Biosystems Engineering最新文献

Accurate oestrous detection is critical to optimise sows' reproductive performance. The conventional method of oestrous detection relies on the laborious back-pressure test. This study presents an automated oestrous detection method for sows housed in individual stalls using a robotic imaging system and neural networks. A robotic imaging system consisting of a LiDAR camera was used to monitor a group of stall-housed sows at a 10-min interval to capture their postures and vulva volume. Imagery data were analysed using a previously developed pipeline. Results showed that significant changes were observed in daily standing index, sternal lying index, lateral lying index, posture change frequency, and vulva volume before the onset of oestrous. A 1-D convolutional neural network model architecture for oestrous detection was developed using Days from Weaning (DFW), behaviour features, and vulva volume features as inputs. The oestrous detection models were evaluated using 10-fold cross validation. The training and testing accuracies of the oestrous detection model were 96.1 ± 2.0% and 92.3 ± 10.1% when using the DFW and behaviour features as input. The model's training and testing accuracies increased to 98.1 ± 2.4% and 98.0 ± 4.2% when vulva volume features were added to the input. While it is difficult to trace the behaviour of sows housed in group conditions, combining vulva volume features with DFW could be a suitable method to detect the onset of oestrous in these sows. The training and testing accuracies of this method of oestrous detection were 97.9 ± 1.4% and 95.2 ± 4.8%. However, further validation under real group house conditions is needed.

To build a general model for wheat root systems, this study tests and analyses the geometric morphology of wheat root systems in soil. On this basis, a geometric model of the wheat root system is constructed, and a discrete element model of the wheat root system is established using the bonding model. Additionally, through the analysis of the shape of the soil particles used, it is determined that the soil particles can be simplified to spheroidal and prismatic shapes, based on which a discrete element model of the soil particles is established using the Edinburgh Elasto-Plastic Adhesion model. Meanwhile, the parameters of the soil model at two water contents (11% and 14%) are obtained by the soil angle of repose test and simulation. On the basis of the above work, the accuracy of soil model parameters is verified by soil direct shear test, and the accuracy of the root system bonding model and parameters, as well as the root-to-soil contact model and parameters, are verified by pulling tests and simulations of actual single roots in soil. At the same time, the feasibility and effectiveness of the general model of the wheat root system established are proven through the comparison between the pulling test and simulation of the actual root system in soil, which provides a reference for the study of the overall modelling of the wheat plant and the simulation of the contact interaction between the root system and agricultural machinery components during postharvest tillage.

This paper develops and verifies a multiscale computational fluid dynamics (CFD) model to investigate the airflow and heat transfer in kiwifruit cold storage under forced-air cooling (FAC). The CFD model incorporates the material properties, geometry, position of kiwifruit and ventilated packaging box, and the detailed structure of the cooling unit. For the multiscale modeling, the material characteristics are described using three interconnected sub-models, focusing each on different spatial scales: the warehouse-scale, packaging-scale, and kiwifruit-scale. In the FAC experiment, the measured airflow and temperature on these three spatial scales were obtained and compared with the simulated results. The data analysis shows that the environmental fluctuations at different scales weaken significantly in a stepwise manner with the cushioning of packaging and fruit flesh, indicating coupling between the spatial scales, which is well reflected in the numerical simulation. The average kiwifruit temperature decreases from 20 to 6 °C in 14.5 h (experimental) and 15.6 h (simulated). Specifically, the average mean absolute error, mean absolute percentage error, and root mean squared error of the predicted airflow and kiwifruit temperature were 0.116 m s⁻¹, 1.26 °C; 26.8%, 14%; and 0.124 m s⁻¹, 1.54 °C, respectively. These results indicate that the multiscale CFD model accurately and efficiently simulates the airflow and spatiotemporal temperature distribution in the given kiwifruit FAC system. Finally, this study provides a reference for accurately simulating large-scale industrial FAC systems and supports optimal decision-making for the design of sustainable kiwifruit cold chains.

