rural population depends on agriculture for its livelihood, yet difficulties with manual processes and disease management persist. To address these problems, this study presents an automated framework for detecting leaf damage and identifying plant diseases using advanced deep learning methods. The introduced technique unfolds in six steps: first, YOLOv8 is harnessed to pinpoint regions of interest from drone-acquired images; second, DeepLabV3+ is applied for background stripping to aid disease categorization; third, a CNN model is deployed for reliable disease sorting, registering strong training and validation accuracy figures (96.97% and 92.89%, respectively); fourth, UNet-based semantic segmentation is utilized for meticulous pixel-scale damage identification, with an assessment accuracy of 99%; fifth, disease severity is gauged; and sixth, customized remedial measures are put forward according to the disease category and the extent of damage. Experimental trials using the Plant Village dataset confirm the technique’s ability to detect various defects across crops, including apple, tomato, and corn. This mechanized strategy carries significant potential to boost agricultural output and strengthen disease management both within India and globally.

Achieving food availability, advancing the long-term viability of farming, and propelling economic progress. However, current forecasting frameworks often overlook the specific characteristics of rice yield distributions, which vary across regions, exhibit skewness, and are constrained by fixed lower and upper thresholds, conforming to a four-parameter beta distribution. The consequence is outputs that are less reliable and can potentially misguide decision-making. Furthermore, a majority of techniques fall short in modeling the intricate variable interactions that arise when satellite observations are fused with field survey measurements, an integration widely viewed as essential for boosting forecast fidelity and streamlining cultivation methods. To remedy these deficiencies, the present research proposes a four-parameter beta Generalized Linear Mixed Model (GLMM) and advances it into a four-parameter beta Generalized Mixed Effect Tree (GMET). This four-parameter beta GMET extends the four-parameter beta GLMM architecture by embedding a regression tree mechanism, thereby providing superior flexibility for representing non-linear and interaction-laden dependencies. Deploying this analytical framework on empirical data from Central Kalimantan and Karawang yields significant gains over earlier approaches, as evidenced by markedly lower AIC and RRMSE metrics. Significantly, the investigation finds that time-lagged reflectance values from bands 4 and 8 and the NDVI index captured by Sentinel-2A satellites are the dominant determinants of rice productivity, overshadowing the explanatory contribution of farmer questionnaire responses. This finding underscores the promise of remotely sensed data as a cornerstone for rice yield estimation, offering a more streamlined, less resource-intensive alternative to reliance on farmer-reported data. By further refining satellite-based technologies, upcoming initiatives in yield forecasting can realize superior efficiency and precision, thereby underpinning evidence-based choices in agricultural governance.

Achieving optimal cannabis crop performance and product quality hinges on dependable, automated systems for leaf disease identification that allow for swift diagnosis and action. Present-day automated methods fall short of being purpose-built for the distinct difficulties inherent in spotting cannabis diseases, rarely holding up across shifting environmental contexts and varying phases of plant maturity. This paper proposes a novel Hybrid Adaptive Multi-Intelligence System for Deep Learning Ensembles (HyAMIS-DLE), grounded in an extensive data collection that mirrors the breadth of cannabis leaf pathologies and their unfolding. The design pairs non-population-based decision merging within the image preparation pipeline with population-based decision merging at the classification stage, calling upon a variety of CNN backbones. The outcome is a meaningful leap in key performance measures: HyAMIS-DLE achieves 99.58% accuracy, besting standard architectures by up to 4.16%, while also offering markedly improved resilience and a boosted Area Under the Curve (AUC) metric, enabling a clear distinction between unaffected and afflicted foliage. Rolling HyAMIS-DLE into our Automated Cannabis Leaf Disease Classification System (A-CLDC-S) confirms its on-the-ground value, helping achieve higher harvest volumes, reduce waste, and advance ecologically sound farming practices.

