In this paper, a video-based behavioural recognition dataset for beef cattle is constructed. The dataset covers five behaviours of beef cattle: standing, lying, drinking, feeding, and ruminating. Six beef cows in a captive barn were selected and monitored for 168 h. Different light conditions and nighttime data were considered. The dataset was collected by one surveillance video camera. The data collection process required deploying cameras, memory, routers and laptops. Data annotation was automated using the YOLOv8 target detection model and the ByteTrack multi-target tracking algorithm to annotate each beef cow’s coordinates and identity codes. The FFmpeg tool cut out individual beef cow video clips and manually annotated them with behavioural labels. The dataset includes 500 video clips, 2000 image recognition samples, over 4000 target tracking samples, and over 10G of frame sequence images. 4974 video data of different behavioural types are labelled, totalling about 14 h. Based on this, a TimeSformer multi-behaviour recognition model for beef cattle based on video understanding is proposed as a baseline evaluation model. The experimental results show that the model can effectively learn the corresponding category labels from the behavioural category data of the dataset, with an average recognition accuracy of 90.33% on the test set. In addition, a data enhancement and oversampling strategy was adopted to solve the data imbalance problem and reduce the risk of model overfitting. The dataset provides a data basis for studying beef cattle behaviour recognition. It is of great significance for the intelligent perception of beef cattle health status and improvement of farming efficiency.
The primary objective of this research was to develop an efficient method for accurately identifying and localizing weeds in ecological tea garden environments, aiming to enhance the quality and yield of tea production. Weed competition poses a significant challenge to tea production, particularly due to the small size of weed plants, their color similarity to tea trees, and the complexity of their growth environment. A dataset comprising 5366 high-definition images of weeds in tea gardens has been compiled to address this challenge. An enhanced U-Net model, incorporating a Double Attention Mechanism and an Atrous Spatial Pyramid Pooling module, is proposed for weed recognition. The results of the ablation experiments show that the model significantly improves the recognition accuracy and the Mean Intersection over Union (MIoU), which are enhanced by 4.08% and 5.22%, respectively. In addition, to meet the demand for precise weed management, a method for determining the center of weed plants by integrating the center of mass and skeleton structure has been developed. The skeleton was extracted through a preprocessing step and a refinement algorithm, and the relative positional relationship between the intersection point of the skeleton and the center of mass was cleverly utilized to achieve up to 82% localization accuracy. These results provide technical support for the research and development of intelligent weeding equipment for tea gardens, which helps to maintain the ecology of tea gardens and improve production efficiency and also provides a reference for weed management in other natural ecological environments.
In modern agriculture, reducing the internal moisture content of crop seeds is essential to enhance the activity and mobility of seed oil molecules, thereby increasing oil yield while minimizing the risk of mold and deterioration. However, traditional drying methods often result in uneven heating, leading to seed scorching and diminished drying efficiency and economic returns. To address these limitations, this study proposes a novel thin-layer seed drying system incorporating a redesigned drying tray structure. Specifically, the system places the seed-bearing tray beneath a vibration module operating at a predetermined frequency. The vibration mechanism induces the uniform motion of the seeds, thereby preventing localized overheating (scalding) and enabling automatic weighing for the real-time monitoring of moisture reduction during the drying process. The advancement of wireless sensor technologies in agriculture has enabled the deployment of more refined, large-scale monitoring networks. In this work, a commercial chip-based piezoelectric vibration detection device was integrated into the experimental setup to collect time-domain response signals resulting from interactions among seeds, impurities, and the drying tray. These signals were used to construct a comprehensive database of seed collision signatures. To mitigate discontinuities in signal transmission caused by vibration and potential equipment failure, the shortest routing protocol (SRP) was implemented. Additionally, the system outage probability (OP) and a refined closed-form solution for signal transmission reliability were derived under a Rayleigh fading channel model. To validate the proposed method, a series of experiments were conducted to determine the optimal vibration frequencies for various seed types. The results demonstrated a reduction in seed scalding rate to 1.5%, a decrease in seed loss rate to 0.4%, and an increase in moisture monitoring accuracy to 97.0%. Compared to traditional drying approaches, the vibrating drying tray substantially reduced seed loss and effectively distinguished between seeds and impurities. Furthermore, the approach shows strong potential for broader applications in seed classification and moisture detection across different crop types.
In complex mountainous environments, the asynchronous development between external color turning and internal sugar accumulation (often termed “false maturity”) in coffee cherries poses a severe challenge to post-harvest quality sorting and the consistency of final coffee products. To overcome the limitations of single-phenotype detection in raw material screening, this study proposed a multimodal quality discrimination framework integrating fruit hyperspectral imaging, micro-topography, and plant physiological characteristics. Taking typical mountain-grown fresh coffee cherries as the research object, and after comparing various spectral preprocessing and feature dimensionality reduction algorithms, the multimodal fusion efficacy of nine machine learning classifiers was systematically evaluated. The results demonstrated that: (1) Full-spectrum difference analysis quantitatively confirmed the limitations of visual harvesting; spectral reflectance differences between high- and low-sugar fruits were highly concentrated in the red and red-edge regions, with the maximum difference precisely located at 676 nm. (2) Compared to the single-spectrum model (mean accuracy of 75.93%), the fully fused Multilayer Perceptron (MLP) network effectively mitigated background noise induced by heterogeneous environments, improving the mean classification accuracy to 77.22% with a mean Area Under the Curve (AUC) of 0.827. (3) Correlation analysis clarified the quantitative association between topography and quality; micro-topographic slope (r = 0.346) was identified as the key environmental driver of spatial differentiation in fruit sugar content, while plant chlorophyll A content (r = 0.183) exhibited a corresponding physiological response trend. This study not only explains the root cause of visual assessment failure from a physical optics perspective but also reveals the spatial variation laws of quality driven by micro-topography, providing preliminary data support for the intelligent sorting of raw materials and ensuring post-harvest quality consistency of mountainous crops.
Yunnan ‘Bingtang’ oranges possess significant market competitiveness due to their unique taste and high sugar content. However, the limited efficiency of traditional quality inspection methods severely hampers the industrial grading process. Sugar content (Brix), as a core indicator of internal quality, is also a key factor determining consumer repurchase rates. This study focuses on the characteristic ‘Bingtang’ orange variety “Yunguan Orange” in the Longling Chu Orange Demonstration Base in Yunnan Province. Hyperspectral imaging data and corresponding sugar content values were collected during three critical stages from fruit set to harvest, and sugar content prediction research was conducted within the 400–1000 nm spectral range. Based on the extraction of regions of interest (ROI), Multispectral Scattering Correction (MSC) coupled with the Savitzky-Golay first-order derivative (SG-1st Deriv) method was employed to preprocess the raw spectral data (RAW). This study focused on investigating the performance of the Competitive Adaptive Weighted Sampling (CARS) algorithm across six prediction models. Raw full spectra (RAW) and the Sequential Projection Algorithm (SPA) were adopted as comparative schemes in ablation experiments to systematically analyze the influence of different feature extraction strategies. Experimental results showed that the coefficient of determination (R2) for most models exceeded 0.8 on the test set, while the root mean square error (RMSE) remained below 1.0. The findings indicate that under the current experimental conditions, this system can realize acceptable quantitative analysis for tested Yunguan oranges and greatly lower the overall model complexity. It clarifies the compatibility of the CARS algorithm with various prediction models, and systematically integrates three innovations of multi-stage sampling, SPA/CARS comparison and multi-type model evaluation, providing critical theoretical support and technical parameters for the development of low-cost, real-time online sorting equipment.