Root architecture systems (RAS) reflect the spatial structure of roots in soil. To clarify the structure and distribution of rice roots and investigate the coupling between roots and soil, wetland rice was selected as the experimental object, and a three-dimensional (3D) growth model of rice root environment-roots (ERoots) based on the parameter Lindenmayer system (L-system) was proposed. ERoots combines a root morphological structure model with a growth model and defines L-system grammar iteration rules with the unit time and unit step length as parameters. At the same time, the basic growth parameters of rice roots were obtained via destructive detection, and 3D growth visualisation of roots was realised via MATLAB. In the soil coupling process, a soil nutrient simulation map was constructed based on the spatial soil characteristics per unit volume, and an adjustment strategy for roots reaching the growth boundary was designed. The flexibility of the model coupled with soil was reflected in the tropisms of root growth, growth rate and root branching strategy. Finally, combined with soil spatial characteristic simulation, geometric growth boundary and 3D root growth model, the ability of 3D growth visualisation of rice roots was verified under three soil conditions: (1) unconfined root growth, (2) confined spatial root growth, and (3) root growth with tropisms. The results indicated that the ERoots root model basically realised coupling with soil and achieved a satisfactory simulation effect in regard to the rice morphological structure. This study provides a reference for 3D growth modelling and visualisation of other crop roots.

Addressing heat stress in dairy farming is a substantial challenge, and there is an increasing need for efficient cooling systems, even in regions with moderate climates. Accurately predicting the efficacy of diverse cooling options under different climatic conditions is crucial for reducing heat stress in modern high-producing dairy cows, aligning with sustainability goals. This study assessed the effectiveness and feasibility of different cooling measures, including fans, sprinklers with fans, and evaporative air cooling, using a dynamic thermoregulatory model. This 3-node dynamic model was developed based on recent animal data simulating the processes of dairy cows' physiological regulation and heat dissipation under various environmental conditions. The cooling methods were based on two principles: enhancing heat loss from cows using fans with/without sprinklers; lowering the ambient temperature by evaporative air cooling. The predicted results were discussed and partly validated using the experimental data from the literature. The predictions indicated that fan cooling alone was effective in ambient temperatures below 26 °C, while higher temperatures required a combination of fans and sprinklers for effective heat stress alleviation. Consideration of individual cow characteristics and environmental factors, including fan speed and wetting area, is crucial for optimal cooling. In regions with high relative humidity, evaporative air cooling could be counterproductive to some extent. The model's predictions largely aligned with experimental data, demonstrating its capability to forecast cooling effects under various climatic conditions. Future model improvements included refining calculations for water holding capacity, wetted skin area, and dry time, depending on the influence of spraying time and rate.

Proactive dietary control of blood vitamin A levels is crucial for the intramuscular fat development in cattle worldwide. However, cattle become susceptible to either vitamin A deficiency or excessive state during fattening stage, influencing cattle performance, health, and beef quality. A good understanding and modelling of vitamin A levels throughout the whole cattle growth phase is needed. This study aims to assist in controlling the fattening process for production of high-marbling beef through a non-invasive monitoring of blood vitamin A levels. Using an automatic double imaging system, this study captured both surface and fundus images of cattle eyes, and based on this, predicted blood vitamin A levels through a novel dynamic analysis of 29 eye features. The best PLS model had a prediction of R² = 0.82 and RMSE = 6.50 IU·dL⁻¹ (equivalent to 0.02 μg · mL⁻¹), which is of a clinically meaningful accuracy. This system can greatly facilitate vitamin A levels management in cattle raising, contributing to the effective control of beef marbling for both the market and industry.

