基于机器学习的生鲜农产品供应链

摘要: 农产品供应链稳定对保障市场供应、稳定需求、扩大居民消费需求具有重要意义. 新兴技术的应用可以提高农产品供应链的价值，降低传统供应链的风险. 因此，本文在机器学习算法的基础上，将贝叶斯算法融入决策树模型，建立了一种用于农产品供应链风险评估预测的模型. 首先，基于贝叶斯算法设计了用于风险划分的决策树模型. 然后，使用随机变量的权重构建决策树模型用于风险预测评估. 最后，将决策树模型应用于农产品供应链(APSC)，提出了APSC管理框架，帮助决策者进行评估分析，以降低农产品供应链风险. 实验结果表明，对比模型预测风险的性能达到90%左右，而本文提出的模型可以达到95%左右，证明其具有较高的预测准确性. 敏感性分析证实，该模型对市场变化具有较高的灵敏度，可以帮助决策者更好地评估农产品供应链风险并及时调整供应链机制.

摘要: 为破解特大城市“菜篮子”安全传统监管面临的多重困境，数字化转型已成必然发展路径。该文以广州市定点供穗生猪养殖基地管理信息系统为例，剖析信息工程技术支撑下的数字化转型模式，在保障生猪及相关产品供给安全中的实践应用。结果显示，该系统通过构建“养殖-屠宰”全链条数字化追溯体系、重构多主体协同业务流程、赋能数据驱动精准监管，破解了传统监管模式中存在的信息孤岛、响应滞后与责任追溯不畅等难题。该案例为我国同类特大城市“菜篮子”安全治理提供经验，同时揭示了利用前沿技术迈向智慧化、主动型监管的未来发展方向。

The digital and intelligent transformation of the agricultural product supply chain is considered crucial for improving agricultural efficiency， enhancing supply chain resilience， and promoting sustainable development. In this paper， 689 English - language literatures published by foreign scholars and 753 Chinese and English literatures published by domestic scholars， which are indexed in the Web of Science and CNKI from 1998 to 2023， are selected as samples. Bibliometric analyses on aspects such as the number of publications， keywords， and research hotspots are carried out on the data using CiteSpace and VOSviewer， and the historical evolution of the research on the digital and intelligent transformation of the agricultural product supply chain is sorted out. The research results show that： in terms of the number of publications， an upward trend year by year has been shown in the research on the digital and intelligent transformation of the agricultural product supply chain both at home and abroad； in terms of high - frequency keywords， the research focus of the digital and intelligent transformation of the agricultural product supply chain has been closely associated with technologies such as blockchain， the Internet of Things， and big data； in terms of the clustering of high - frequency keywords， eight clusters have been formed in the research on the digital and intelligent transformation of the agricultural product supply chain， and there is a common category of "blockchain" technology. Finally， the new models and new trends emerging in the digital and intelligent transformation of the agricultural product supply chain are analyzed， and prospects for future research were made.

… , and cloud computing can change production management practices. The current study presents a systematic review of machine learning (ML) applications in the agri-food supply chain…

Abstract - The agri-food industry faces persistent challenges in managing supply chains, including inefficiencies, traceability issues, and compliance concerns. Leveraging blockchain technology offers transparency, yet optimizing decision-making remains a challenge. This research explores the integration of deep reinforcement learning (DRL) with blockchain-based agri-food supply chains to enhance efficiency, transparency, and compliance. The study focuses on developing DRL models, integrating them with blockchain, optimizing operations, ensuring quality control, and evaluating performance. Key Words: Agri-Food Supply Chains, Blockchain Technology, Deep Reinforcement Learning, Decision-Making Optimization, Transparency, Compliance

In agri-food supply chains (AFSC), ensuring both agri-food safety for consumers and increased profitability for farmers remains a complex challenge. The intricacies and dynamics of AFSC demand effective traceability and management solutions that existing methods often fall short of providing. To address these challenges, we introduce a novel approach that combines blockchain technology for agri-food traceability with a Deep Belief Network-based secure Block Chain Management named DBCM for profit optimization. The proposed blockchain-based AFSC framework serves as the foundation for product traceability, offering decentralized security for agri-food tracing data within AFSC. This secure and transparent platform ensures that consumers can trust the safety and origin of the products they purchase, while farmers can leverage it to increase their profit margins. For enhancing the security an advanced hybrid cryptographic mechanism named HEDS (Homomorphic Encryption with Digital-Ring Signatures). The objective of the Data-Based Decision Making with Hierarchical Ensemble Decision Support (DBCM with HEDS) method that has been proposed is to arrive at safe data-driven decisions concerning the manufacturing and warehousing of agri-food products. By optimizing these operations, it facilitates higher profits for farmers. Unlike traditional methods, the DBCM-HEDS employs advanced machine learning techniques to adapt to the dynamic nature of AFSC, resulting in improved profit margins. the effectiveness of our blockchain-based framework is rigorously evaluated through extensive simulation experiments conducted in various AFSC environments. Furthermore, the DBCM-HEDS method consistently outperforms traditional methods, consistently delivering higher product profits. This innovative combination of blockchain technology and deep belief network offers a promising solution to the intricate challenges faced by agri-food supply chains, benefiting both consumers and farmers.

The agri-food supply chain is a complex network enclosing various stakeholders, from farmers to consumers, with multifaceted interactions and dependencies. Traditional supply chain management approaches often need help adapting to dynamic environments and optimising decision-making processes. Deep reinforcement learning is employed by integrating value-based and policy-based models, enhanced by advanced learning techniques, to tackle these challenges. This paper explores applying Deep Reinforcement Learning (DRL) approaches, including Q-learning, Deep Q-Learning (DQL), and the Actor-Critic method, to optimise the efficiency of the agri-food supply chain. The actor-critic model significantly enhances decision-making processes across various supply chain stages by improving efficiency and increasing profit margins. A specific scenario of sugar processing and distribution is incorporated, considering real-world scenarios to validate our model. DRL methods optimise sugar production, storage and distribution, ensuring timely deliveries and enhancing profitability. The models address fluctuating demand and transportation logistics challenges, resulting in a more streamlined and responsive sugar distribution network. The findings reveal that Actor-Critic and DQL methods significantly outperform traditional Q-learning considering product profitability, offering unique advantages in handling complex state-action spaces.

This research proposes a new framework for agri-food capacity production by considering resiliency and robustness and paying attention to disruption and risk for the first time. It is applied robust stochastic optimization by adding robustness to the constraint's objective function and resiliency situation. This research minimizes the mean absolute deviation and coefficient of standard deviation errors by linear function in the agri-food capacity production. This study suggests agri-food managers and decision-makers use this mathematical method to forecast and improve production management. The results of this research lead to better decision-making and are compared with other sine functions. The main model's Robust and Resiliency Mean Absolute Deviation (RRMAD) value is 1.28% lower than other sine-type functions. The conservativity coefficient, confidence level, weight factor, resiliency coefficient, and probability of the scenario vary. The main model's RRMAD value is 1.28% lower than other sine-type functions. Growing the weight factor will result in an increase in RRMAD and a smooth decline in R-squared . Additionally, as the resilience coefficient rises, the RRMAD function increases while the R-squared declines. By altering the probability of the scenario, the RRMAD function drops, and the R-squared goes up.

Abstract The agri-food sector contributes significantly to economic and social advancements globally despite numerous challenges such as food safety and security, demand and supply gaps, product quality, traceability, etc. Digital technologies offer effective and sustainable ways to these challenges through reduced human interference and improved data-accuracy. Innovations led by digital transformations in the agri-food supply chains (AFSCs) are the main aim of ‘Agri-Food 4.0’. This brings significant transformations in the agri-food sector by reducing food wastage, real-time product monitoring, reducing scalability issues, etc. This paper presents a systematic review of the innovations in the agri-food for digital technologies such as internet-of-things, artificial intelligence, big data, RFID, robotics, block-chain technology, etc. The employment of these technologies from the ‘farm to fork’ along AFSC emphasizes a review of 159 articles solicited from different sources. This paper also highlights digitization in developing smart, sensible, and sustainable agri-food supply chain systems.