The surveillance of animal behavior is a vital tool in animal production systems. This approach to monitoring behavior can serve as an indicator of animal welfare and health status, potentially translating into superior productive outcomes. The present investigation was designed to determine the optimal accelerometer placement (on the halter compared to a neck collar) and data transmission sampling periods (from 6 up to 600 seconds) for the forecasting of behavioral repertoires, which included frequencies of water and feed consumption along with additional activities, in young beef cattle bulls under feedlot conditions. To meet the study’s goals, a suite of machine learning classifiers was implemented, including random forests, support vector machines, multilayer perceptrons, and naive Bayes. Every model examined achieved high classification performance (above 0.90) regardless of placement site, except those trained with the naive Bayes classifier. Accordingly, integrating accelerometers into collars represents a more feasible solution for on-animal use, given the relative ease of application compared with halter mounting. The adoption of a larger dataset (i.e., shorter sampling windows) did not substantially improve the performance of the fitted models. Consequently, operating with datasets of reduced observational density is more beneficial, as it can reduce both the computational load and time investment required for model fitting while simultaneously extending the device’s battery longevity.

Around the world, livestock producers are under pressure to curb greenhouse gas emissions to limit climate change. One potential avenue for grazing-based dairy enterprises involves adding substances that suppress methane production into the animals’ feed. A novel method for supplying these substances at the required regularity and accuracy is through smart feeders, a current smart farming tool already used to automatically dispense supplements to livestock while they are in the paddock. For this innovation to take hold, however, it must work within existing farm systems and deliver a clear, beneficial value proposition to farmers. This study set out to investigate the farm system and technology attributes that could shape the potential acceptance of in-paddock smart technologies for administering methane inhibitors in pasture-based settings. An adoption prediction instrument (ADOPT) was applied to simulate the diffusion patterns of smart feeders functioning as methane-inhibition delivery mechanisms on dairy operations, with contributions from industry specialists and farmers gathered through focus group discussions. The outcomes revealed that smart-feeder uptake within a pasture-based system framework would be modest. A deeper examination was conducted through a sensitivity analysis of seven decisive ADOPT variables identified as highly influential during the farmer focus group sessions. We modelled the effects of these seven essential ADOPT variables across two distinct smart-feeder designs to compare their adoption prospects. The adoption modelling indicated that, while considerations such as technological cost and capability were significant, uptake would also be profoundly affected by future regulatory directions, uncertainty surrounding the innovation, and how well it fits with farmers’ values and philosophies about their farm system. This investigation emphasized that the in-paddock delivery apparatus and the practical routines for its use on-farm constitute an early-phase innovation; consequently, farmers and other interested parties must be brought into the further refinement process to ensure that adoption barriers are properly addressed.

Fungi represent a highly nutritious dietary staple that can be essential for supplying cost-effective protein, vitamins, and minerals to populations around the globe; however, their successful cultivation requires extensive training and specialized knowledge to carefully calibrate a range of environmental variables. Internally displaced populations located in Northern Syria depend on extremely limited traditional oyster mushroom farming, which falls short of meeting local consumption requirements. Numerous international and local non-governmental organizations (NGOs) supporting Syrian refugees run mushroom-farming programs. They have encountered considerable difficulties and obstacles in mushroom farming among the intended participants. Thus, the two core questions steering this investigation are: (1) How might organic mushroom farming be advanced via a reliable and low-cost intelligent cultivation system? (2) In what ways can organic mushroom farming methods be streamlined to assist internally displaced persons and Syrian refugees? This investigation examines the mechanization of mushroom farming by designing and implementing an intelligent oyster (Pleurotus ostreatus) cultivation system for remote surveillance and adjustment of environmental conditions—temperature, humidity, air quality, and light—within the growing space. Additionally, ready-made and purpose-built, intuitive web interfaces were created to enable growers to survey and operate their farms via the Internet remotely. Consequently, a trustworthy and economically viable intelligent oyster mushroom cultivation system was designed and successfully implemented in this study. The system encompasses remote surveillance and operation via intuitive interfaces. This makes mushroom farming more straightforward not only for displaced populations and refugees but also for mushroom growers in nations with limited financial resources. This research can aid in the elimination of poverty and hunger, in accordance with the United Nations Sustainable Development Goals one and two.

