Real-time monitoring of fish physiological stress during green waterless low-temperature transportation remains a key technical challenge for aquatic food quality and safety. Such conditions induce complex multi-scale physiological responses, leading to high mortality and quality deterioration of aquatic products. In this study, we report an additively manufactured flexible dual-modal biosensing system based on biointerface engineering, which integrates breath peak (BrP) and bioimpedance (BI) detection modules and is tailored for non-invasive, continuous evaluation of fish stress in practical aquatic food cold-chain scenarios. A key interfacial optimization, validated through systematic electric field simulations, demonstrates that a two-electrode BI configuration generates a more uniform electric field distribution at the sensor-tissue interface, which markedly enhances signal stability and reproducibility for long-term cold-chain biosensing applications. We further establish a hierarchical multi-level information fusion framework. Firstly, Granger causality analysis verifies the bidirectional temporal causality between BrP and BI signals, laying a mechanistic foundation for effective biological data fusion. Secondly, a stacking ensemble model with GRU and SVM as base models was built to achieve multi-level information fusion for accurate stress assessment. Experimental results show the proposed dual-modal biosensing system reaches a stress evaluation accuracy of 0.9669, markedly outperforming single-modal BrP (0.9256) and BI (0.8182). The data-driven non-invasive biosensing method delivers superior performance for target fish under waterless low-temperature cold-chain conditions, with great potential for broader application. It is poised to facilitate aquatic food monitoring, underpin sustainable high-quality cold-chain management for aquatic products, and reduce post-harvest losses while safeguarding aquatic food quality in green logistics.

The increasing demand for lamb products has highlighted the need for reliable non-destructive testing technologies to evaluate meat freshness indicators during storage and distribution. This study developed multi-parameter prediction models based on bioimpedance technology to assess lamb quality parameters and monitor quality changes throughout storage and transportation. Multi-parameter non-destructive bioimpedance data (bioimpedance, phase angle, frequency, time) were collected from lamb samples stored at different temperatures (4 degrees C, 15 degrees C, and 25 degrees C) and tested under various input frequencies (1, 10, 100, 1000, and 10000 Hz). These data were correlated with conventional meat quality parameters like color difference (Delta E), pH, total volatile basic nitrogen (TVB-N), tenderness, and shear force. Linear regression (LR), support vector regression (SVR), random forest regression (RF), and gradient boosting regression (GBR) models were developed to analyze quality changes, with a focus on predicting meat quality at 4 degrees C. The results showed that different models demonstrated optimal performance for predicting specific quality parameters. The RF model performed best in predicting Delta E, tenderness, and shear force, with coefficients of determination (R2) of 0.9965, 0.9955, and 0.9977, and mean squared errors (MSE) of 0.0410, 4.8224, and 4.0706, respectively. The standard errors for these predictions were 0.331, 3.8163, and 5.1193, respectively. Meanwhile, the GBR model excelled in predicting pH and TVB-N, with R2 values of 0.9640 and 0.9740, MSE of 0.0003 and 0.0582, and standard errors of 0.0107 and 0.1611, respectively. By selecting the most appropriate model for each specific quality parameter, this model selection strategy significantly improved the accuracy of lamb freshness indicator prediction, particularly under cold chain transportation and storage conditions, providing a reliable algorithmic foundation for the construction of future lamb quality monitoring systems.