PurposeThis study aims to analyse how the adoption of Industry 4.0 technologies can help different types of agri-food supply chains introduce and manage innovations in response to the challenges and opportunities that emerged following the COVID-19 pandemic.Design/methodology/approachA systematic literature review methodology was used to bring together the most relevant contributions from different disciplines and provide comprehensive results on the use of I4.0 technologies in the agri-food industry.FindingsFour technological clusters are identified, which group together the I4.0 technologies based on the applications in the agri-food industry, the objectives and the advantages provided. In addition, three types of agri-food supply chains have been identified and their configuration and dynamics have been studied. Finally, the I4.0 technologies most suited for each type of supply chain have been identified, and suggestions on how to effectively introduce and manage innovations at different levels of the supply chain are provided.Originality/valueThe study highlights how the effective adoption of I4.0 technologies in the agri-food industry depends on the characteristics of the supply chains. Technologies can be used for different purposes and managers should carefully consider the objectives to be achieved and the synergies between technologies and supply chain dynamics.

This work consists of a systematised review examining recent models for agri-food supply chain management (AFSCM), focusing on artificial intelligence (AI) applications and sustainability integration. Through PRISMA methodology, we conducted a descriptive statistical analysis of 183 articles published between 2019 and 2023. The dimensional analysis revealed that strategic modelling represented 55% of approaches, while perishable products were considered in 58% of studies. Sustainability principles were integrated into 74% of the models, while AI applications showed a limited 14% adoption across AFSCM. The analysis of AI implementations showed distribution in agricultural production management at 27%, transportation and logistics at 8%, and sustainability planning at 19%. A novel AFSCM taxonomy was developed through systematic classification, combining quantitative bibliometric analysis, qualitative thematic analysis and expert interpretation. The results identified gaps in mixed methods approaches and tactical or operational modelling, providing a structured framework for future research directions in AI integration and sustainable practices across agri-food supply chains.

… machine learning (ML) emerges as a revolutionary technology that can address the challenges faced by agricultural supply chains … present in the agricultural supply chain [18, 19]. …

Global agri-food systems face a critical conflict between the need to feed a growing population and the imperative to mitigate its substantial environmental impact, including 23% of global greenhouse gas emissions and 70% of freshwater withdrawals. This bibliometric review maps the scientific landscape of Machine Learning (ML) research applied to sustainable agri-food systems. Using a structured bibliometric protocol, we analyzed 648 scientific documents from Scopus (2010–2025) to map the evolution, collaborative networks, and thematic trends in this domain. Results reveal a field that has grown exponentially until 2021, primarily driven by contributions from Computer Science (26%) and Engineering (21%), with key publications in journals such as Computers and Electronics in Agriculture (22 papers, 2631 citations). While China and India lead in productivity (80% of top authors), high-impact research remains strongly linked to international collaborations with institutions in the U.S. and EU. Current ML efforts focus on technical optimization—such as precision irrigation, pest detection, and yield prediction—but fall short in addressing social equity and climate resilience. The study concludes that while ML holds significant promise for sustainable agri-food processing and system optimization, future progress depends on overcoming fragmented regional collaborations and integrating holistic frameworks, such as life-cycle assessment, to ensure resilient and equitable food systems.

The agri-food industry, a cornerstone of human existence and economic well-being, faces unprecedented challenges and opportunities in the 21st century. This chapter explores the profound influence of machine learning (CL) and computer vision (CV) technologies in addressing the critical demands of the agriculture and food sectors. Highlighting the industry’s global significance, we emphasize its pivotal role in safeguarding food supplies, fostering economic development, and addressing concerns related to nutrition, healthcare, and ecological sustainability. The pressures of a growing global population, resource constraints, and climate change necessitate innovative technological advancements in agriculture. ML and CV emerge as key drivers of this transformation, enabling evidence-based decision-making, precision agriculture, and continuous monitoring, thus ushering in an era of unmatched efficiency and productivity. Through examples from precision agriculture, crop management, and food quality assurance, this chapter demonstrates how these technologies can revolutionize farming practices, reduce crop-related risks, and enhance food safety standards. Offering valuable insights into the future of the agri-food industry, this chapter envisions a world where artificial intelligence, ML, and CV play crucial roles in creating a sustainable and thriving global food ecosystem. It serves as a vital resource for scholars, practitioners, and policymakers navigating the evolving landscape of technology-infused agriculture, shedding light on the path toward a more robust, efficient, and sustainable agri-food sector.

… Here, ΔTcorrected denotes the corrected temperature value after calibration, Traw represents the raw temperature reading from the sensor, and XGBoost(...) is the machine learning …

In recent years, the Industry 4.0 concept has gained considerable attention from professionals, researchers and decision makers. For its part, the COVID-19 pandemic has highlighted the importance of managing the agri-food supply chain to ensure the food that the population needs. Industry 4.0 and its extensions can address the needs of the agri-food supply chain by bringing new features such as security, transparency and traceability in line with sustainable development goals. This study aims to systematically analyze the literature to address the challenges and barriers against the application of industry 4.0 and its related technologies in the management of an agri-food supply chain. Currently, despite the large number of publications, there is no clear agreement on what Industry 4.0 is, and even less its extensions. The next revolution that includes new technologies and improves several existing technologies brings additional conceptual and practical complexity. Consequently, in this work we first determine the main components of I 4.0 and their extensions by studying the literature, and then, in the second step, define the agri-food supply chain on which I 4.0 technologies are applied. Two well-known databases—Web of Science and Scopus—were chosen to extract data for the systematic review of the literature. For the final evaluation, we identified 24 of 100 reviewed publications. The results provide an exhaustive analysis of the different I 4.0 technologies and their extensions that are applied in regards to the agri-food supply chain. In addition, we find 15 challenges that are classified into five major themes in the agri-food supply chain: technical, operational, financial, social and infrastructure. The four most important challenges identified are technological architecture, security and privacy, big data management and IoT (internet)-based infrastructure. Only a few articles addressed sustainability, which reaffirms and demonstrates a considerable gap in terms of the sustainable agri-food supply chain, with waste management being the one that has attracted the most attention. This review provides a roadmap for academics and practitioners alike, showing the gaps and facilitating the identification of I 4.0 technologies that can help address the challenges facing the efficient management of an agri-food supply chain.

The agri‐food sector is facing increasing sustainability pressures owing to environmental and socioeconomic issues and regulatory efforts in the European context. Given this effervescent background and the growing interest of academia and players in this important industry, a systematic literature review was conducted to organise the contributions of researchers to sustainability business practises and processes, investigate current and emerging trends, and highlight possible directions for future studies. Short food supply chain (SFSC) and circular economy (CE) models have emerged as the main thematic findings and promising strategies for improving sustainability from both the environmental and social perspectives. Moreover, digitalisation, including IoT, blockchain, AI, and machine learning, is crucial for implementing SFSC and CE practises, and consequently, sustainability in the agri‐food sector.

Approximately one-third of fresh food is wasted globally throughout the supply chain. Machine learning (ML), a key branch of artificial intelligence, enhances postharvest logistics and preservation of fresh food by enabling intelligent sensing, precise evaluation, and adaptive control. However, its application is challenged by data standardization, sensor limitations, product variability, and limited model generalizability. This review summarizes current advanced ML applications in the food supply chain, emphasizing their transformative potential for quality control. We explore ML's ability to integrate multi-omics data for deeper insights into molecular changes during transportation and storage, enabling the development and evaluation of management strategies. Practical applications in grading, sensor technology, and intelligent preservation materials are also evaluated. ML models, such as support vector machine (SVM) and convolutional neural network (CNN), enhance precise grading and quality prediction by analyzing sensory attributes and chemical composition. By capturing complex molecular interactions, ML enables innovative sensor surface design with enhanced sensitivity and specificity. ML-driven sensors further support real-time environmental monitoring, while intelligent packaging materials powered by ML maintain freshness and reduce spoilage through adaptive responses to internal conditions. To ensure model robustness and generalizability, appropriate validation strategies such as cross-validation and external validation are essential. Despite its substantial potential, the widespread adoption of ML still faces challenges, including limited varietal and regional generalization, decision-making transparency, computational demands, limited data availability, and algorithm selection. Addressing these issues is critical for achieving effective and sustainable ML integration in postharvest quality control systems.

