This research presents an IoT-enabled smart aquaponics system integrating multi-modal sensor networks with advanced machine learning algorithms to achieve autonomous optimization of symbiotic crop–fish production environments. The system employs a distributed sensor architecture incorporating pH, dissolved oxygen, ammonia, nitrate, temperature, humidity, light intensity and turbidity sensors, interfaced through ESP32-based edge computing nodes with real-time data transmission capabilities via LoRaWAN and WiFi protocols. The core innovation lies in the implementation of a hybrid deep learning framework combining Convolutional Neural Networks (CNNs) for image-based plant health assessment, Long Short-Term Memory (LSTM) networks for temporal pattern recognition in water quality parameters, and Reinforcement Learning (RL) agents for dynamic system optimization. The AI model processes over 15,000 data points hourly, enabling predictive analytics for disease detection, nutrient deficiency identification and growth trajectory forecasting with 94.7% accuracy. Advanced computer vision algorithms utilizing YOLOv8 architecture perform real-time fish behavior analysis and biomass estimation, while hyperspectral imaging integrated with transformer-based attention mechanisms monitors plant stress indicators at cellular resolution. The system's autonomous control mechanisms regulate LED spectrum optimization (380–780 nm), nutrient dosing through precision peristaltic pumps, water circulation via variable-speed pumps and climate control through HVAC integration. Experimental validation demonstrates a 43% increased crop yield, 28% enhanced fish growth rates and a 35% reduction in water consumption compared to conventional systems. The platform achieved 99.2% uptime with sub-second response times for critical parameter adjustments. Machine learning models successfully predicted system failures 72 hours in advance, enabling proactive maintenance protocols. The system's scalability is demonstrated through blockchain-based data integrity verification, edge-to-cloud hybrid processing architecture, and standardized API interfaces enabling seamless integration with existing agricultural management systems. This breakthrough technology represents a paradigm shift toward sustainable, intelligent food production systems capable of addressing global food security challenges while minimizing environmental impact.

The rapid spread of the Panama disease outbreak presented significant difficulties for banana growers during the COVID-19 pandemic ("The ‘Banana Pandemic’ Destroying the World’s Favourite Fruit," n.d.). The growers' insufficient knowledge and skills in cultivating and managing their banana trees worsened the effects of multiple diseases, such as the absence of a definitive cure for Panama wilt disease. This study explores the experiences of banana growers affected by Panama disease during the COVID-19 pandemic. The participants of this qualitative phenomenological study were seven banana growers, and the data were obtained through in-depth interviews. The responses were classified and grouped into themes developed by a data analyst. The findings indicated that financial hardships are a central focus, with growers contending with risks to their livelihoods and diminishing sources of income. Banana growers face heightened emotional strain due to uncertainties, but amidst these challenges, they show resilience and adaptability, highlighting their ability to endure tough times. The researchers found that the participants' concerns must be addressed through targeted interventions, comprehensive mental health approaches, financial support, promoting interdisciplinary collaboration, and resilience-building to support banana growers in the face of persistent challenges. Furthermore, this study encompassed Sustainable Development Goal (SDG) 3, Good Health and Well-being, emphasizing mental health support for banana growers, and SDG 12, Responsible Consumption and Production, promoting sustainable agricultural practices for long-term productivity and environmental health.

Abstract
Introduction:
The cultivation of Stevia rebaudiana, a plant of increasing nutritional and economic value, requires strict control of environmental conditions to ensure high leaf quality and optimal glycoside content. The TELEMETRY project aims to develop a remote telemetry system for the precision monitoring of Stevia cultivation, enabling sustainable agricultural practices through real-time decision support. The system also supports proactive management through a rule-based alerting mechanism and neural networks, enabling the forecasting of future environmental and cultivation conditions.

Methods:
The system integrates Narrow Band IoT (NB-IoT) wireless sensors to measure critical environmental (temperature, humidity, rainfall, and soil moisture) parameters in a 4-hectare experimental plot managed by Stevia Hellas Coop. Sensor data are transmitted to central nodes and further relayed to a cloud-based storage and alert system. At the same time, local farmers perform traditional manual measurements (e.g., using analog hygrometers), which serve as a reference baseline for validating the sensor data. This comparative process enhances sensor reliability and contributes to the improvement of data accuracy for downstream machine learning models, including neural networks. The experimental layout ensures data uniformity across replicated plots.