Heat stress, intensified by high summer temperatures and humidity, severely impacts the growth, reproduction, and welfare of sheep, particularly in intensive farming systems. This study aims to develop a lightweight, comfortable, and multifunctional monitoring solution with strong decision-making capabilities to address the shortcomings of existing wearable monitoring devices. We designed and validated a multi-sensor wearable electronic system (MWES) that attaches to a belt around the animal’s thorax. This system integrates flexible strain sensors, silicon-based sensors, and a hybrid machine learning model to enable real-time, wireless monitoring of respiratory activity, body posture, and body temperature, along with heat stress-level assessment. The experimental results show that: (1) The flexible strain sensor achieved high-quality, long-term monitoring of respiratory signals. (2) The integrated 3-axis digital compass and NTC thermocouple sensors accurately measured body posture and temperature. (3) The Long short-term memory (LSTM) model enabling behavioral analysis achieves 99% accuracy. (4) The environmental node reliably monitored ambient temperature and humidity, allowing comprehensive stressor analysis. (5) By fusing multimodal data streams, the XGBoost model classified five levels of heat stress with ahigher accuracy of 97.5% in test set. (6) Interpretable machine learning methods further revealed the relationships between physiological indicators and stress levels, enhancing model transparency and credibility. The MWES offers a pioneering tool for real-time heat stress monitoring in livestock, representing a major advancement in wearable electronic applications for animal science and agricultural management. This system promises to significantly improve personalized livestock health management in intensive farming environments.

ABSTRACT Grouper faces quality inspection challenges during cold chain logistics transportation, with traditional inspection methods suffering from labor‐intensive and damaging drawbacks. This study employs bioimpedance and bioimpedance imaging technology (multimodal detection) combined with machine learning algorithms to construct a quality classification model for chilled grouper, aiming to achieve rapid and precise monitoring. Research design of a multimodal bioimpedance detection system for grouper. Analysis of multimodal bioimpedance parameters and quality parameters. Construction of multimodal bioimpedance detection models using four machine learning classification models: K‐Nearest Neighbors (KNN), Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Network (ANN). Average classification accuracies were 1.00, 0.96, and 0.96, respectively, and compared the performance of the models. The developed system achieves efficient classification of chilled grouper quality with an average accuracy ≥ 96%, establishing a new paradigm for non‐destructive testing and intelligent classification. This approach is scalable to other perishable foods, enhancing supply chain traceability and sustainability.

IntroductionKoi herpesvirus (KHV or CyHV-3) causes a highly contagious and lethal disease in common carp and koi, resulting in substantial economic losses in global aquaculture. Rapid point-of-care testing (POCT) of KHV is critical for curbing its spread and preventing outbreaks. However, conventional diagnostic approaches, such as TaqMan quantitative PCR (qPCR), require professional personnel and specialized equipment, limiting their on-site applicability.MethodsA rapid visual POCT assay was developed by combing loop-mediated isothermal amplification (LAMP) with non-invasive mucus swab sampling. Two optimized release solutions enabled one-step nucleic acid preparation within 5 min. Among five primer sets designed, the optimal set targeting the thymidine kinase (TK) gene was selected for isothermal amplification at 65 °C for 60 min. The assay was further adapted for naked-eye result interpretation using a pH-sensitive indicatior dye.ResultsThe assay demonstrated a detection limit of 21.42 copies/μL, with a 95% confidence interval of 14.88—41.83 copies/μL. No cross-reactivity was observed with seven common fish pathogens. Comparative evaluation with TaqMan qPCR using parallel clinical specimens yielded 100% concordant results. The total detection time was 65 min, enabling the POCT implementation across diverse scenarios.DiscussionThis non-invasive, visual LAMP-based POCT platform provides a sensitive, specific and equipment-free diagnostic strategy for KHV detection, facilitating timely disease surveillance, outbreak control management, and aquaculture biosecurity.

Global livestock systems are undergoing a technology-driven transformation, and wearable sensors are now widely used for physiological monitoring, behavioral analysis, early disease detection, and production management. Following PRISMA guidelines, this review synthesizes studies published between 2015 and 2025, including 98 studies retrieved from major academic databases. Unlike earlier reviews that focus mainly on single species or single functions, this study provides a cross-species, multi-sensor synthesis of wearable sensor applications in animal husbandry, covering motion, acoustic, optical, thermal, and emerging sensor technologies. Beyond listing sensor types, it examines how sensor performance and deployment strategies shape core monitoring functions, including behavioral monitoring, physiological assessment, and spatial tracking across livestock species. These functions include detecting feeding, locomotion, rumination, and activity patterns, as well as measuring of body temperature, heart rate, respiratory rate, and location-related parameters. Despite major technological progress, current wearable systems still face challenges in detection accuracy, long-term stability, wearer comfort, and cost. This review synthesizes current research findings, summarizes technological advances, identifies persistent technical and application barriers, and outlines future research directions. It provides an integrated account of the current state and the main directions field.