<div class="section abstract"><div class="htmlview paragraph">In the food industry, there is an increased demand for generic pharmaceutical products and perishable food without compromising with the changes in texture and taste that occur in the transit. With this demand, there is a need for better visibility of products in the logistics network, to minimize wastage, to ensure product integrity, influence productivity, transparently track the fleet and to identify pathogens before a potential outbreak. In Cold Chain Management, information is power: with potentially billions of dollars’ worth of cargo (such as food items, vaccines, serums, tests or chemicals) at stake worldwide. Hence, careful live monitoring, inspection, supervision, validation and documentation of business-critical information is essential. In this paper, we have proposed a framework for Cold Chain Management using Internet of Things (IoT) combined with other technological innovations such as: Cloud Computing, Machine Learning and Big Data Analytics to revolutionize the cold transport industry. By establishing such an architecture, we have tried to monitor, visualize, track and control various platform dependent parameters thereby providing a complete solution across the fleet cycle with assured freshness and palpability.</div></div>

ABSTRACT Effective cold chain management is crucial for ensuring food safety, but traditional monitoring approaches remain reactive and fragmented. Industry 4.0 technologies, such as the Internet of Things (IoT) and artificial intelligence (AI), offer transformative potential for real‐time tracking, automation, and predictive analytics. Many studies have explored the application of IoT and AI in the cold chain, but the existing literature focuses on isolated technologies. This review presents an integrated perspective, examining the synergistic potential of IoT‐enabled data acquisition and AI‐driven predictive analytics in the cold chain. We reviewed 97 peer‐reviewed papers published between 2010 and 2025, following PRISMA guidelines, covering IoT sensors, AI applications, and their integration across the food cold supply chain. Our analysis reveals a rapid growth in IoT and AI adoption, driven by regulatory and consumer demands for transparency, quality, and predictive risk assessment. IoT sensors enable real‐time monitoring, providing early detection of potential safety risks. AI‐powered models process sensor data to predict temperature deviations, assess food safety, and optimize logistics, reducing spoilage and contamination risks. We also highlight current limitations and future research directions, such as the limited number of studies on closed‐loop systems, where IoT sensors provide real‐time data and AI models respond dynamically. This review provides a comprehensive resource for selecting IoT and AI systems to enhance food safety and ensure a more resilient, transparent, and sustainable cold chain.

The assessment of tomato firmness is pivotal in determining optimal harvest time, evaluating shelf life, and gauging ripeness. This attribute plays a crucial role in guiding the distribution and transportation processes. Post-harvest, tomatoes tend to lose firmness and can deteriorate into a rotten state during transportation within the supply chain, mainly due to environmental fluctuations. To mitigate such losses and uphold tomato quality, the cold supply chain, with its controlled environmental conditions, proves instrumental. Monitoring this cold supply chain is imperative to combat the adverse impact of ambient temperatures on tomatoes during logistics. This research introduces an innovative approach, employing an Internet of Things (IoT) framework and the Whale Optimization Algorithm for temperature prediction within the cold supply chain. Ambient and tomato temperatures, along with stable temperature calculations under variable conditions using the Whale Optimization Algorithm, were collected. The predictions were executed using the Extreme Learning Machine of Artificial Intelligence. The data is collected during tomato cold storage for experimentation. The proposed technique with mean average precision 84.957%, mean average recall 96.9% and accuracy 99.83%. Evaluation through precision, recall, and F-measure accuracy metrics demonstrates the superior performance of the proposed approach compared to conventional models such as Decision Tree, Linear Model, Naïve Bays, Random Forest, and Support Vector Machine.

The pharmaceutical cold chain sector in the United States brings in $100 billion in revenue but temperature violations cost $50 billion a year and endanger patient safety. Machine learning (ML) is studied in this research to boost both compliance rates and quality performance in cold chain logistics. Our study using both stakeholder interviews and Narrative case study observations on MediCool distributor alongside Random Forest and LSTM Model evaluations demonstrated how the ML approach decreased temperature deviations by 25% at the same time it cut expenses by 20% and exceeded 98% of FDA compliance standards for a period of twelve months. The approach enables 150 million Americans to obtain equal drug access due to its integration of IoT sensors, agile workflows, and human-centered dashboards. The proposed framework targets mid-sized companies by breaking down training limitations and outdated systems to advance AI capabilities which strengthen U.S. healthcare quality and fairness.

PurposeIn the cold supply chain (SC), effective risk management is regarded as an essential component to address the risky and uncertain SC environment in handling time- and temperature-sensitive products. However, existing multi-criteria decision-making (MCDM) approaches greatly rely on expert opinions for pairwise comparisons. Despite the fact that machine learning models can be customised to conduct pairwise comparisons, it is difficult for small and medium enterprises (SMEs) to intelligently measure the ratings between risk criteria without sufficiently large datasets. Therefore, this paper aims at developing an enterprise-wide solution to identify and assess cold chain risks.Design/methodology/approachA novel federated learning (FL)-enabled multi-criteria risk evaluation system (FMRES) is proposed, which integrates FL and the best–worst method (BWM) to measure firm-level cold chain risks under the suggested risk hierarchical structure. The factors of technologies and equipment, operations, external environment, and personnel and organisation are considered. Furthermore, a case analysis of an e-grocery SC in Australia is conducted to examine the feasibility of the proposed approach.FindingsThroughout this study, it is found that embedding the FL mechanism into the MCDM process is effective in acquiring knowledge of pairwise comparisons from experts. A trusted federation in a cold chain network is therefore formulated to identify and assess cold SC risks in a systematic manner.Originality/valueA novel hybridisation between horizontal FL and MCDM process is explored, which enhances the autonomy of the MCDM approaches to evaluate cold chain risks under the structured hierarchy.

Ensuring vaccine integrity requires rigorous cold-chain management, as even minor temperature deviations can compromise safety, cause financial losses, and endanger public health. Conventional monitoring systems, dependent on centralized databases and manual oversight, remain prone to delays, inaccuracies, and tampering. To address these limitations, we propose an integrated framework that combines deep learning and blockchain for secure, predictive, and real-time cold-chain monitoring. A distributed network of IoT sensors captures temperature, humidity, vibration, and GPS data at one-minute intervals. Long Short-Term Memory (LSTM) networks forecast short-term temperature trends, Autoencoders (AE) detect anomalies, and one-dimensional Convolutional Neural Networks (1D CNN) classify shipment states as Safe, At Risk, or Spoiled. A decision engine fuses outputs from these models into a unified risk score, enabling timely and data-driven interventions. Critical events and metadata are immutably recorded on the Ethereum blockchain, while raw sensor data is stored off-chain using IPFS to ensure auditability and efficiency. The system also supports automated alerts, real-time monitoring through Grafana dashboards, and adaptive model retraining for continuous improvement. Evaluation covers forecasting accuracy, anomaly detection precision, classification performance, and blockchain efficiency. By uniting predictive analytics with tamper-proof logging, the proposed framework reduces vaccine spoilage, strengthens supply chain resilience, and offers a scalable solution for broader temperature-sensitive logistics applications.

Abstract Radio Frequency Identification (RFID) technology has significantly improved in the past few years and is presently sought for implementation in the identification and traceability of perishable food in the food sector to safeguard food safety and quality. It is currently considered a worthy successor to the barcode system and has significant advantages for monitoring products in the perishable food supply chain (PFSC). The present study proposes a traceability system that utilizes RFID and Internet of Things (IoT) sensors. RFID technology can be used to track and trace perishable food while IoT sensors can be used to measure temperature and humidity during storage and transportation. Furthermore, it is important that RFID gates can identify the direction of tags and whether products are being received or shipped through the gate. In this study, machine-learning models are utilized to detect the direction of passive RFID tags. The input features are derived from receive signal strength (RSS) and the timestamp of tags. The proposed system has been tested in the perishable food supply chain and has revealed significant benefits to managers and customers by providing real-time product information and complete temperature and humidity history. In addition, by integrating a machine-learning model into the RFID gate, tagged products that move in or out through a gate can be correctly identified and thus improve the efficiency of the traceability system.

Postharvest losses of perishable commodities remain a major global challenge due to inefficiencies in conventional cold chain systems, necessitating intelligent, adaptive technologies for improved preservation. This study aims to develop a machine learning framework using an Artificial Neural Network (ANN) to optimize refrigeration performance for vegetable preservation, with specific objectives to model nonlinear interactions among operational parameters, improve prediction accuracy for quality indicators, and identify optimal operating conditions that balance energy use and product longevity. Using a 30-run Design of Experiments (DOE) dataset, the ANN was trained in Python (TensorFlow/Keras) with inputs including evaporator temperature, cooling duration, insulation thickness, airflow rate, and storage load, and outputs comprising coefficient of performance (COP), moisture loss, energy consumption, and shelf life. Model evaluation using MSE, RMSE, MAE, and R² revealed inconsistent performance, with some outputs initially achieving high predictive accuracy while later metrics showed negative R² for moisture loss and shelf life, indicating overfitting and data limitations; however, feature importance analysis and 3-D response surfaces confirmed meaningful nonlinear relationships, and optimization results identified settings such as −3.69°C evaporator temperature and moderate storage load for maximizing COP, and −2.14°C for minimizing moisture loss. The findings demonstrate the potential of ANN in multi-objective cold-chain optimization but highlight the need for larger datasets, improved network regularization, and integration with IoT-based real-time monitoring. It is therefore recommended that future work employ expanded experimental data, robust ANN architectures, and sensor-driven dynamic updating to enhance generalization and practical deployment in commercial cold-chain environments.