Maize threshing is a complex and dynamic process, and optimisation of operating parameters is essential to improve threshing quality and efficiency. In this study, machine learning was combined with interpretability analysis to investigate the dynamic effects of operating parameters on maize threshing quality and to optimise the threshing process. The maize cob model used to simulate threshing was validated by stacking angle and tensile test. Real-time drum operating parameters and threshing quality data obtained through Discrete Element Method (DEM) threshing simulation were used to train a threshing quality prediction network. The prediction accuracy was improved by incorporating an attention mechanism into the Long Short-Term Memory (LSTM) model with an optimised Root Mean Square Error (RMSE) of 0.0041. The global feature importance and dynamic Shapley Additive Explanations (SHAP) value analyses demonstrated that rotational speed is a key determinant of unthreshed and damaged rates and that its effect varies significantly at different stages of the threshing process. Guided by these analyses, a staged speed adjustment experiment was conducted. Specifically, an increase in rotational speed during the initial threshing phase markedly lowered the initial unthreshed rate for medium and high-speed groups to 6.63% and 2.73%, respectively, a significant improvement over the 67.70% observed in the low-speed group. The final damage rate in the high-speed group decreased by 9.79% relative to the low-speed group. This dynamic analysis approach provides a novel paradigm for optimising complex agricultural processes under varying conditions, offering interpretable insights for precise process control and improvement.

Disease and nutrient deficiency disorders significantly impact the productivity of rice crops. Timely identification of these conditions is essential for effective mitigation of potential crop damage. To address this challenge, considerable research is happening in the field of rice crop monitoring and maintenance, using cutting-edge techniques like Machine learning (ML)/Deep learning (DL). This study aims to address critical aspects of the research landscape, including publication trends, data modalities, ML/DL models, pre-processing methods, segmentation techniques, and feature selection approaches in the context of rice crop's health. By presenting both research findings and existing gaps, this systematic literature review (SLR) offers valuable insights to direct future research endeavours in this domain. Our investigation involves a comprehensive review of articles sourced from Scopus, IEEE Xplore, Science Direct and Google Scholar resulting in a dataset of 91 unique articles spanning from the year 2013–2023. Following rigorous selection criteria, these 91 articles have been considered for in-depth analysis. Through an extensive examination of this corpus, our study seeks to provide answers to seven key questions pertaining to the past, present, and future directions of research of ML/DL application in rice crop health monitoring and disease/disorder diagnosis. The review adheres to the agricultural science-based PRISMA systematic review methodology and incorporates statistical analysis to explore relationships among variables such as dataset sample size, experimental accuracy, and classification models employed in various studies.

The queen bee is the core member of a bee colony, and her loss will pose a great threat to the survival of the colony that may cause colony collapse. However, the process by which queen bee loss affects the internal social state of the bee colony remains unclear. In this paper, we used a multi-sensors system to continually monitor colonies with queen loss and regularly checked their biological status. Our results show that the queen loss initially caused a rapid decrease in brood rearing and changed the foraging strategy of the colony, leading to an increase in food storage. Also the population decline is difficult to reverse in a short time, even if the queen is naturally replaced. This study emphasises the impact of queen bee loss on the operation of the bee colony social system, and elucidates the interconnectedness of the bee colony social system.

Leaf wetness duration is a crucial factor in plant disease management. Current optical methods use standard RGB images to classify leaf wetness as a binary problem, i.e., wet or dry. Green leaves absorb red light, whereas water reflects it. Based on this difference, an experimental platform was built to semi-automatically measure droplet deposition on grape leaves while capturing red laser images using an RGB camera. The setup measured changes in leaf mass and area of scanned leaves to determine the water mass per leaf area as a measure of leaf wetness. A sprayer was used to apply water droplets to the leaves. As the amount of deposited water increased, the mean red channel intensity decreased, with more bright spots in the images. These bright spots were more distinguishable as droplets in the green channel. Segmented leaf area, mean red channel intensity, and the number of identified droplets were used as image features. A generalised additive model was employed to predict the leaf wetness value with extracted features. The R-squared value for the prediction of the validation dataset was 0.71. Image resolution and leaf orientation were identified as factors that influenced the model accuracy. The measurement method introduced in this study shows potential for accurately quantifying leaf wetness, and implies that in practice detecting leaf wetness can be integrated into a multi-classification problem, thereby broadening the potential applications of optical methods.