Facilitating the rapid and straightforward fabrication of flexible sensing device emerges as a pivotal endeavor. Laser-based direct write processing is a promising technology. A detailed analysis is presented for understanding the laser-graphene-based materials interactions while considering lattice vibrational properties, charge carrier scattering with phonons, impurities, defects, and edge boundaries of flakes. The proposed methodology enables the large-scale production of 3D porous reduced graphene oxide (rGO), leveraging it as temperature-sensitive layer for real-time temperature monitoring. The extensive dimensions of LrGO facilitate efficient electron transport, resulting in remarkable electrical conductivity and long-term stability. Experimental results show that the proposed wearable sensing patch exhibits sensitivity of 0.517 %degrees C-1, fast response time, coupled with exceptional stability under bending stress or humidity fluctuations. The comprehensive performance is superior to that of most reported temperature-sensitive devices based on graphene materials. Furthermore, the patch enables accurate long-term temperature tracking and humidity-insensitive respiration monitoring, highlighting its potential for applications in body temperature sensing, respiratory diagnostics, and non-contact humancomputer interfaces. Therefore, the design and methodology presented in this work are important for the future development and application of wearable respiratory sensors and temperature patch.

Jujube is susceptible to biotic and abiotic adversity stresses resulting in abnormal phenotypic defects. Therefore, abnormal phenotype fruits should be removed during postharvest sorting to increase added value. An improved maximum horizontal diameter linear regression (MHD-LR) method for size grading of jujube prior to detection of abnormal phenotypic defects was developed. The accuracy of the MHD-LR model is 95%, with an error of only 0.95 mm. In addition, a method for detecting abnormal phenotypic defects in jujube was established. It can effectively and accurately classify seven kinds of jujube phenotypes (regular, irregular, wrinkled, moldy, hole-broken, skin-broken, and scarred). The data augmentation method based on linear interpolation can effectively expand the dataset with a variance of only 0.0006. Support vector machine-decision tree (SVMDT), logistic regression, back propagation neural network, and long short-term memory network models were established to classify jujube samples with different phenotypes, with accuracies of 99.57%, 99.00%, 99.14%, and 99.29%, respectively. The results showed that the SVMDT model had higher accuracy and explainability. This research is expected to provide a new method to improve the precise classification of abnormal phenotypic defects in postharvest jujube.

Vibrations during transportation inevitably lead to mechanical damage, endangering grape freshness and directly impacting their economic worth. While adequate packaging serves as a viable solution, current studies on packaging efficacy lack depth. Moreover, conventional methods for forecasting fruit freshness fail to accommodate the varying freshness levels of grapes across different packaging techniques. Consequently, a novel approach for predicting fruit freshness leveraging multi-sensing technology and machine learning algorithms is introduced. By reasonably evaluating packaging performance, the automation, intelligence, and accuracy of fruit freshness prediction are enhanced. Initially, critical control points in grape supply chain logistics were scrutinized using the HACCP method to identify key environmental parameters (vibration, temperature, and humidity) and their interaction with grape freshness. Subsequently, an environmental monitoring platform was devised for the grape supply chain, facilitating environmental surveillance under distinct packaging types (corrugated carton, foam box, plastic box, and inflatable package). Through a blend of environmental monitoring outcomes and physical-chemical indicators, the protective efficacy of diverse transport packaging was meticulously analyzed and appraised alongside finite element analysis. Notably, environmental data proved capable of characterizing grape freshness in lieu of quality data, with vibration metrics exhibiting strong correlations with quality metrics. Machine learning models were developed to predict grape freshness based on environmental cues, yielding prediction accuracies of 92.512% (SVM) and 94.334% (GA-ANN). The automated, non-destructive data acquisition and novel machine learning approaches offer a fresh avenue for evaluating packaging, predicting freshness, and managing food quality within grape logistics operations.