The management of a food supply chain is difficult and complex because of the product's short shelf-life, time-sensitivity, and perishable nature which must be carefully considered to minimize food waste. Temperature-controlled perishable food supply chain provides the highly crucial facilities necessary to maintain the quality and safety of the product. The storage temperature is the most vital factor in maintaining both the quality and shelf-life of a perishable food. Adequate storage temperature control ensures that perishable foods are transported to the end-users in good quality and safe to consume. This paper presents perishable food storage temperature control through mathematical optimal control model where the storage temperature is regarded as the control variable and the deterioration of the perishable food’s quality follows the first-order reaction. The optimal storage temperature for a single perishable food is determined by applying the Pontryagin's maximum principle to solve the optimal control model problem. For multi-temperature commodities supply chain, an unsupervised machine learning (ML) method, called k-means clustering technique is used to determine the temperature clusters for a range of perishables. Based on descriptive analysis, it is observed that the k-means clustering technique is effective in identifying the best suitable storage temperature clusters for quality control of multi-commodity supply chain.

Cold chain logistics play an important role in ensuring the quality and safety of temperature sensitive goods such as vaccines, pharmaceuticals and perishable foods. This paper presents the implementation and architecture of an IoT-based cold chain monitoring system that integrates sensors, cloud computing and predictive analytics to ensure real-time tracking and proactive decision-making. The designed system uses sensors to monitor humidity, temperature, vibration and GPS location all connected to ESP32 micro-controller to provide real-time data collection, automated alerts and remote monitoring. Predictive analytics equipped with historical sensor data and trend analysis, forecasts potential threshold breaches of humidity and temperature enhancing the reliability of the system. The results of all mentioned above are presented in results section. For temperature forecasting, the LSTM model achieved a Mean Absolute Error (MAE) of 2.59° C, indicating that, on average, the predicted temperatures deviate by 2.59° C from actual values. The Root Mean Squared Error (RMSE) of 2.64° C suggests that while larger deviations exist, they remain close to the MAE, highlighting a well-balanced prediction model. Similarly, for humidity forecasting, the model attained an MAE of 0.57 %, meaning the predicted humidity levels differ by 0.57 % on average, while an RMSE of 0.65 % reflects minimal extreme variations, confirming the model's reliability in forecasting both temperature and humidity over baseline systems. The proposed solution addresses the limitations of traditional cold chain systems, such as delayed responses, data inaccuracies, and lack of predictive maintenance. The proposed system offers a scalable and cost-effective solution for modern cold chain management. Existing systems focus only on realtime monitoring, but the proposed system integrates predictive analytics with low-cost IoT hardware, enabling proactive interventions.

Fresh fruits and vegetables are highly perishable due to their short shelf life and sensitivity to temperature fluctuations. To solve this problem, this study compares the predictive performance of traditional machine learning models such as XGBoost (XGB), and Extreme Learning Machine (ELM) and deep learning models such as Gated Recurrent Units (GRU) and Multilayer Perceptron (MLP) for temperature. Optimal parameters were found using Grid Search and Random Search. Performance was evaluated using metrics such as MSE, MAE and $\mathrm{R}^{2}$. Networking sensors captured data, including internal humidity, temperature, and $\text{CO}_{2}$ levels, from a cold room storing fresh apples in the Western Cape, South Africa. The data was divided using TimeSeriesSplit, with different proportions for each model. GRU, XGB, and MLP models used an 88 % training and 12 % testing split, while ELM used a 75 % training and 25 % testing split. The models showed strong performance, achieving $\mathrm{R}^{2}$ values between 0.9669 and 0.9914, MAE between 0.0562 and 0.1343, and MSE between 0.0119 and 0.1029, all with 95 % confidence. Statistical tests ($\mathrm{t}$ - test and Wilcoxon signed-rank test) show no significant differences in performance between MLP and ELM predictions. These findings offer valuable insights for improving cold chain management and reducing food waste and its associated impacts. It is crucial to incorporate continuous learning capabilities into these models to adapt to dynamic environments like the cold chain. Developing models for other phases, such as refrigerated transport and retail storage, is also recommended for future efforts.

Vaccines are a unique category of drugs sensitive to temperature and humidity and whose effectiveness directly impacts public health. There has been an increase in vaccine-related adverse events worldwide, particularly in developing countries, attributed to suboptimal temperatures during transport and storage. At the same time, the Internet of Things (IoT) has ushered in a paradigm shift in vaccine information and storage monitoring, enabling continuous 24/7 tracking. This further reduces the dependence on limited human resources and significantly reduces the associated errors and losses. This paper presents an IoT-driven framework that aims to improve the sustainability of medical cold chain management. The framework promotes trust and transparency in vaccine surveillance data by accessing and authenticating IoT devices. The proposed system aims to improve the safety and sustainability of vaccine management. Moreover, we provide detailed insights into the design and hardware components of the proposed framework. In addition, the specific use of the framework in a particular province is highlighted, covering the design of the software platform and the analysis of the hardware equipment.

… application of sensors and other IoT devices. For instance, the … to provide insight into the digital cold chain of strawberries and … Ultimately, a first-expired-first-out distribution chain can …

Agriculture and livestock play a vital role in social and economic stability. Food safety and transparency in the food supply chain are a significant concern for many people. Internet of Things (IoT) and blockchain are gaining attention due to their success in versatile applications. They generate a large amount of data that can be optimized and used efficiently by advanced deep learning (ADL) techniques. The importance of such innovations from the viewpoint of supply chain management is significant in different processes such as for broadened visibility, provenance, digitalization, disintermediation, and smart contracts. This article takes the secure IoT–blockchain data of Industry 4.0 in the food sector as a research object. Using ADL techniques, we propose a hybrid model based on recurrent neural networks (RNN). Therefore, we used long short-term memory (LSTM) and gated recurrent units (GRU) as a prediction model and genetic algorithm (GA) optimization jointly to optimize the parameters of the hybrid model. We select the optimal training parameters by GA and finally cascade LSTM with GRU. We evaluated the performance of the proposed system for a different number of users. This paper aims to help supply chain practitioners to take advantage of the state-of-the-art technologies; it will also help the industry to make policies according to the predictions of ADL.

Cold chain logistics plays a critical role in preserving the quality and safety of temperature-sensitive products in the food, pharmaceutical, and chemical industries. This study proposes a …

… Therefore, it is crucial to maintain the cold chain during the transportation and storage of … Subsequently, the machine learning model operates on the cloud to predict the remaining …

ABSTRACT This research investigates how Artificial Intelligence (AI) and Machine Learning (ML) forecasting methodologies can be leveraged for cold chain capacity planning, specifically utilising Prophet and Seasonal Autoregressive Integrated Moving Average parametrised through grid search. In collaboration with Americold, the world's second-largest refrigerated logistic service provider, the study explores the challenges and opportunities in applying AI/ML techniques to complex operations covering 385 customers and a capacity of 73,296 pallet positions. We train and test several AI/ML and traditional statistical models using extensive data for every customer over 3.5 years. Based on the results, MAPE of 5.28% was achieved on the whole site level, and SARIMA outperformed ML models in most cases. Next, we show that developing and applying a Customer Segmentation Matrix has enabled more accurate forecasting and planning across various customer segments, addressing the issue of forecasting inaccuracies. This approach effectively improves forecasting inaccuracies, underscoring the significance of tailoring AI/ML models for demand forecasting within the cold-chain industry. Ultimately, this research presents an AI-driven approach that transcends mere forecasting, offering a practical pathway to manage capacity in light of the constraints.