Results:
Initial deployments confirmed the system's robustness under field conditions. Sensor-based monitoring enabled early identification of disease-favoring microclimates (e.g., high dew point and humidity), facilitating timely phytosanitary interventions. Compared to traditional irrigation scheduling, the telemetry-guided regime achieved significant water savings while maintaining efficient plant growth. Deviations in soil microclimate were detected and addressed through localized management.

Conclusions:
The TELEMETRY system demonstrates the potential of IoT-based solutions for precision agriculture in specialty crops such as Stevia. By integrating real-time data with grower decision-making, the system contributes to input reduction, disease prevention, and high-value product traceability. Future work will focus on increasing the monitored parameters and scaling up the system for broader deployment.

Introduction: The wine industry is facing significant challenges, with only sparkling wine showing growth. Southern Italy is particularly impacted due to its focus on white and full-bodied red wines. To address this, we have created new sparkling wines using the Champenois method and six typical Apulian white grape varieties: Bombino, Antinello, Falanghina, Fiano, Greco Bianco, and Montonico Pinto, along with native yeast strains. The concept of “microbial terroir” highlights the connection between the microbial environment, climate, and production area, linking wines to their cultural and historical roots. Utilizing indigenous yeast strains enhances the wine's identity and ties it to the notions of heritage and terroir.

Methods: Grape, base wines, and sparkling wines were analyzed for conventional parameters (pH, titratable acidity, volatile acidity, alcohol content, and residual sugars). The aromatic profile of final sparkling wines was assessed through GC-MS and sensory analyses. Base wines were produced through sequential inoculation of a native L. thermotolerans (Lt) and a S. cerevisiae strain (VB1, Oenobrands), or only with the commercial S. cerevisiae. For the second fermentation, two different S. cerevisiae strains were used: an indigenous yeast strain (S21) isolated from Apulian vineyards, or a commercial one (18-2007 IOC).

Results: The most appreciated sparkling wines were those made from Antinello, Fiano, and Falanghina in combination with the native S. cerevisiae S21 strain. These wines showed floral and white fruit notes, linked to the chemical compounds identified through GC/MS analysis (alcohols, carboxylic acids, esters, terpenoids, lactones, and others). Additionally, visual characteristics, such as color and perlage, contributed to the positive evaluation of these wines. While native S21 yeast allowed for the production of well-received products, unfortunately, the wines produced with Lt during the first fermentation did not perform well in sensory evaluations.

Conclusion: Offering sparkling wines with a background, based on the use of local varieties and native yeast, can be a strategy to capture consumer interest and drive demand.

The citrus industry worldwide has been devastated by widespread diseases, particularly greening, canker, and black spot, leading to significant tree losses, orchard closures, and reduced orange production. The traditional inspection methods for detecting such diseases are expensive and inefficient, thus warranting a better solution. This study aims to compare the effectiveness of different AI-powered, real-time computer vision architectures in accurately detecting and classifying citrus diseases through imagery. Two object detection models were compared: YOLOv8, a Convolutional Neural Network (CNN), and RT-DETR, a Vision Transformer (ViT). Both models were trained and cross-validated on a custom benchmark dataset, which featured 6,000 citrus images. This included 1,500 original images, as well as 4,500 images through augmentation, which were split into three difficulty levels to test the model response to varying simulations of real-world conditions such as lighting and motion blur. Initial training across the original dataset revealed that YOLOv8 outperformed RT-DETR in its accuracy and real-time speed by a slight margin. The weight decay, learning rate, and batch size were finetuned via Bayesian optimization. Additional few-shot learning on several other datasets boosted the performance and speed, resulting in 92.5% mean average precision (mAP) for YOLOv8 versus 87.07% mAP for RT-DETR. While YOLOv8 performed better overall, RT-DETR demonstrated a better performance on the hardest set, which displayed the model’s robustness in difficult environmental conditions. The optimized models were deployed on a Raspberry Pi 5 with a camera module and several sensors. Field tests in a citrus grove confirmed successful real-time detection and accurate classification of diseased leaves and fruits, with visual explainability through Grad-CAM analysis. This research showcases the viability of low-cost platforms for object detection and introduces a novel data framework for future research into AI implementation in the citrus industry, allowing for the early detection and rapid treatment of such diseases.