It is expected that waterless low-temperature stressful environments will induce stress responses in fish and affect their vitality. In this study, we developed a laser-activated, stretchable, highly conductive liquid metal (LM) based flexible sensor system for fish multi-scale bioimpedance detection. It has excellent conformability, electrical conductivity, bending and cyclic tensile stability. Meanwhile, test result showed that wireless power supply is a potential solution for realizing safe power supply for devices inside waterless low-temperature packages. In addition, a hierarchical regression model (GC-HRM) based on Granger causality was established. The result showed that tissue bioimpedance can induce changes in individual bioimpedance with unidirectional Granger causality. The R2 of the linear regression (LR), support vector regression (SVR) and artificial neural network (ANN) models under single-scale individual bioimpedance were 0.85, 0.90 and 0.78, respectively. By adding the multi-scale bioimpedance features, the R2 of the LR, SVR and ANN models were improved to 0.95, 1.00 and 0.98, respectively.

The health status of livestock will directly or indirectly affect meat quality and farmers' income. Therefore, it is necessary to monitor the physiological indicators of livestock to timely reflect their health status. The respiratory rhythm of animals can provide early warning for the occurrence of diseases, but there is currently limited research on the characteristics of the rhythm. Therefore, this paper conducted the following research. PVDF flexible piezoelectric sensors were prepared by spin-coating, laser-induced graphene, and transfer printing processes, in which three common flexible substrate materials were compared, three key parameters of laser scanning were explored, while practical application tests were carried out on model animals, and the collected respiratory rhythm signals were analyzed in the time and frequency domains. The results showed that (1) posilicone of 25 durometer was identified as the flexible substrate material, and the LIG process with a scanning speed of 120 mm/min, a scanning power of 12 %, and a scanning interval of 0.02 mm as the optimal parameters was determined; (2) the sensor exhibited a response speed of 50 ms, an output voltage of 3.7 v, and still had a stable output; (3) in the monitoring of the respiratory rhythm of rabbits, the monitoring accuracy reached 96.7 %, while the respiratory signal collected by the sensor was randomly intercepted for 10 s, and its respiratory rhythm prediction accuracy reached 95.85 %. At the same time, the abnormal respiratory rhythm change of the animal was found 10 min in advance, giving the farmer more time to deal with the problem. (4) The monitoring of the respiration of farm animals such as Hu sheep is being realised. At the same time, the time domain as well as frequency domain waveforms of the Hu sheep were analysed and compared with the time domain frequency domain plots of the rabbit, which showed that the sensor could achieve good results for both higher frequency breathing and low-frequency breathing. It provides potential technical support for future health monitoring and early prediction of diseases in large farm animals.

Temperature fluctuations at different stages of the supply chain increase the frozen-thawed cycle of perishable foods, potentially leading to quality and safety issues. For raw edible salmon in particular, it is not possible to ignore the issue of adulteration when frozen-thawed flesh is sold as fresh flesh. It is a challenge to achieve real-time detection of frozen-thawed salmon adulteration in fresh salmon. Existing impedance change ratio (Q-value) and PCA models cannot accurately authenticate frozen-thawed cycle adulterated salmon. In this paper, a flexible bioimpedance based non-destructive detection system was designed to authenticate adulterated salmon by online monitoring of changes in bioimpedance signals, ambient temperature, and relative humidity. The system provided a high level of monitoring accuracy and stability. Furthermore, an improved machine learning classification model based on principal component analysis - Bayesian optimization algorithm - support vector machine (PCA-BOA-SVM) was developed to effectively identify frozen-thawed adulterated salmon. The optimised model performance enhanced with prediction accuracy, precision, recall and F1 score of 0.9683, 0.9708, 0.9683 and 0.9679, respectively. This work could provide an effective solution to improve the authentication of food adulteration in the perishable food supply chain by improving traceability at all stages of the supply chain and sustainability of food industry development.