Agricultural artificial intelligence is increasingly shaped by a precision paradigm that treats good farming as the increasingly fine-grained optimization of inputs, outputs, risks, and operational decisions. This paradigm is attractive because it promises measurable efficiency, predictive control, and scalable decision support across complex farming systems. In climate-sensitive agriculture, however, the same orientation can narrow what counts as a good decision. The problem is that climate-sensitive farming is not merely a technical setting in which better data automatically produce better outcomes. Farmers must act under uncertainty, ecological variability, long time horizons, and ethically contested trade-offs involving productivity, resilience, livelihoods, and environmental stewardship. These conditions expose the limits of optimization as a sufficient guide for responsible agricultural AI. This article develops a conceptual framework that shifts the guiding principle of agricultural AI from precision to prudence. Prudence is understood as practical wisdom: the capacity to deliberate well about what is fitting, responsible, and contextually appropriate when action is necessary but certainty is unavailable. The framework therefore reorients AI from prescribing optimal decisions to supporting wise and accountable judgment. The article is based on a conceptual synthesis peer-reviewed publications across agricultural ethics, responsible innovation, precision agriculture, AI governance, climate adaptation, and science and technology studies. It critiques precision agriculture as an optimization paradigm, articulates prudence as a responsible AI principle, and develops a framework for climate-sensitive decision support. Two tables contrast the precision and prudence paradigms and present the proposed conceptual framework. The central argument is not that agricultural AI should abandon data, modelling, prediction, or decision support. Rather, responsible AI for climate-sensitive farming should use these capacities to enlarge human judgment rather than replace it. A prudent AI system helps farmers deliberate under uncertainty, compare trade-offs, question recommendations, and remain accountable for decisions that are ecological, social, and moral as well as technical.

Smart farming increasingly embeds artificial intelligence, machine learning, predictive analytics, remote sensing, robotics, and automated decision-support systems into agronomic decision-making. These technologies promise more precise, timely, and data-intensive forms of farm management. Their growing role suggests that agronomic decisions are no longer shaped only by farmer observation, seasonal experience, and advisory knowledge, but also by algorithmic interpretation. This article addresses the theoretical problem that AI-supported farming systems may improve technical efficiency while simultaneously reshaping farmer judgment, experiential knowledge, trust, and decision authority. Smart farming is therefore not merely a matter of technological adoption. It is also a reconfiguration of how farming problems are recognised, evaluated, prioritised, and acted upon. The objective of the article is to develop a theory of human–AI decision dependence in smart farming. The theory explains how repeated interaction with algorithmic systems can create new forms of reliance, cognitive offloading, skill transformation, and redistributed authority. It also explains why dependence on AI can be supportive, excessive, or contested depending on the quality of system design, institutional context, and farmer capacity to interpret recommendations. The article contributes a conceptual model of human–AI decision dependence in smart farming. It argues that dependence is not simply a user attitude toward technology but a sociotechnical process through which judgment, skill, trust, evidence, and authority are redistributed. This perspective helps future smart farming research examine not only whether AI systems perform accurately, but also how they change the human capacities and decision relations on which farming depends.

Intelligent farming systems increasingly depend on data generated through sensors, machinery, satellites, digital platforms, advisory tools, farm management software, and artificial intelligence-enabled decision-support systems. These systems promise more precise monitoring, improved resource allocation, and better responsiveness to environmental and operational uncertainty. As farm operations become more data-intensive, farm data shift from being a by-product of agricultural activity to becoming a strategic asset within intelligent farming systems. This transformation creates governance problems that cannot be resolved through technical data management alone. Farm data raise contested questions about ownership, consent, access, interoperability, platform dependence, accountability, and farmer control. These questions are especially important because data generated on farms may be used beyond the original farming context for commercial analytics, platform development, benchmarking, insurance, input recommendation, credit scoring, or supply-chain governance. The objective of this article is to develop a data stewardship framework for intelligent farms. The framework is designed to clarify how ownership, consent, access, interoperability, accountability, and farmer-control mechanisms can be organised in ways that support responsible digital agricultural innovation. It treats data stewardship as a governance function that mediates between innovation, commercial data use, and the protection of farmers’ decision-making authority. The article is based on a governance framework synthesis  peer-reviewed articles published. It does not present new empirical data. Instead, it synthesises literature on digital agriculture, smart farming, agricultural data governance, platformisation, responsible innovation, interoperability, and farmer data rights to construct a practical and conceptual governance framework. The article maps farm data as a governance problem, analyses ownership and consent issues, examines interoperability and platform dependence, and proposes a stewardship framework for intelligent farms. Three tables summarise ownership and control issues, interoperability risks, and the proposed framework. Responsible intelligent farming requires data stewardship arrangements that protect farmer control while enabling trustworthy data sharing and innovation.