Mauritania’s agricultural sector, particularly rice production, faces significant challenges due to climate variability, market inefficiencies, and limited access to technological resources. In response, this study proposes a smart agriculture framework that integrates Artificial Intelligence (AI), Big Data, and simulated Internet of Things (IoT) technologies to enhance rice yield forecasting and farm climate monitoring. Building upon our prior work—including yield prediction via Random Forest and LSTM models, IoT-based digital twin climate monitoring, and national agricultural datasets—we present an improved architecture combining real-time data analytics and scalable decision support.

The framework utilizes historical datasets spanning 1960–2023, covering rice yields, agricultural value added, fertilizer use, population, employment, trade, and rice prices. Simulated IoT data, based on historical temperature and humidity records, serve as a proxy for field sensors in data-scarce environments. Three predictive models were benchmarked: Random Forest achieved the highest R² (0.87), followed by XGBoost and LSTM. Feature importance analysis highlighted temperature, rainfall, and fertilizer use as key yield predictors. A weak correlation (0.08) between retail and wholesale prices indicates limited market integration, which may affect production planning.

Beyond model accuracy, the study emphasizes the practical value of integrating AI-driven prediction with IoT-based monitoring to support precision agriculture in Sub-Saharan Africa. This approach enables farmers to make data-informed decisions on crop scheduling and resource allocation, potentially increasing yields by 10–15% under climate uncertainty. It also lays the groundwork for a future SaaS platform tailored to smallholder needs, incorporating real-time alerts, decision dashboards, and mobile access.

The proposed solution aligns with the goals of climate-smart agriculture and sustainable development. Future work will focus on deploying real sensors, validating predictions in the field, and expanding to other staple crops critical to regional food security.

Pesticides comprise a broad category of chemical substances with diverse chemical compositions. They are predominantly employed by agricultural producers to protect crops from pests and enhance yields. However, the presence of pesticide residues in agricultural products used for food production represents a potential health risk, particularly for vulnerable consumer groups such as infants and young children. From a safety control perspective, the focus is on pesticide residues that accumulate in agricultural products intended for food production.

According to the World Health Organization, pesticide residue is defined as any substance or mixture of substances found in food for humans or animals, resulting from the application of pesticides. This includes "any specific derivatives such as degradation products, conversion products, metabolites, and reaction products that are considered to be toxicologically significant." The accumulation of such residues can trigger a range of adverse health effects, including sensitivities, allergies, and the onset of serious diseases, including oncological, neurodegenerative, and chronic conditions.

The control of infant food involves the processing and analysis of samples of infant formula for children up to 6 months of age, follow-up formula for children from 6 months to 1 year, cereal-based food products for infants and young children, and samples of fruit- and vegetable-based baby food.

The National Reference Center for Pesticide Residues has established and validated a multi-residue method for detecting over 200 pesticide residues in infant food, employing liquid and gas chromatography coupled with mass spectrometric detection (LC-MS/MS, GC-MS/MS). The samples were pre-treated using the QuEChERS method.

Recent results from official regulatory controls and European monitoring efforts demonstrate the safety of infant food in terms of pesticide residue levels. No tested sample over the past few years has exceeded the legally permissible limit for pesticide residues.

Addressing the persistent issue of food insecurity—particularly in regions experiencing agricultural deficits—requires accurate and timely forecasting of crop yields. Reliable yield predictions are vital for policymakers in key agricultural regions, as they facilitate strategic planning for the redistribution of surplus commodities through international trade. This, in turn, contributes to regional food security and enhances the economic stability of exporting nations.