Ensuring the freshness of salmon in the face of freezing and thawing cycles during cold chain transportation is a critical challenge, particularly as some merchants may mislabel frozen-thawed salmon as fresh for higher profits. In this study, we developed an advanced monitoring system to track changes in bioimpedance signals and quality parameters of salmon. Bioimpedance data at 1 kHz and 16 kHz were analyzed and fed into four machine learning classification models: support vector machine (SVM), linear discriminant analysis (LDA), K-nearest neighbors (KNN), and random forest (RF). Following correlation analysis between quality parameters and bioimpedance signals, the classification models achieved average accuracy, precision, recall, and F1 scores of 98%, 97%, 97%, and 97%, respectively. The findings of this study can be applied to other fish and meat quality classification tasks, offering a reliable method for identifying freeze-thaw cycles in these products.

Waterless and low temperature transportation is a green and efficient way for the transportation of live fish. However, waterless and low temperature conditions could lead to a stress response in live fish, resulting in reduced transport survival rates. It is still a challenge to intelligently monitor the breath stress state of live fish under adversity stress. Temperature (T), relative humidity (RH), oxygen (O2) and carbon dioxide (CO2) signals can reflect changes in adversity stress environment; while the breath angle sensors can monitor the gill opening and closing angle (breath angle) to reflect changes in fish breath. In this work, microenvironment and breath angle sensor systems were designed and developed to comprehensively evaluate the breath stress state of fish. Meanwhile, the Kalman filter-quaternion-fast Fourier transform method was established to process the breath angle signal. The breath angle signal indicated that the sturgeon had three levels of breath stress: acute fluctuation stage (0-2.5h), organismal regulation stage (2.5-16h) and cumulative stress stage (>16h). In addition, linear regression (LR), back propagation neural network (BPNN), support vector regression (SVR), and radial basis function neural network (RBFNN) models were established for breath efficiency signal prediction. The R-2 of the RBFNN (0.9544) model was significantly higher than the LR (0.8092), BPNN (0.9289), and SVR (0.9428) models. This study provided a reference for further intelligent monitoring and management of the fish breath stress state under waterless and low temperature conditions.

Crabs have a high nutritional and economic value and there is an increased demand for live crabs. However, live crabs have limited shelf life and they are susceptibility to death and spoilage during transportation. Monitoring live crab vitality during transportation in the supply chain to meet consumer acceptability is vital. In this study, an information fusion enabled live crab viability monitoring system was developed. Hazard analysis and critical control point (HACCP) analysis was used to identify potential hazards and critical control points in the transportation supply chain that affect live crab vitality. Multi-source information during live crab transportation was collected by integrating temperature, relative humidity, oxygen, alcohol, aldehyde, and impedance sensors. The predictive modelling of live crab vitality based on information fusion effectively improved, the utilisation of information and the accuracy of vitality prediction. An ensemble learning based soft-voting classifier outperformed the individual performances of other models (i.e. support vector machine, random forest, k-nearest neighbour) and it achieved accuracy above 99% and 86% at 4 degrees C and 25 degrees C. The system evaluation indicated that the developed information fusion-based vitality monitoring offers the possibility of ubiquitous monitoring of the vitality of aquatic products in the transportation supply chain and improves the economic efficiency of the supply chain.(c) 2023 IAgrE. Published by Elsevier Ltd. All rights reserved.

The sowing depth decision is a critical link in variable-depth sowing (VDS). Compared with rule-based sowing depth decisions, model-based methods are more intelligent and flexible under different growing conditions. This study develops a prediction model that predicts the germination and establishment of maize at different soil depths, enabling the rapid determination of the optimum sowing depth by comparing the predicted results. Dependent variables of the prediction model were selected from five seedling quality indexes, which were the emergence rate index (ER) and seedling uniformity index (Un). Independent variables were soil parameters, namely bulk density, soil depth, and soil moisture. Datasets for modelling were obtained from two experimental sites: one (240 data groups) for model training and the other (48 data groups) for model testing. After the classification of dependent variable data, a support vector machine (SVM), random forest (RF), and extreme learning machine (ELM) were used to process the training dataset and establish 186 models. Further assessment of the three selected models was performed using a test dataset. Finally, a model embedded in the SVM algorithm was developed, and achieved accuracies of 70.83% and 72.92% when predicting ER and Un. Maize emergence and establishment were effectively improved in the field by applying the developed model to select the optimum sowing depths.

