Biosystems Engineering最新文献

Livestock feeding behaviour is an influential research area in animal husbandry and agriculture. In recent years, there has been a growing interest in automated systems for monitoring the behaviour of ruminants. Current automated monitoring systems mainly use motion, acoustic, pressure and image sensors to collect and analyse patterns related to ingestive behaviour, foraging activities and daily intake. The performance evaluation of existing methods is a complex task and direct comparisons between studies is difficult. Several factors prevent a direct comparison, starting from the diversity of data and performance metrics used in the experiments. This review on the analysis of the feeding behaviour of ruminants emphasise the relationship between sensing methodologies, signal processing, and computational intelligence methods. It assesses the main sensing methodologies and the main techniques to analyse the signals associated with feeding behaviour, evaluating their use in different settings and situations. It also highlights the potential of the valuable information provided by automated monitoring systems to expand knowledge in the field, positively impacting production systems and research. The paper closes by discussing future engineering challenges and opportunities in livestock feeding behaviour monitoring.

Accurately identifying the fish feeding intensity plays a vital role in aquaculture. While traditional methods are limited by single modality (e.g., water quality, vision, audio), they often lack comprehensive representation, leading to low identification accuracy. In contrast, the multimodal fusion methods leverage the fusion of features from different modalities to obtain richer target features, thereby significantly enhancing the performance of fish feeding intensity assessment (FFIA). In this work a multimodal dataset called MRS-FFIA was introduced. The MRS-FFIA dataset consists of 7611 labelled audio, video and acoustic dataset, and divided the dataset into four different feeding intensity (strong, medium, weak, and none). To address the limitations of single modality methods, a Multimodal Fusion of Fish Feeding Intensity fusion (MFFFI) model was proposed. The MFFFI model is first extracting deep features from three modal data audio (Mel), video (RGB), Acoustic (SI). Then, image stitching techniques are employed to fuse these extracted features. Finally, the fused features are passed through a classifier to obtain the results. The test results show that the accuracy of the fused multimodal information is 99.26%, which improves the accuracy by 12.80%, 13.77%, and 2.86%, respectively, compared to the best results for single-modality (audio, video and acoustic dataset). This result demonstrates that the method proposed in this paper is better at classifying the feeding intensity of fish and can achieve higher accuracy. In addition, compared with the mainstream single-modality approach, the model improves 1.5%–10.8% in accuracy, and the lightweight effect is more obvious. Based on the multimodal fusion method, the feeding decision can be optimised effectively, which provides technical support for the development of intelligent feeding systems.

This study investigates the passive vibration dynamics of a sweep tool in a laboratory soil bin test, employing various spring configurations. A discrete element method (DEM) model of simulating the passively vibrating sweep tool was developed based on the laboratory soil bin tests. Ensuring precision in the DEM model parameters was achieved by applying a genetic algorithm tailored for this purpose. The genetic algorithm revealed that within the particle assemblies of the three geometries used in the DEM, several parameter sets were suitable for accurately describing the modelled soil. The final parameter set was chosen by integrating the DEM model with results from the laboratory direct shear box test. Employing Fast Fourier Transformation, both the laboratory soil bin test and the calibrated DEM model of the soil and the vibrating sweep tool facilitated an examination of frequencies and amplitudes during force and displacement measurements. The results indicated that, compared to a rigid tool, the draught force required by the 16 spring sweep tool was reduced by 6–9%. The absence of DEM would have limited the investigation of kinetic energy in the sweep tool and the dynamics of energy dissipation in the soil, if measurement equipment alone was used. This research successfully demonstrated that the reduced draught force with the 16 spring passively vibrating sweep tool, operating near the system's eigenfrequency, resulted from its ability to generate higher kinetic energy in the sweep tool while minimising energy dissipation in the soil.

The real-time and high-precision detection methods on embedded platforms are critical for harvesting robots to accurately locate the position of the table grapes. A novel detection method (ESP-YOLO) for the table grapes in the trellis structured orchards is proposed to improve the detection accuracy and efficiency based on You Only Look Once (YOLO), Efficient Layer Shuffle Aggregation Networks (ELSAN), Squeeze-and-Excitation (SE), Partial Convolution (PConv) and Soft Non-maximum suppression (Soft_NMS). According to cross-group information interchange, the channel shuffle operation is presented to modify transition layers instead of the CSPDarkNet53 (C3) in backbone networks for the table grape feature extraction. The PConv is utilised in the neck network to extract the part channel's features for the inference speed and spatial features. SE is inserted in backbone networks to adjust the channel weight for channel-wise features of grape images. Then, Soft_NMS is modified to enhance the segmentation capability for densely clustered grapes. The algorithm is conducted on embedded platforms to detect table grapes in complex scenarios, including the overlap of multi-grape adhesion and the occlusion of stems and leaves. ELSAN block boosts inference speed by 46% while maintaining accuracy. The [email protected]:0.95 of ESP-YOLO surpasses that of other advanced methods by 3.7%–16.8%. ESP-YOLO can be a useful tool for harvesting robots to detect table grapes accurately and quickly in various complex scenarios.