The convergence of artificial intelligence and the Internet of Things is creating new possibilities for adaptive farming systems. In agricultural settings, this convergence allows sensor networks, machine learning models, communication infrastructures, and automated decision support tools to operate as interconnected cyber-physical systems. Yet many current smart farming architectures remain technically fragmented and do not fully connect operational decisions with sustainability feedback. A central problem is that sensing, intelligence, agronomic reasoning, and sustainability assessment are often treated as separate system layers. Sensor data may support monitoring, machine learning may support prediction, and sustainability indicators may be reported after implementation, but these functions are rarely organised as a continuous adaptive loop. This separation limits the ability of farming systems to learn from environmental consequences and adjust decisions in real time. This article proposes a conceptual architecture for AIoT-enabled adaptive farming systems. The architecture integrates distributed sensor networks, edge-based intelligence, agronomic knowledge models, and sustainability feedback mechanisms into a closed-loop system. It is designed for farming environments where decisions must be locally responsive, scientifically grounded, and aligned with long-term resource stewardship. The article is based on a conceptual design synthesis of peer-reviewed literature published. The synthesis draws on agricultural IoT, edge computing, machine learning, crop monitoring, remote sensing, smart farming platforms, and sustainability-oriented digital agriculture. No new empirical data are presented. The resulting architecture specifies four interacting layers: sensing, edge intelligence, agronomic knowledge integration, and sustainability feedback. It defines how sensor observations move through edge analytics and knowledge-informed decision support before being evaluated against sustainability indicators that feed back into future control actions. The article contributes a systems blueprint for adaptive farming architectures that are not only intelligent and data-driven but also environmentally responsive and sustainability-oriented.

Intelligent agriculture is often presented as the next great leap in farming: a data-rich, sensor-driven, algorithmically guided transformation of how food is produced. By combining artificial intelligence, Internet of Things infrastructures, remote sensing, robotics, and big data analytics, it promises to make farming more precise, efficient, and responsive. Yet much of this promise remains tied to an older and narrower ambition: producing more output per hectare. This perspective argues that intelligent agriculture has inherited and digitally extended the yield-maximization logic of twentieth-century productivist agriculture. The problem is not that yield is unimportant, but that yield has become too dominant as the organizing measure of agricultural progress. When farming systems are judged primarily by short-term productivity, the underlying ecological foundations of production can be degraded while the system still appears technically successful. The article advances a reframing of intelligent agriculture around three interconnected goals: soil health, biodiversity, and long-term farm resilience. These are not secondary environmental concerns to be considered after productivity has been secured. They are the living infrastructure of agriculture and the conditions through which future productivity, livelihoods, and food security become possible. The central claim is that the future of intelligent agriculture should not be defined by producing more at any cost. It should be defined by producing with care: sustaining soil function, enhancing biodiversity, and strengthening the adaptive capacity of farms under climate uncertainty. This requires not simply better technologies, but a deliberate change in the goals, metrics, and values that guide agricultural innovation.

Smart farming has entered a phase in which agricultural fields, machines, livestock systems, greenhouses, and advisory platforms increasingly generate continuous streams of digital information. Sensors promise greater precision, faster monitoring, and more responsive management, yet their proliferation has not automatically improved farmer decision-making. Many farms now face a paradox in which more data are available than can be interpreted, trusted, or acted upon in routine agronomic practice. The central problem addressed in this article is the widening gap between sensor abundance and actionable agronomy. Smart farming systems often excel at data collection but remain less effective at filtering, prioritising, and translating data into decisions that farmers can use under time, labour, weather, and resource constraints. This creates a data-to-decision gap that reduces the practical value of agricultural digitalisation. The objective of the article is to propose a conceptual pathway for reducing data overload in smart farming systems. The pathway moves from sensor data acquisition to information filtering, decision relevance, decision translation, and actionable agronomic delivery. It is designed to help researchers, developers, advisors, and farm managers evaluate not only whether data are collected, but whether they are converted into usable decisions. The proposed pathway identifies four sequential stages: sensor data acquisition, information filtering and relevance assessment, decision translation, and actionable agronomy delivery. Three tables support the framework by cataloguing sensor sources, classifying filtering strategies, and presenting the full pathway. The article concludes that the value of smart farming should be judged less by the amount of data generated and more by the quality, usability, and timeliness of the agronomic decisions enabled.