Environmental changes are expected to induce a stress response in oysters, affecting their vitality and meat quality. However, the real-time monitoring of oyster stress levels remains challenging. In this study, we developed a flexible bioimpedance sensor system for real-time monitoring of the bioimpedance and phase angle of oysters under different stress regimens. To accommodate the bivalve structure of the oyster specimens, three bioimpedance measurement positions were utilized, namely, right-right shell, left-left shell, and left-right shell. A predictive model based on principal component analysis combined with the use of a support vector machine (PCA-SVM) was first set up to evaluate stress levels, and the performance of a flexible sensor system was evaluated. The results showed that the flexible sensor system possessed a high level of accuracy and stability. The average precision, recall and F1-score for the best predicted position (right-right-shell) were 95.15%, 93.94% and 93.94%, respectively. The results of this study may be extended to other live animal stress monitoring applications, providing a means of monitoring the health status of live animals.

During the 14th Five-Year Plan period, the total fishery production in China is expected to continue to grow, and aquatic products further become an important dietary component for consumers. However, it is highly likely to lead to food safety incidents due to irregularities in the breeding, processing and cooking processes. Spectroscopy has become a hot spot for aquatic product testing technology because of its advantages of rapid, nondestructive, and high-test reproducibility, reflecting both the spectral properties of objects and the spatial information of samples, but mostly focusing on freshness testing. This paper reviews the literature related to the application and progress of spectroscopic techniques in foreign matter residues of aquatic products in the past 10 years. It introduces the common spectroscopic techniques in their application and progress from four aspects: fishbone detection, adulteration analysis, parasite detection and heavy metal detection, mainly including X-Ray technology, visible imaging, near-infrared imaging, hyperspectral imaging, etc. While introducing the current problems, we look forward to the development of spectroscopic techniques in aquatic products. The development prospect of foreign substance residue detection : traditional detection algorithms are further optimized, and multi-spectral technology is used for foreign substance residue detection of aquatic products; the great advantage of deep learning in feature extraction is applied, and the application field of spectral technology in foreign substance residue detection of aquatic products is studied more deeply; the organic integration of spectral technology and multiple detection technologies becomes an inevitable trend, and online real-time detection becomes possible.

The of monitoring the Internet of Things (IoT) in the cold chain allows process data, including packaging data, to be more easily accessible. Proper optimization modelling is the core driving force towards the green and low-carbon operation of cold chain logistics, laying the necessary foundation for the development of a data-driven modelling system. Since efficient packaging is necessary for loss control in the cold chain, its final efficiency during circulation is important for realizing continuous loss prevention and efficient supply. Thus, it is urgent to determine how to utilize these continuously acquired data and how to formulate a more accurate packaging efficiency control methodology in the agri-products cold chain. Through continuous monitoring, we examined the feasibility of this topic by focusing on the concept of data-driven evaluation modelling and the dynamic formation mechanism of comprehensive packaging efficiency in cold chain logistics. The packaging efficiency in the table grape cold chain was used as an example to evaluate the comprehensive efficiency evaluation index system and data-driven evaluation framework proposed in this paper. Our results indicate that the established methodology can adapt to the continuity of comprehensive packaging efficiency, also reflecting the comprehensive efficiency evaluation of the packaging for different times and distances. Through the evaluation of our results, the differences and the dynamic processes between different final packaging efficiencies at different moments are effectively displayed. Thus, the continuous improvement of a low-carbon system in cold chain logistics could be realized.