Camellia oleifera is an oilseed crop with high economic value. The short optimum harvest period and high labour costs of C. oleifera harvesting have prompted research on intelligent robotic harvesting. This study focused on the determination of grabbing points for the robotic harvesting of C. oleifera fruits, providing a basis for the decision making of the fruit-picking robot. A relatively simple 2D convolutional neural network (CNN) and stereoscopic vision replaced the complex 3D CNN to realise the 3D positioning of the fruit. Apple datasets were used for the pretraining of the model and knowledge transfer, which shared a certain degree of similarity to C. oleifera fruit. In addition, a fully automatic coordinate conversion method has been proposed to transform the fruit position information in the image into its 3D position in the robot coordinate system. Results showed that the You Only Look Once (YOLO)v8x model trained using 1012 annotated samples achieved the highest performance for fruit detection, with mAP₅₀ of 0.96 on the testing dataset. With knowledge transfer based on the apple datasets, YOLOv8x using few-shot learning realised a testing mAP₅₀ of 0.95, reducing manual annotation. Moreover, the error in the 3D coordinate calculation was lower than 2.1 cm on the three axes. The proposed method provides the 3D coordinates of the grabbing point for the target fruit in the robot coordinate system, which can be transferred directly to the robot control system to execute fruit-picking actions. This dataset was published online to reproduce the results of this study.

To address the automatic navigation issue of unmanned agricultural tractors affected by unknown disturbances, a path-tracking control scheme is proposed by utilising fixed-time nonsingular terminal sliding mode and adaptive disturbance observer technique. Firstly, a path-tracking kinematic model is established, which considers the unknown disturbances. Secondly, unlike conventional sliding mode controllers, a novel fixed-time terminal sliding mode controller is proposed for the unmanned agricultural tractor, which effectively enhances the dynamic performance and reduce the chattering effect. Furthermore, to reduce the detrimental effects of unknown disturbances, a new adaptive disturbance observer is designed to estimate and compensate these unknown disturbances. Subsequently, a strict Lyapunov analysis is conducted to confirm that the lateral and heading offsets of the unmanned agricultural tractor under the adaptive disturbance observer-based fixed time nonsingular terminal sliding mode control scheme can be stabilised to the arbitrarily small neighbourhood near the origin within a fixed time. Finally, extensive experiments were carried out to verify the effectiveness and advantages of the proposed control scheme.

A novel method for the mechanical detection of wheat biomass, based on the mechanical properties of wheat plants, is proposed to enable the quick assessment of wheat biomass. The mechanical model developed for the wheat plants, based on the variable cross-section beam elastic bending theory, can be used to analyse the interactive forces between the reel and wheat plants, and predict wheat biomass based on the magnitude of the force. The influence of wheat ears on deflection was incorporated into the model. The accuracy of wheat plant deflection forces obtained using the model was confirmed through theoretical analyses, simulations and experimental measurements. Moreover, deflection tests and posture analysis were performed on the wheat plants for different locations at which the deflection forces were acting and for different plant densities. Experiments focusing on reel operation demonstrated that the deflection forces exerted by the reel rod on wheat plants could be used to predict the number of bent plants, which would subsequently help in wheat biomass estimation. The study found that the influence of the wheat ear on the deflection force significantly increased as the plant deflection increased. The deflection force was most effective at two-thirds of the height of the wheat plant. Moreover, the higher the plant density, the greater the deflection force, which was closely correlated with wheat biomass. A model was established based on the results of the linear regression performed to determine the relationship between the deflection force acting on a wheat plant and its biomass. The model with a determination coefficient of 0.9155 provided a theoretical basis for detecting the feed quantity of the combine harvester.