Accurate demand forecasting is crucial for reducing food waste, enhancing resilience, and promoting sustainability in food systems. Relying only on historical sales data is inadequate to predict the demand for perishable food products. The remaining shelf-life of food products and their qualities need to be considered as they play a vital role in the forecasting process, leading to accurate and reliable predictions. This paper proposes a four-stage conceptual framework to predict the daily demand for fresh food items, incorporating two critical variables: health class and remaining shelf life. The framework is applied to a real industry case to monitor the freshness quality of bananas, and important variables are collected during run-time. The collected data are then processed and used to train several classification models to classify the health class of bananas into either fresh, ripening, or spoiled. Not very surprisingly as many other application domains, results reveal that the random forest model outperforms other models in predicting the health class of bananas placed in boxes, achieving about 91 % prediction accuracy. However, this paper presents a first research of its kind with the abundance of real data, and specific comprehensiveness and focus on the overdue industrial problem.

Purpose The purpose of this study is to select the appropriate forecasting model at the retail stage for selected vegetables on the basis of performance analysis. Design/methodology/approach Various forecasting models such as the Box–Jenkins-based auto-regressive integrated moving average model and machine learning-based algorithms such as long short-term memory (LSTM) networks, support vector regression (SVR), random forest regression, gradient boosting regression (GBR) and extreme GBR (XGBoost/XGBR) were proposed and applied (i.e. modeling, training, testing and predicting) at the retail stage for selected vegetables to forecast demand. The performance analysis (i.e. forecasting error analysis) was carried out to select the appropriate forecasting model at the retail stage for selected vegetables. Findings From the obtained results for a case environment, it was observed that the machine learning algorithms, namely LSTM and SVR, produced the better results in comparison with other different demand forecasting models. Research limitations/implications The results obtained from the case environment cannot be generalized. However, it may be used for forecasting of different agriculture produces at the retail stage, capturing their demand environment. Practical implications The implementation of LSTM and SVR for the case situation at the retail stage will reduce the forecast error, daily retail inventory and fresh produce wastage and will increase the daily revenue. Originality/value The demand forecasting model selection for agriculture produce at the retail stage on the basis of performance analysis is a unique study where both traditional and non-traditional models were analyzed and compared.

The fresh produce supply chain sector is a vital pillar of any society and an indispensable part of the national economic structure. As a significant segment of the agricultural market, accurately forecasting vegetable prices holds significant importance. Vegetable market pricing is subject to a myriad of complex influences, resulting in nonlinear patterns that conventional time series methodologies often struggle to decode. Future planning for commodity pricing is achievable by forecasting the future price anticipated by the current circumstances. This paper presents a price forecasting methodology for tomatoes which uses price and production data taken from 2008 to 2021 and analyzed by means of advanced deep learning-based Long Short-Term Memory (LSTM) networks. A comparative analysis of three models based on Root Mean Square Error (RMSE) identifies LSTM as the most accurate model, achieving the lowest RMSE (0.2818), while SARIMA performs relatively well. The proposed deep learning-based method significantly improved the results versus other conventional machine learning and statistical time series analysis methods.

In the complex landscape of fresh food retail, accurate demand forecasting stands as a critical challenge, pivotal to both financial optimization and sustainable practices. This paper introduces an advanced approach to daily retail order predictions by applying multiple machine learning paradigms: Long Short-Term Memory (LSTM) networks, Transformers, Prophet, Feedforward Neural Networks (FNN), and eXtreme Gradient Boosting (XGBoost). Leveraging LSTM’s adeptness at processing sequential sales data, we harness its capability to discern temporal patterns and long-term dependencies inherent in the sales trajectory. Transformers can handle longer dependencies, which have been shown to be beneficial for our yearly seasonal data. On the other hand, XGBoost, with its robustness against overfitting and ability to capture nonlinear relationships, provides a granular understanding of the domain-related factors. FNNs are included as a strong baseline, and the included residual connections have proven crucial due to our sales data’s correlation with the next day’s sales being the most pronounced. Through experimentations with a dataset of sales data from the past seven years, our research highlights the strength of these methodologies. The results reveal that all models perform in acceptable ranges, however, Transformers, XGBoost and FNN models outperform LSTMs. Additionally, XGBoost exhibits slightly lower variation, which may be preferable.

The current review examines the state of knowledge and research on machine learning (ML) applications in horticultural production and the potential for predicting fresh produce losses and waste. Recently, ML has been increasingly applied in horticulture for efficient and accurate operations. Given the health benefits of fresh produce and the need for food and nutrition security, efficient horticultural production and postharvest management are important. This review aims to assess the application of ML in preharvest and postharvest horticulture and the potential of ML in reducing postharvest losses and waste by predicting their magnitude, which is crucial for management practices and policymaking in loss and waste reduction. The review starts by assessing the application of ML in preharvest horticulture. It then presents the application of ML in postharvest handling and processing, and lastly, the prospects for its application in postharvest loss and waste quantification. The findings revealed that several ML algorithms perform satisfactorily in classification and prediction tasks. Based on that, there is a need to further investigate the suitability of more models or a combination of models with a higher potential for classification and prediction. Overall, the review suggested possible future directions for research related to the application of ML in postharvest losses and waste quantification.

Adequately priced orders and time for fresh produce (FP) are two factors that bring commercial benefits to vendors and minimizes waste. However, many factors, such as income, labor, and other trade issues, affect the price that include uncertainties due to climate change, making decisions on FP procurement prices and quantities extremely challenging. Two artificial intelligence-based forecasting tools, i.e., a single variate and a multivariate model, are trained, tested, and compared in this study to predict future daily offer prices up to 7 days ahead for strawberries using mutual transactions for the distribution centers of Loblaws Companies Limited (LCL) in Canada. Results reveal that the developed multivariate model, utilizing both prices of the LCL dataset and California's strawberries yield dataset as predictors, outperforms the best single variate model.

… novel ML target prediction algorithm to inform the farmers about the market target product and … Producers can sell fresh pulses and buy processed pulses for family consumption at these …

This study presents a novel hybrid model that combines two different algorithms to increase the accuracy of short-term berry yield prediction using only previous yield data. The model integrates both autoregressive integrated moving average (ARIMA) with Kalman filter refinement and neural network techniques, specifically support vector regression (SVR), and nonlinear autoregressive (NAR) neural networks, to improve prediction accuracy by correcting the errors generated by the system. In order to enhance the prediction performance of the ARIMA model, an innovative method is introduced that reduces randomness and incorporates only observed variables and system errors into the state-space system. The results indicate that the proposed hybrid models exhibit greater accuracy in predicting weekly production, with a goodness-of-fit value above 0.95 and lower root mean square error (RMSE) and mean absolute error (MAE) values compared with non-hybrid models. The study highlights several implications, including the potential for small growers to use digital strategies that offer crop forecasts to increase sales and promote loyalty in relationships with large food retail chains. Additionally, accurate yield forecasting can help berry growers plan their production schedules and optimize resource use, leading to increased efficiency and profitability. The proposed model may serve as a valuable information source for European food retailers, enabling growers to form strategic alliances with their customers.

With the continuous expansion of fresh agricultural products market and the diversification of consumer demand, the operational efficiency of cold chain logistics and the accuracy of demand forecast have become the key to the development of the industry. Aiming at the limitations of traditional forecasting methods in dealing with complex nonlinear data relations, this paper proposes a hybrid model (CNN-LSTM) combining Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN) to forecast the demand of cold chain logistics. This model uses CNN's powerful feature extraction ability to capture local features of data and uses LSTM's time series processing ability to remember historical information to predict future demand. The experimental results show that compared with the traditional forecasting methods, CNN-LSTM hybrid model is more accurate in the overall trend forecasting, especially in the peak sales season and the product types with stable demand. However, in the off-season of sales and the product types with large fluctuations in demand, the prediction error increases slightly. The research in this paper provides new ideas and methods for the accurate prediction of cold chain logistics demand of fresh agricultural products, which is helpful to improve the operational efficiency and market responsiveness of cold chain logistics.

… model to predict the yield of a characteristic regional fruit (the … field of demand prediction based on deep learning is extensive… A scientific and accurate prediction of fruit production is of …

Expired foods turning into waste has always been an important issue. In Taiwan, more than 36,000 metric tons of unopened expired food, worth more than $130 million, are thrown away from retail stores as waste each year. Insufficient inventory results in the loss of business prospects for retailers, whilst excessive inventory results in abandoned merchandise. Foods with a short shelf life are particularly vulnerable. Typically, food producer and retailer would form team merchandising (MD). The team MD mechanism is responsible for ensuring safety and quality, not for forecasting demand. This study uses artificial neural networks (ANNs) to analyze sales data to forecasting purchase volume in response to changes in store environment, weather, events, and consumer attributes. The study object is a sort of cream puff with a short shelf life cobranded by a retailer. According to the experimental results, the adopted proposed model in this study effectively reduces the error in purchasing; the mean-square percentage error (MSPE) of the forecast values is less than 6%. The importance of this study is on promoting the team MD’s green energy management capabilities in food production and verifiably achieving the goal of environmental sustainability.

