This study aimed to develop a residual error-based artificial neural network integrated mathematical modelling framework to improve the prediction of moisture ratio and effective moisture diffusivity during convective hot-air drying of Coccinia grandis. The drying was conducted using slices of 5, 10, 15, and 20 mm thickness at different temperatures of 60, 70, and 80°C in convective hot air dryer. The experimental drying data were fitted using 10 thin-layer drying models and three two-factor effective moisture diffusivity mathematical models. Furthermore, the residual errors from each mathematical model were used to train an artificial neural network, which was then integrated with the corresponding mathematical model to minimize prediction error. Findings suggested that increasing the drying temperature accelerated the moisture removal and increased the effective moisture diffusivity. Furthermore, an increase in the thickness of the sample prolonged the drying time. Among different mathematical models, the “Midilli model” and “Balbay and Sahin model” showed the best performance for moisture ratio, while the quadratic polynomial equation best predicted effective moisture diffusivity. Subsequently, the integration of residual error-based artificial neural networks with different thin-layer models and two-factor mathematical models improved the overall prediction performance, with only slight reductions under several drying conditions, although unseen-data validation indicated differences in predictive performance across parameters, with relatively higher errors observed for effective moisture diffusivity. The developed model can improve mathematical model predictions during hot-air drying of Coccinia grandis. However, its effectiveness depends on the underlying mathematical model, drying conditions, and dataset characteristics.
Hydroponic lettuce is prone to rot at the bottom leaves, which are unevenly distributed and may overlap with each other. Manual removal is labor-intensive and inefficient, while mechanical removal is hindered by occlusion. This study aims to develop a visual method for intelligent rotten leaf detection and occlusion discrimination to enable automated removal. An improved YOLOv11n_CGS model incorporating SPDConv, SHSA, and GD mechanism was developed for oriented rotten leaf detection. The region of interest (ROI) was extracted using the YOLOv11s model, and the occlusion relationship of rotten leaves was determined based on the curvature features of leaf contours within the ROI. The rotational angle of the flexible gripper was then adjusted according to the visual detection results. The YOLOv11n_CGS model achieved a mean average precision at IoU@0.5 (mAP@0.5) of 96.4
Ornamental crop production is a rapidly developing sector of US agriculture. Accurate inventory management and quality assessment are critical tasks for monitoring sales, yield, space requirements, production timing, and quality improvements of ornamental nurseries. Current inventory practices rely on manual counting, which is labor-intensive and often inaccurate. This study develops an automated system for plant counting and quality assessment of container nursery stock. An unmanned teleoperated ground vehicle, “NurseryBot,” was developed for plant counting and quality assessment of container nursery stock. The NurseryBot utilized RGB cameras to collect video data from a downward-facing perspective. A deep learning framework based on YOLOv11 model variants was trained and tested on a panoramic dataset to detect and segment plants efficiently. Cross-species knowledge adaptation approaches were explored to improve segmentation generalization. Morphological parameters, including area, perimeter, eccentricity, and circularity, were extracted to systematically evaluate plant quality using image-based parameters. The YOLOv11-m instance segmentation model showed the highest precision, recall, and counting accuracy among the evaluated variants for plant detection and segmentation. For bounding boxes, the model attained precision, recall, and mAP50 values of 0.971, 0.970, and 0.941 for azalea and 0.968, 0.966, and 0.915 for sunshine, respectively. For mask segmentation, the model achieved precision, recall, and mAP50 values of 0.969, 0.968, and 0.901 for azalea and 0.967, 0.966, and 0.907 for sunshine, respectively. The proposed system achieved counting accuracies of 95.59
This study aimed to develop gelatin-based nanoparticles (GNPs) as multifunctional nanoplatforms that function both as intrinsic nanofertilizers and as nanocarriers for controlled urea delivery to enhance plant growth and nitrogen utilization efficiency. GNPs were fabricated using a two-step desolvation method and characterized by scanning electron microscopy (SEM) and Fourier-transform infrared (FT-IR) spectroscopy. Urea release behavior was evaluated through in vitro release studies. The growth-promoting effects of GNPs and urea-loaded GNPs were assessed using cabbage (Brassica oleracea) and radish (Raphanus sativus) grown on Murashige and Skoog (MS) medium. The synthesized GNPs exhibited a uniform spherical morphology with an average particle size of approximately 152 nm. FT-IR analysis confirmed successful nanoparticle formation and structural integrity of the gelatin matrix. Urea-loaded GNPs provided sustained nitrogen release over 15 days. GNP treatment significantly enhanced germination, root growth, leaf expansion, shoot development, and biomass accumulation in both crop species. Moreover, urea-loaded GNPs outperformed free urea and GNP-only treatments, suggesting that GNP-mediated urea delivery may have contributed to the observed improvements in plant growth. GNPs represent an eco-friendly nanomaterial with potential for promoting plant growth and improving fertilizer use efficiency. The observed growth-promoting effects of GNP treatment highlight their potential application as a nanofertilizer for sustainable crop production.
Alfalfa is a perennial legume forage crop harvested several times a year, and identifying the primary factors that influence stem cutting effectiveness could substantially improve harvesting efficiency. While stem cutting directly impacts mechanical energy consumption and operational efficiency, existing research has struggled to clarify the complex interactions between simultaneously acting multiple blade variables. To address these limitations, we developed a dataset focusing on maximum cutting force for alfalfa stems by evaluating the blade’s approach angle, shear angle, sliding angle, cutting speed, and sharpness index through one-factor-at-a-time (OFAT) experiments and fractional factorial and Box–Behnken designs. Random forest, extra trees, and CatBoost regression models were then trained on this dataset and compared with a conventional second-order response-surface model. CatBoost outperformed the other models, achieving a coefficient of determination of 0.7013 and a root-mean-square error of 0.5723 on the Yeo–Johnson transformed scale, corresponding to 2.885 N on the original force scale. Variable importance and SHAP analyses identified stem cross-sectional area as the most important predictor of maximum cutting force overall and, among the controllable blade design variables, the sharpness index and approach angle as the dominant predictors. Among the geometric blade angle pairs, the shear angle × approach angle interaction yielded the widest predicted force range. While the individual importance of the shear angle was low, it showed significant interaction effects when combined with other blade variables that substantially altered predictions. The CatBoost model was employed as the fitness function of a genetic algorithm to minimize cutting force, yielding optimal design ranges of 40.2–42.3° for shear angle, 33.9–40° for approach angle, 24.1–29.54° for sliding angle, 400–404.41 mm/min for cutting speed, and 186.19–233.04 g for the sharpness index. These ranges are preliminary, model-predicted optima obtained under quasi-static conditions of 100–500 mm/min and require experimental validation before application to field harvesters.