Autonomous farming systems are moving rapidly from experimental prototypes into commercial and operational agricultural settings. Autonomous tractors, robotic weeders, drones, harvest-assist systems, and sensor-guided machines promise to address labour shortages, improve operational efficiency, and support more precise field management. Yet the governance arrangements needed to ensure that these systems remain safe, understandable, and accountable have not developed at the same pace. The central problem is not only whether autonomous machines can perform agricultural tasks accurately. It is whether farmers, workers, contractors, and rural communities can understand, supervise, question, override, and hold these systems accountable when they act in complex farm environments. Without deliberate attention to trust calibration, human override, explainability, and accountability, autonomous farming may produce technically impressive but socially fragile systems. This article develops a Human-Centred Governance Framework for autonomous farming. The framework is designed around four pillars: calibrated trust, meaningful human override, explainable decision-making, and clear accountability structures. Its purpose is to shift governance attention from machine capability alone toward the conditions under which autonomous systems can be responsibly integrated into agricultural work. The proposed framework shows that human-centred governance is not an optional ethical supplement to autonomous farming. It is a practical requirement for safe deployment, legitimate adoption, and accountable operation. By integrating trust, override, explainability, and accountability, the framework provides guidance for farm managers, technology developers, regulators, and agricultural service providers.

Intelligent agriculture is often presented as a technological pathway for increasing productivity, reducing input waste, and improving farm-level decision-making. In this view, the farm is commonly treated as an optimisation target in which sensors, analytics, and automated recommendations are used to improve decisions at a particular moment. Although this perspective has advanced digital agriculture, it does not fully explain how farms accumulate experience, revise practices, and become more capable over time. This article argues that the intelligent farm should be conceptualised not only as a data-driven production system but also as a learning system. A learning system is one that senses changing conditions, interprets feedback, adapts action, and modifies its decision logic across repeated cycles of experience. This framing shifts attention from isolated optimisation events to the longer-term evolution of farm knowledge, adaptive capacity, and agronomic intelligence. The objective of the article is to develop a theory of the intelligent farm as a learning system. The theory integrates three mechanisms: continuous feedback, local adaptation, and agronomic intelligence evolution. Together, these mechanisms explain how intelligent farms can move from collecting data to building situated knowledge that improves future decision quality. The proposed theory shows that intelligent farms become smarter when feedback streams are interpreted through local context and converted into evolving decision rules. Two tables are included: one contrasts optimisation and learning system views of the intelligent farm, and the other presents the full theoretical model. The article concludes that treating farms as learning systems offers a more realistic and resilient paradigm for intelligent agriculture.

Artificial intelligence is increasingly presented as a transformative force in agriculture, promising more precise decisions, better resource use, improved productivity, and more resilient farming systems. Yet much of this promise has been shaped around large-scale, capital-intensive, data-rich farms in high-income settings. This creates a structural mismatch between dominant agricultural AI imaginaries and the everyday realities of smallholder farming. Smallholder farmers operate under conditions that are often treated as barriers to technological adoption rather than as core design conditions. Limited connectivity, intermittent electricity, restricted access to smartphones, low disposable income, linguistic diversity, and uneven access to advisory support all shape how digital tools are encountered and used. When these realities are ignored, agricultural AI risks widening rather than narrowing existing inequalities. This article develops a Smallholder-Centred AI Model for agriculture. The model reorients agricultural AI around four foundational principles: low-resource design, inclusion and affordability, context-awareness, and farmer participation. It argues that these principles must not be added after technical development but must structure the design, governance, and evaluation of AI from the beginning. The article concludes that AI can serve smallholder agriculture only when smallholders are positioned as active participants in the intelligence of the system. A smallholder-centred approach requires technologies that work with limited infrastructure, respond to local agroecological and social contexts, respect farmer knowledge, and remain affordable and accountable. The future of agricultural AI must therefore be judged not only by predictive accuracy but by its capacity to support inclusive, locally grounded, and farmer-led transformation.