Food waste in perishable supply chains represents a critical sustainability challenge, with approximately one-third of global food production lost annually. Traditional demand forecasting methods often fail to capture the complex patterns inherent in perishable goods, leading to suboptimal inventory decisions that increase waste. This paper proposes a machine learning-based demand forecasting framework specifically designed to reduce food waste by addressing stockout-induced demand censoring and incorporating contextual factors. Using the FreshRetailNet-50K dataset comprising 382,500 training records from 898 stores and 863 products, we implement a comprehensive pipeline that recovers latent demand during stockout periods and applies advanced forecasting models including Random Forest, XGBoost, and Long Short-Term Memory (LSTM) networks. Our Random Forest model achieves exceptional performance with 99.35% accuracy, demonstrating 98.79% service level and 98.79% waste reduction potential. The proposed framework provides actionable insights for sustainable supply chain management by significantly improving inventory optimization while maintaining operational profitability.

… these losses by enhancing food spoilage predictions and optimizing supply chain … machine learning models, predictive analytics, and advanced algorithm in predicting food spoilage …

Food spoilage is a global problem which causes food waste, economic loss and foodborne illness. The shelf life and spoilage estimation of food is traditionally done with fixed expiration dates and this leads to disposal of still eatable food or eating spoiled food. Recently, with the development of the Artificial Intelligence (AI), the predictive models have been developed to better evaluate the food spoilage based on such factors as temperature, humidity, microbial activities and gas emissions. This paper discusses the part played by AI in the prediction of food spoilage, while also outlining various machine learning and deep learning models (regression, classification, convolutional neural network – CNN and hybrid AI). Food spoilage estimation powered by AI relies on multiple sources of data including IoT enabled sensors, Spectroscopy as well as real time environmental monitoring. The practical use in the food industry of such data driven models is in the context of real life applications as smart packaging, AI powered quality in supply chains, retail inventory product optimization. However, the adoption of AI in this field is limited as the data is scarce and of low quality, the models have limited accuracy, ethical concerns exist, and implementation is expensive. In this review, potential for AI in transforming food spoilage estimation is highlighted and this could be achieved by working on obtaining greater accuracy, scalability, and adoption of the model in different food sectors. The role of AI in enhancing food security, sustainability and efficient use of resources, waste reduction and increasing accessibility of good quality perishables to every consumer will gain increasing feasibility with the improvement in AI.

Over the past decade, analytical approaches based on vibrational spectroscopy, hyperspectral/multispectral imagining and biomimetic sensors started gaining popularity as rapid and efficient methods for assessing food quality, safety and authentication; as a sensible alternative to the expensive and time-consuming conventional microbiological techniques. Due to the multi-dimensional nature of the data generated from such analyses, the output needs to be coupled with a suitable statistical approach or machine-learning algorithms before the results can be interpreted. Choosing the optimum pattern recognition or machine learning approach for a given analytical platform is often challenging and involves a comparative analysis between various algorithms in order to achieve the best possible prediction accuracy. In this work, "MeatReg", a web-based application is presented, able to automate the procedure of identifying the best machine learning method for comparing data from several analytical techniques, to predict the counts of microorganisms responsible of meat spoilage regardless of the packaging system applied. In particularly up to 7 regression methods were applied and these are ordinary least squares regression, stepwise linear regression, partial least square regression, principal component regression, support vector regression, random forest and k-nearest neighbours. MeatReg" was tested with minced beef samples stored under aerobic and modified atmosphere packaging and analysed with electronic nose, HPLC, FT-IR, GC-MS and Multispectral imaging instrument. Population of total viable count, lactic acid bacteria, pseudomonads, Enterobacteriaceae and B. thermosphacta, were predicted. As a result, recommendations of which analytical platforms are suitable to predict each type of bacteria and which machine learning methods to use in each case were obtained. The developed system is accessible via the link: www.sorfml.com.

Machine learning assists with food process optimization techniques by developing a model to predict the optimal solution for given input data. Machine learning includes unsupervised and supervised learning, data pre-processing, feature engineering, model selection, assessment, and optimization methods. Various problems with food processing optimization could be resolved using these techniques. Machine learning is increasingly being used in the food industry to improve production efficiency, reduce waste, and create personalized customer experiences. Machine learning may be used to improve ingredient utilization and save costs, automate operations such as packing and labeling, and even forecast consumer preferences to develop personalized products. Machine learning is also being used to identify food safety hazards before they reach the consumer, such as contaminants or spoiled food. The usage of machine learning in the food sector is predicted to rise in the near future as more businesses understand the potential of this technology to enhance customer experience and boost productivity. Machine learning may be utilized to enhance nano-technological operations and fruit and vegetable preservation. Machine learning algorithms may find trends regarding various factors that impact the quality of the product being preserved by examining data from prior tests. Furthermore, machine learning may be utilized to determine optimal parameter combinations that result in maximal produce preservation. The review discusses the relevance of machine learning in ready-to-eat foods and its use as a safety tool for preservation were highlighted. The application of machine learning in agriculture, food packaging, food processing, and food safety is reviewed. The working principle and methodology, as well as the principles of machine learning, were discussed.

Food rotting is an essential concern contributing to significant food waste, and novel early detection and prevention methods are needed. Existing manual monitoring procedures are generally time-consuming and prone to errors, resulting in spoiling detection delays. The study presents an IoT-based food spoilage detection system to solve these issues. Using IoT sensors and machine learning (ML) algorithms, the proposed System provides continuous monitoring and predictive analysis of essential variables such as temperature, humidity, gas levels, and food availability. The proposed System uses real-time monitoring and alerting to enable proactive interventions to prevent spoilage, decrease waste, and promote sustainability throughout the food supply chain. The performance evaluation shows that the proposed System outperforms existing systems in terms of accuracy (0.92), precision (0.90), recall (0.89), and F1-score (0.91). Overall, the IoT-based Food Spoilage Detection System provides an integrated strategy for reducing food waste and increasing sustainability, representing a big step forward in food waste reduction initiatives.

Based on both new and previously utilized experimental data, the present study provides a comparative assessment of sensors and machine learning approaches for evaluating the microbiological spoilage of ready-to-eat leafy vegetables (baby spinach and rocket). Fourier-transform infrared (FTIR), near-infrared (NIR), visible (VIS) spectroscopy and multispectral imaging (MSI) were used. Two data partitioning approaches and two algorithms, namely partial least squares regression and support vector regression (SVR), were evaluated. Concerning baby spinach, when model testing was performed on samples randomly selected, the performance was better than or similar to the one attained when testing was performed based on dynamic temperatures data, depending on the applied analytical technology. The two applied algorithms yielded similar model performances for the majority of baby spinach cases. Regarding rocket, the random data partitioning approach performed considerably better results in almost all cases of sensor/algorithm combination. Furthermore, SVR algorithm resulted in considerably or slightly better model performances for the FTIR, VIS and NIR sensors, depending on the data partitioning approach. However, PLSR algorithm provided better models for the MSI sensor. Overall, the microbiological spoilage of baby spinach was better assessed by models derived mainly from the VIS sensor, while FTIR and MSI were more suitable in rocket. According to the findings of this study, a distinct sensor and computational analysis application is needed for each vegetable type, suggesting that there is not a single combination of analytical approach/algorithm that could be applied successfully in all food products and throughout the food supply chain.

Food spoilage is a critical concern in the industry, leading to economic losses and food safety issues. We propose a comprehensive system that leveraging the synergy of Internet of Things (IoT) and Machine learning. By deploying IoT sensors across the food supply chain to collect environmental data, such as temperature and humidity, and then processing this data with machine learning algorithms, we can promptly identify spoilage patterns and anomalies. This approach enables early spoilage detection, reducing food waste, enhancing food safety, and optimizing supply chain management. Ultimately, it offers a comprehensive and cost-effective solution to the food industry's spoilage-related challenges, benefitting both businesses and consumers.