Data scarcity is a hurdle for physiology-based precision agriculture. Measuring nutrient uptake by visible-near infrared spectroscopy implies collecting spectral and compositional data from low-throughput, such as inductively coupled plasma optical emission spectroscopy. This paper introduces data augmentation in spectroscopy by hybridisation for expanding real-world data into synthetic datasets statistically representative of the real data, allowing the quantification of macronutrients (N, P, K, Ca, Mg, and S) and micronutrients (Fe, Mn, Zn, Cu, and B). Partial least squares (PLS), local partial least squares (LocPLS), and self-learning artificial intelligence (SLAI) were used to determine the capacity to expand the knowledge base. PLS using only real-world data (RWD) cannot quantify some nutrients (N and Cu in grapevine leaves and K, Ca, Mg, S, and Cu in apple tree leaves). The synthetic dataset of the study allowed predicting real-world leaf composition of macronutrients (N, P, K, Ca, Mg and S) (Pearson coefficient correlation (R) ∼ 0.61–0.94 and standard error (SE) ∼ 0.04–0.05%) and micro-nutrients (Fe, Mn, Zn, Cu and B) (R ∼ 0.66–0.91 and SE ∼ 0.88–3.98 ppm) in grapevine leaves using LocPLS and SLAI. The synthetic dataset loses significance if the real-world counterpart has low representativity, resulting in poor quantifications of macronutrients (R ∼ 0.51–0.72 and SE ∼ 0.02–0.13%) and micronutrients (R ∼ 0.53–0.76 and SE ∼ 8.89–37.89 ppm), and not allowing S quantification (R = 0.37, SE = 0.01) in apple tree leaves. Representative real-world sampling makes data augmentation in spectroscopy very efficient in expanding the knowledge base and nutrient quantifications.

Mounting behaviour is an important characteristic of cows during oestrus. Real-time and accurate detection of cow mounting behaviour can shorten the calving-to-conception period and increase the economic benefits for dairy farms. Cow mounting behaviour occurs more often at night, and drastic scale changes in surveillance images caused by different distances between cows and camera, influence the detection of cow mounting. Existing methods are unable to address these challenges effectively. To address these challenges, this study collected 9392 images of Holstein cow mounting behaviour under intensive farming conditions using cameras and proposed an IATEFF-YOLO that is more suitable for cow mounting behaviour detection at nighttime and drastic scale changes in surveillance images caused by different distances between cows and camera. IATEFF-YOLO comprises an Illumination Adaptive Transformer (IAT) and an efficient feature fusion detector. The IAT enhances low-light images at night to enrich the cow mounting features, facilitating the subsequent detection of mounting behaviour. The efficient feature fusion detector, EFF-YOLO, enhances the feature fusion capability and further enable the detector to adapt to drastic scale changes in surveillance images caused by different distances between cows and camera. IATEFF-YOLO achieved a mean Average Precision of 99.3% and a detection speed of 102.0 f/s on test set. Compared with existing methods, IATEFF-YOLO achieved higher detection accuracy and faster detection speed during nighttime and drastic scale changes in surveillance images caused by different distances between cows and camera. IATEFF-YOLO can assist ranch breeders in achieving round-the-clock monitoring of cow oestrus, thereby enhancing oestrus detection efficiency.

The rice dry direct-seeding planting mode is a typical shallow sowing operation, and the traditional compacting mechanism with only longitudinal profiling ability is difficult to ensure the seedbed uniformity, resulting in the seedling emergence rate always lower than 80%. This study innovatively proposed a novel bidirectional-micro-profiling compacting device (BMPCD). In this study, the coupled DEM-MFBD simulation technique was utilised to find that the core design parameters k (elasticity coefficient of the reset spring) and t (thickness of the elastic sheet) of the BMPCD would significantly affect the seedbed uniformity by changing the resistance value F_r during the profiling process (P ≤ 0.01). The simulation results showed that when k was taken as 7.8 N mm⁻¹ and t was taken as 1.6 mm, the seedbed uniformity could be most greatly improved. The field experiments showed that compared with the bidirectional profiling compacting device (BPCD) and longitudinal profiling compacting device (LPCD), BMPCD could reduce the coefficient of variation of soil firmness (CVSF) by 33.1% and 40.1%, and the coefficient of variation of sowing depth (CVSD) by 37.1% and 51.8%, respectively, and then improve the seedling emergence rate of dry direct-seeded rice by 5.8% and 12.2%. This indicated that bidirectional and micro-profiling compaction technology could tackle the problem of low seedling emergence rate in rice dry direct-seeding. Meanwhile, the results of the DEM-MFBD coupling simulation were not significantly different from the test results of the field experiments (P > 0.05), indicating that it could be used as an efficient and accurate new method to study the dynamic characteristics between the soil and machinery.