Bruise damage is a very commonly occurring defect in apple fruit which facilitates disease occurrence and spread, leads to fruit deterioration and can greatly contribute to postharvest loss. The detection of bruises at their earliest stage of development can be advantageous for screening purposes. An experiment to induce soft bruises in Golden Delicious apples was conducted by applying impact energy at different levels, which allowed to investigate the detectability of bruises at their latent stage. The existence of bruises that were rather invisible to the naked eye and to a digital camera was proven by reconstruction of hyperspectral images of bruised apples, based on effective wavelengths and data dimensionality reduced hyperspectrograms. Machine learning classifiers, namely ensemble subspace discriminant (ESD), k-nearest neighbors (KNN), support vector machine (SVM) and linear discriminant analysis (LDA) were used to build models for detecting bruises at their latent stage, to study the influence of time after bruise occurrence on detection performance and to model quantitative aspects of bruises (severity), spanning from latent to visible bruises. Over all classifiers, detection models had a higher performance than quantitative ones. Given its highest speed in prediction and high classification performance, SVM was rated most recommendable for detection tasks. However, ESD models had the highest classification accuracy in quantitative (>85%) models and were found to be relatively better suited for such a multiple category classification problem than the rest.

Spectroscopy data are useful for modelling biological systems such as predicting quality parameters of horticultural products. However, using the wide spectrum of wavelengths is not practical in a production setting. Such data are of high dimensional nature and they tend to result in complex models that are not easily understood. Furthermore, collinearity between different wavelengths dictates that some of the data variables are redundant and may even contribute noise. The use of variable selection methods is one efficient way to obtain an optimal model, andthis was the aim of this work. Taking advantage of a non-contact spectrometer, near infrared spectral data in the range of 800–2500 nm were used to classify bruise damage in three apple cultivars, namely ‘Golden Delicious’, ‘Granny Smith’ and ‘Royal Gala’. Six prominent machine learning classification algorithms were employed, and two variable selection methods were used to determine the most relevant wavelengths for the problem of distinguishing between bruised and non-bruised fruit. The selected wavelengths clustered around 900 nm, 1300 nm, 1500 nm and 1900 nm. The best results were achieved using linear regression and support vector machine based on up to 40 wavelengths: these methods reached precision values in the range of 0.79–0.86, which were all comparable (within error bars) to a classifier based on the entire range of frequencies. The results also provided an open-source based framework that is useful towards the development of multi-spectral applications such as rapid grading of apples based on mechanical damage, and it can also be emulated and applied for other types of defects on fresh produce.

Introduction: Globally, 1.3 billion tons of food is lost or wasted each year, negatively impacting food security, the economy, and the climate. Fresh fruits and vegetables (FFVs), with their short shelf life and temperature sensitivity, are the most affected. This study systematically evaluates the integration of Machine Learning (ML), Adaptive Learning (AL), the Internet of Things (IoT), and Fog computing for temperature-break detection and prediction in FFVs supply chains. It critically evaluates their individual and combined capabilities, identifying compounding barriers that prevent genuine real-time integration of these technologies, while assessing their performance and operational readiness for real-time cold chain monitoring. Additionally, the role of fog computing in enabling efficient ML/AL deployment at the network edge is investigated for real-time applications in dynamic environments, with an implementation framework provided. Methods: Based on the PRISMA framework, searches of Scopus, Web of Science, IEEE Xplore, ACM Digital, supplemented by citation and reference chasing, produced 830 pre?deduplication records, identifying 14 relevant studies. Results: From the 12 analysed unique-dataset studies, 7 (58.3%) collected data using Basic Sensors, 3 with WSN (25%), and 2 with IoT (16.7%). Of the 14 ML studies, 5 (35.7%) detected temperature breaks, 5 (35.7%) predicted FFVs' temperature values, 3 (21.4%) predicted internal temperature (IT) values of a cold room or container, and 1 predicted IT values and time-to-temperature breaks. None of the studies predicted temperature breaks (event occurrence) or even their causes, while a few detected these breaks and their predefined causes. Four (28.6%) studies used IoT data, but none enabled live ML inference. None of the reviewed studies includes Fog or AL. Discussion: An integrated IoT-Fog-AL framework is thus proposed to address these gaps. These areas require focus to proactively reduce temperature breaks, thereby minimising food wastage and its associated effects, while also enhancing supply chain resilience and food security.

Postharvest losses of fresh produce threaten food security, particularly in resource-constrained supply chains where quality is still assessed by subjective manual inspection. This study presents a two-component system: a MobileNetV2 convolutional neural network that classifies images of produce as fresh or rotten, and a browser-based web application that runs the classifier entirely client-side using TensorFlow.js, without a server or specialised hardware. Using a public dataset of roughly 30,000 labelled images spanning nine produce types, the model achieved 98.46% accuracy on a held-out test set, with a macro-averaged F1-score of 0.985 and an ROC-AUC of 0.9989. The confidence output, presented to users as a Visual Freshness Index (VFI), was well calibrated against the binary label (Expected Calibration Error: 0.0051), making it a dependable indicator of classification confidence. Trained with native preprocessing, the heavier baselines ResNet50, EfficientNetB0, and VGG16 achieved comparable accuracy, so MobileNetV2 was retained for its small footprint and suitability for deployment. The shelf-life and storage suggestions shown by the application are indicative heuristics, not validated model outputs. Its offline, browser-based design makes the system potentially suited to resource-constrained and low-connectivity agricultural settings.

Introduction: This study proposes an integrated framework combining deep learning, causal discovery, explainability, and domain thresholds for proactive monitoring in temperature-sensitive fresh fruits and vegetables cold chains, a domain that remains largely unexplored. The framework predicts temperature breaks (a critical physical hazard that compromises food safety and accelerates postharvest losses), prospectively identifies their causal drivers, and prioritizes countermeasures. Methods: Using two IoT-gathered datasets from an Apple cold room in South Africa, eight machine learning models and three baselines were compared for temperature prediction, including a proposed LSTM-GRU hybrid (LSGR) model. Causal relationships between measured parameters and predicted temperature were identified using Latent Peter & Clark Momentary Conditional Independence (LPCMCI; compared with PCMCI). At the same time, SHapley Additive exPlanations quantified feature importance, and domain-based thresholds complemented them for countermeasure prioritization after temperature-break prediction (TBP). Results: On the extended dataset, deep neural networks' (DNN) R ² and normalized mean squared error improved by up to 30% and to 61%, respectively, over the smaller dataset, demonstrating strong generalizability to substantially different cold room operating conditions. The LSGR demonstrated the best overall performance across both temperature and TBP tasks among the recurrent models assessed, raising the fewest false alarms per day. All DNNs flagged breaks in advance; no baseline did. LPCMCI recovered most direct causal links after break prediction, especially for lagged temperature, identifying it as the dominant factor associated with predicted breaks. Notably, post-TBP, causality, importance, and thresholds strongly agreed on the same top two critical parameters or countermeasures (temperature and humidity), providing consistent and actionable insights for decision-making. Discussion: These findings demonstrate the viability of combining causal discovery, explainability and thresholds for cold chain management, with implications for reducing food wastage and enhancing quality, safety, and sustainability from farm to fork in temperature-sensitive FFVs supply chains. LPCMCI recovered a few links overall, indicating the influence of unmeasured drivers and warranting further investigation, including additional features, their interactions, and seasonal data, as well as a fuller characterization of causality-regime dependence.