To control spoilage by lactic acid bacteria (Leuconostoc spp.) in cooked deli food, various combinations of environmental and/or intrinsic factors have been employed based on hurdle technology. Since many factors and their combinations greatly influence Leuconostoc spp. growth, this study aimed to develop a machine learning model based on the experimentally obtained growth kinetic data using extreme gradient boosting tree algorithm to quantitatively and flexibly predict Leuconostoc spp. growth. In particular, effects of sodium acetate (0-1.5%) and glycine (0-1.5%), which are frequently used food additives in the Japanese food industry, on the growth of Leuconostoc spp. in cooked deli foods were examined with combination of temperature (5-25 °C) and pH (5.0-6.0) conditions. The developed machine learning model to predict the number of Leuconostoc spp. over time successfully demonstrates comparable accuracy in culture media to the conventional Baranyi model based prediction. Furthermore, while the accuracy of the prediction by the machine learning model for cooked deli foods such as potato salad, Japanese simmered hijiki, and unohana evaluated by the proportion of relative error within the acceptable prediction range was 98%, the accuracy of the conventional Baranyi model based prediction was 89%. The developed machine learning model successfully and flexibly predicted the growth of Leuconostoc spp. in various cooked deli foods incorporating the effect of food additives, with an accuracy comparable to or better than that of the conventional kinetic-based model.

Ensuring a safe and adequate food supply is a cornerstone of human health and food security. However, a significant portion of the food produced for human consumption is wasted annually on a global scale. Reducing harvest and postharvest food waste, waste during food processing, as well as food waste at the consumer level, have been key objectives of improving and maintaining sustainability. These issues can range from damage during processing, handling, and transport, to the use of inappropriate or outdated systems, and storage and packaging-related issues. Microbial growth and (cross)contamination during harvest, processing, and packaging, which causes spoilage and safety issues in both fresh and packaged foods, is an overarching issue contributing to food waste. Microbial causes of food spoilage are typically bacterial or fungal in nature and can impact fresh, processed, and packaged foods. Moreover, spoilage can be influenced by the intrinsic factors of the food (water activity, pH), initial load of the microorganism and its interaction with the surrounding microflora, and external factors such as temperature abuse and food acidity, among others. Considering this multifaceted nature of the food system and the factors driving microbial spoilage, there is an immediate need for the use of novel approaches to predict and potentially prevent the occurrence of such spoilage to minimize food waste at the harvest, post-harvest, processing, and consumer levels. Quantitative microbial spoilage risk assessment (QMSRA) is a predictive framework that analyzes information on microbial behavior under the various conditions encountered within the food ecosystem, while employing a probabilistic approach to account for uncertainty and variability. Widespread adoption of the QMSRA approach could help in predicting and preventing the occurrence of spoilage along the food chain. Alternatively, the use of advanced packaging technologies would serve as a direct prevention strategy, potentially minimizing (cross)contamination and assuring the safe handling of foods, in order to reduce food waste at the post-harvest and retail stages. Finally, increasing transparency and consumer knowledge regarding food date labels, which typically are indicators of food quality rather than food safety, could also contribute to reduced food waste at the consumer level. The objective of this review is to highlight the impact of microbial spoilage and (cross)contamination events on food loss and waste. The review also discusses some novel methods to mitigate food spoilage and food loss and waste, and ensure the quality and safety of our food supply.

Accurate prediction of food shelf life is critical for ensuring food safety, reducing waste, and delivering reliable products to consumers. Deep learning, as an advanced artificial intelligence (AI) technology, provides transformative solutions for shelf life prediction. This paper systematically reviews research advances in deep learning applications for food shelf life prediction and regulation, including examinations of predictive model architectures, analyses of food quality assessment criteria, explorations of hybrid methods integrating data-driven and mechanistic approaches, and proposals for model-informed optimization strategies. With advancing AI, deep learning will further strengthen food safety systems, enhance resource efficiency, reduce waste, and modernize perishable goods supply chains. PRACTICAL APPLICATIONS: This paper explores the research and application of deep learning in the field of food shelf life, focusing on three key areas: common deep learning models, methods for evaluating preservation quality, and shelf-life prediction techniques. Additionally, the paper introduces the concept of reverse regulation of shelf life, offering innovative solutions for ensuring food safety, enhancing production efficiency, and integrating intelligent supply chain logistics.

A critical engineering analysis is provided regarding the applications of Artificial Intelligence (AI) and Deep Learning (DL) in advancing perishable food shelf life management. …

Ensuring food safety and quality during transportation remains a critical challenge in smart agricultural systems. Traditional empirical models fail to capture the nonlinear relationships between environmental variables and food quality degradation. This paper presents a comprehensive machine learning framework for predicting food spoilage rates under dynamic environmental conditions. We comparatively evaluate eight representative algorithms-Decision Tree, Random Forest, Gradient Boosting, AdaBoost, SVR, KNN, XGBoost, and Light-GBM-on a real-world dataset of $\mathbf{5, 0 0 0}$ transportation records. Feature engineering integrates environmental and temporal features to model nonlinear spoilage dynamics. Experimental results demonstrate that ensemble methods, particularly LightGBM (MAE $=0.0265$) and XGBoost (MAE $=0.0283$), achieve superior predictive performance with $20 \%$ improvement over Decision Tree baselines. The proposed approach highlights the potential of machine learning in intelligent quality monitoring and predictive modeling, contributing to data-driven decision support for food safety and sustainable storage management.

… using Machine Learning(ML) allows food supply chains to change their strategy to quality control to one that proactively minimizes spoilage by … effects on spoilage prediction models in …

For the agricultural food production sector, the control and assessment of food quality is an essential issue, which has a direct impact on both human health and the economic value of the product. One of the fundamental properties from which the quality of the food can be derived is the smell of the product. A significant trend in this context is machine olfaction or the automated simulation of the sense of smell using a so-called electronic nose or e-nose. Hereby, many sensors are used to detect compounds, which define the odors and herewith the quality of the product. The proper assessment of the food quality is based on the correct functioning of the adopted sensors. Unfortunately, sensors may fail to provide the correct measures due to, for example, physical aging or environmental factors. To tolerate this problem, various approaches have been applied, often focusing on correcting the input data from the failed sensor. In this study, we adopt an alternative approach and propose machine learning-based failure tolerance that ignores failed sensors. To tolerate for the failed sensor and to keep the overall prediction accuracy acceptable, a Single Plurality Voting System (SPVS) classification approach is used. Hereby, single classifiers are trained by each feature and based on the outcome of these classifiers, and a composed classifier is built. To build our SPVS-based technique, K-Nearest Neighbor (kNN), Decision Tree, and Linear Discriminant Analysis (LDA) classifiers are applied as the base classifiers. Our proposed approach has a clear advantage over traditional machine learning models since it can tolerate the sensor failure or other types of failures by ignoring and thus enhance the assessment of food quality. To illustrate our approach, we use the case study of beef cut quality assessment. The experiments showed promising results for beef cut quality prediction in particular, and food quality assessment in general.

… using machine vision technology. This paper reviews the latest development of computer vision technology with respect to quality inspection in the agricultural and food industry. …

Abstract Computer vision is a field that involves making a machine “see”. This technology uses a camera and computer instead of the human eye to identify, track and measure targets for further image processing. With the development of computer vision, such technology has been widely used in the field of agricultural automation and plays a key role in its development. This review systematically summarizes and analyzes the technologies and challenges over the past three years and explores future opportunities and prospects to form the latest reference for researchers. Through the analyses, it is found that the existing technology can help the development of agricultural automation for small field farming to achieve the advantages of low cost, high efficiency and high precision. However, there are still major challenges. First, the technology will continue to expand into new application areas in the future, and there will be more technological issues that need to be overcome. It is essential to build large-scale data sets. Second, with the rapid development of agricultural automation, the demand for professionals will continue to grow. Finally, the robust performance of related technologies in various complex environments will also face challenges. Through analysis and discussion, we believe that in the future, computer vision technology will be combined with intelligent technology such as deep learning technology, be applied to every aspect of agricultural production management based on large-scale datasets, be more widely used to solve the current agricultural problems, and better improve the economic, general and robust performance of agricultural automation systems, thus promoting the development of agricultural automation equipment and systems in a more intelligent direction.

Computer vision (CV) applies to many human activities, and its application is a decisive key for the agri-food industry as it progresses towards Industry 4.0. In the agricultural field, CV systems are applied to seeding, cultivation, farm management, disease control, weed control, robotic harvesting, post-harvest control, quality assessment, composition analysis, sorting, and classification. Therefore, the coupling of CV systems, data from new sensors, and artificial intelligence tools such as machine learning and deep learning can enable the automatic management of many tasks previously depended on humans. Thus, the aim of this paper is to review the state-of-the-art CV applied to food and agricultural products.