Among the most widely cultivated crops globally are rice, wheat, maize, soybean, and pigeon pea. Yield prediction for these crops has traditionally relied on various statistical and machine learning approaches. Recently, Quantum Machine Learning (QML) methods—particularly Quantum Neural Networks (QNNs)—have been proposed as a novel paradigm for forecasting applications due to their potential to capture complex, high-dimensional relationships in large datasets. QNNs, leveraging qubits and quantum gates, offer computational advantages over classical models in specific contexts, especially in handling non-linear interactions and entangled feature spaces.

This study aims to forecast the yields of the aforementioned five major crops for the years 2025, 2026, and 2027 using historical data spanning from 1961 onward. The dataset, curated from the Food and Agriculture Organization (FAO) of the United Nations, comprises officially reported statistics from national bureaus across the five leading producing countries for each crop. A training-to-testing split of 75:25 was utilized to develop and evaluate predictive models.

Several baseline models—including conventional statistical regressors and classical machine learning algorithms—were implemented and benchmarked against quantum counterparts, namely Quantum Neural Networks (QNNs) and the Variational Quantum Regressor (VQR). Model evaluation was conducted using standard performance metrics, including Root Mean Squared Error (RMSE), Relative Root Mean Squared Error (RRMSE), and Minkowski distances (with p-values of 1, 2, and 3).

The results demonstrate that quantum models, particularly QNNs, exhibit competitive or superior performance in yield forecasting tasks compared to classical models, highlighting their potential as effective tools for data-driven agricultural decision-making in the quantum era.

Drought stress, impacting around 66% of arable land, poses a significant threat to global food security, with estimated losses of USD 37 billions, and is expected to rise due to climate change. Pigeonpea, a key legume crop in semi-arid regions, often suffers yield losses of 40–50% under drought. This study assessed 47 genetically diverse pigeonpea genotypes under well-watered and drought-stressed conditions using morpho-physiological and high-throughput phenomic traits to enhance crop resilience. Drought stress significantly reduced most traits (2.77%–68.76%), while vapour pressure deficit and proline content increased by 1–1.5 times. Phenomics-based traits like projected shoot area, convex hull area, calliper length, and near-infrared reflectance proved effective in distinguishing drought responses. Nine genotypes—ICPX140196-B-1, ICPX140203-B-1, ICPX140213-B-3, ICPX140203-B-1-5, ICPX140203-B-2, GRG-152, ICPX140205-B-4, ICPX140217-B-1, and ICPX140188-B-3—emerged as drought-tolerant, showing stable yield (≤5% loss), efficient photosynthesis, and better water status. Photosynthetic traits were strongly correlated with yield under stress and exhibited high genetic variability (10–46%) with moderate heritability, making them reliable selection criteria. This study highlights the value of integrating phenomics with traditional traits to identify drought-tolerant pigeonpea genotypes and guide future breeding strategies efficiently.

Description of the subject: Pedotransfer functions are based on the search for mathematical relationships allowing the water properties of soils to be deduced from known or easily measurable characteristics of the soils.

Objective: This work aims to establish pedotransfer functions (FPTs) at eight levels of soil water retention potential on a north–south transect from Algiers to Djelfa (Algiers, Mitidja, Médéa, Djelfa).

Methods: The measurements and analyses were carried out on a set of 38 soil samples representing the main soils of Algeria (vertisols, fersiallitic soils, calcimagnesic soils, less evolved soils). The importance of the contribution of various soil variables with respect to water retention is estimated by the values ​​of the determination and correlation coefficients. The level of reliability of the established pedotransfer functions was estimated using 20 test samples from the studied soils.

Results: The main results obtained showed that clay, organic matter and fine silt, and coarse silt and total limestone are the factors that contribute the most to the water retention of soils, including at low potentials (-1600 kPa). In addition, it emerges from this study that reworking of the ground provoked an increase in porosity, essentially structural, in spite of small variations in its values. Validation of the pedotransfer functions indicates that the biases of the predictions are low, thus reflecting a good quality estimate of the water contents.

Conclusion: The results of this research show once again the importance that should be given to understanding the water functioning of soils in the Mediterranean environment. The results obtained for this study, covering a very large space and a wide range of Algerian soils, deal with a diversity of soils with different parental materials, located under several bioclimatic floors and with varied occupations.

We highlight several interesting results in particular on the soil factors to take into consideration for the calculation of the equations.