… inspection. Here we consider a highspeed potato grading and quality inspection system … A real-time computer vision assessment and control of thermal comfort for grouphoused pigs…

This paper gives two contributions to the state-of-the-art for viticulture technology research. First, we present a comprehensive review of computer vision, image processing, and machine learning techniques in viticulture. We summarize the latest developments in vision systems and techniques with examples from various representative studies, including, harvest yield estimation, vineyard management and monitoring, grape disease detection, quality evaluation, and grape phenology. We focus on how computer vision and machine learning techniques can be integrated into current vineyard management and vinification processes to achieve industry relevant outcomes. The second component of the paper presents the new GrapeCS-ML database which consists of images of grape varieties at different stages of development together with the corresponding ground truth data (e.g., pH and Brix) obtained from chemical analysis. One of the objectives of this database is to motivate computer vision and machine learning researchers to develop practical solutions for deployment in smart vineyards. We illustrate the usefulness of the database for a color-based berry detection application for white and red cultivars and give baseline comparisons using various machine learning approaches and color spaces. This paper concludes by highlighting future challenges that need to be addressed prior to successful implementation of this technology in the viticulture industry.

With the escalating global challenges of food security and resource sustainability, innovative solutions like deep learning and computer vision are transforming agricultural practices by enabling data-driven decision-making. This paper provides a focused review of recent advancements in deep learning-enabled computer vision techniques tailored specifically for greenhouse environments. First, deep learning and computer vision fundamentals are briefly introduced. Over 100 studies from 2020 to date are then comprehensively reviewed in which these technologies were applied within greenhouses for growth monitoring, disease detection, yield estimation, and other tasks. The techniques, datasets, models, and overall performance results reported in the literature are analyzed. Tables and figures showcase real-world implementations and results synthesized from current research. Key challenges are also outlined related to aspects like model adaptability, lack of sufficient labeled greenhouse data, computational constraints, the need for multi-modal sensor fusion, and other areas needing further investigation. Future trends and prospects are discussed to provide guidance for researchers exploring computer vision in the niche greenhouse domain. By condensing prior work and elucidating the state-of-the-art, this timely review aims to promote continued progress in smart greenhouse agriculture. The focused analysis, specifically on greenhouse environments, fills a gap compared to previous agricultural surveys. Overall, this paper highlights the immense potential of computer vision and deep learning in driving the emergence of data-driven, smart greenhouse farming worldwide.

… and the expectation for improved quality in consumer agricultural and food products which has increased the need for enhanced quality monitoring in field of agricultural and food, there …

… of machine learning and computer vision for agriculture-based applications to help the researchers to have a deep … 3 and 4, different image acquisition techniques and computer vision …

… principles of computer vision for nondestructive quality assessment … for quality inspection and monitoring of the product. … for quality monitoring and measurement of different agricultural […

Efficient quality control in the agriculture sector, particularly regarding the inspection of vegetables and fruits, stands as a critical necessity in today's health-focused industry. Conventional fruit grading methods, ill-suited for large-scale production, demand an automated, non-invasive, and economically feasible substitute. Computer vision emerges as a promising avenue, leveraging image analysis and machine learning algorithms to evaluate the quality of produce. The convergence of computer vision and image processing technologies in contemporary agriculture has brought about a substantial transformation in quality assessment methodologies. This paper conducts an in-depth exploration of the amalgamation of computer vision and image processing techniques for the evaluation of agricultural produce quality. Through a comprehensive review, this scientific analysis investigates the integration of computer vision and image processing techniques in agricultural quality assessment. It scrutinizes key studies, their practical implementations, outcomes, and the research voids they reveal. Technological progressions within the agricultural domain have the potential to amplify productivity and curtail the circulation of flawed or substandard products. Moreover, this study deliberates on the forthcoming trends in computer vision technology applications, accentuating their prospective influence on the vegetables and fruits industry.

Agriculture is the foundational industry for human survival, profoundly impacting economic, ecological, and social dimensions. In the face of global challenges such as rapid population growth, resource scarcity, and climate change, achieving technological innovation in agriculture and advancing smart farming have become increasingly critical. In recent years, deep learning and computer vision have developed rapidly. Key areas in computer vision—such as deep learning-based image processing, object detection, and multimodal fusion—are rapidly transforming traditional agricultural practices. Processes in agriculture, including planting planning, growth management, harvesting, and post-harvest handling, are shifting from experience-driven methods to digital and intelligent approaches. This paper systematically reviews applications of deep learning and computer vision in agricultural growth management over the past decade, categorizing them into four key areas: crop identification, grading and classification, disease monitoring, and weed detection. Additionally, we introduce classic methods and models in computer vision and deep learning, discussing approaches that utilize different types of visual information. Finally, we summarize current challenges and limitations of existing methods, providing insights for future research and promoting technological innovation in agriculture.

Sorting machines use computer vision (CV) to separate food items based on various attributes. For instance, sorting based on size and colour are commonly used in commercial machines. However, detecting external defects using CV remains an open problem. This paper presents an experimental contribution to external defect detection using deep learning. An uncensored dataset with 43,843 images including external defects was built during this study. The dataset is heavily imbalanced towards the healthy class, and it is available online. Deep residual neural network (ResNet) classifiers were trained that are capable of detecting external defects using feature extraction and fine-tuning. The results show that fine-tuning outperformed feature extraction, revealing the benefit of training additional layers when sufficient data samples are available. The best model was a ResNet50 with all its layers fine-tuned. This model achieved an average precision of 94.6 % on the test set. The optimal classifier had a recall of 86.6 % while maintaining a precision of 91.7 % . The posterior class-conditional distributions of the classifier scores showed that the key to classifier success lies in its almost ideal healthy class distribution. The results also explain why the model does not confuse stems/calyxes with external defects. The best model constitutes a milestone for detecting external defects using CV. Because deep learning does not require feature engineering or prior knowledge about the dataset content, the methodology may also work well with other foods.

Quality inspection of fruits and vegetables linked to food safety monitoring and quality control. Traditional chemical analysis and physical measurement techniques are reliable, they are also time-consuming, costly, and susceptible to environmental and sample changes. Hyperspectral imaging technology combined with deep learning methods can effectively overcome these problems. Compared with human evaluation, automated inspection improves inspection efficiency, reduces subjective error, and promotes the intelligent and precise fruit and vegetable quality inspection. This paper reviews reports on the application of hyperspectral imaging technology combined to deep learning methods in various aspects of fruits and vegetables quality assessment. In addition, the latest applications of these technologies in the fields of fruit and vegetable safety, internal quality, and external quality inspection are reviewed, and the challenges and future development directions of hyperspectral imaging technology combined with deep learning in this field are prospected. Hyperspectral imaging combined with deep learning has shown significant advantages in fruit and vegetable quality inspection, especially in improving inspection accuracy and efficiency. Future research should focus on reducing costs, optimizing equipment, personalizing feature extraction, and model generalizability. In addition, the development of lightweight models and the balance of accuracy, the enhancement of the database and the importance of quantitative research should also be brought to attention. These efforts will promote the wide application of hyperspectral imaging technology in fruit and vegetable inspection, improve its practicability in the actual production environment, and bring important progress for food safety and quality management.

… Grain quality control directly affects the efficiency of grain … grain quality assessment has been a focal point of agricultural … the field highlight that computer vision-based classification has …

Abstract Emerging technologies such as computer vision and Artificial Intelligence (AI) are estimated to leverage the accessibility of big data for active training and yielding operational real time smart machines and predictable models. This phenomenon of applying vision and learning methods for the improvement of food industry is termed as computer vision and AI driven food industry. This review contributes to provide an insight into state-of-the-art AI and computer vision technologies that can assist farmers in agriculture and food processing. This paper investigates various scenarios and use cases of machine learning, machine vision and deep learning in global perspective with the lens of sustainability. It explains the increasing demand towards the AgTech industry using computer vision and AI which might be a path towards sustainable food production to feed the future. Also, this review tosses some implications regarding challenges and recommendations in inclusion of technologies in real time farming, substantial global policies and investments. Finally, the paper discusses the possibility of using Fourth Industrial Revolution [4.0 IR] technologies such as deep learning and computer vision robotics as a key for sustainable food production.