Hua Li

Continuously acquired garlic images can contain temporally related observations that affect soil residue grading estimates. This study assessed data-partitioning sensitivity and within-session acquisition-stage transfer using a two-stage You Only Look Once version 26 nano (YOLO26n) detector and EfficientNet-B0 classifier. The dataset comprised 754 red–green–blue (RGB) scene images and 10,721 bulb instances collected during one controlled session in which the conveyor operated at 0.5 m/s, while active vibratory soil removal was disabled (0 Hz vibration frequency and 0° screen inclination); inter-annotator agreement was high (quadratic-weighted Cohen’s κ = 0.8750; 360 instances). For seed 42, macro-averaged F1-score (Macro-F1) was 97.49% with random image splitting, 87.43% with 2 s temporal exclusion, and 76.32% with complete stage isolation. Three additional temporal-exclusion runs yielded 85.92 ± 1.38% Macro-F1. On a fixed S07 test set (26 scenes, 405 bulbs), replacing temporally neighboring training images with more distant same-stage images, while preserving training-scene and class-specific instance counts, changed Macro-F1 from 99.76% to 74.12%. In a three-seed, four-fold leave-one-stage-out (LOSO) evaluation, stage-based group distributionally robust optimization (Stage-GroupDRO) achieved 75.57% Stage-macro Macro-F1 versus 74.11% for class-weighted empirical risk minimization (Weighted-ERM), but the paired 95% confidence interval included zero. The full pipeline achieved 72.82% pooled end-to-end Macro-F1 at 19.32 frames per second (FPS) during offline evaluation. These results demonstrate substantial sensitivity to data partitioning within one acquisition session; external generalization, quantitative soil-removal efficiency, and closed-loop control remain untested.

With increasing industrialization, heavy metal contamination poses a growing threat to ecosystem stability and human health. Cadmium (Cd) is taken up by medicinal plants through the soil–plant continuum, altering rhizosphere microbial communities and thereby affecting plant metabolism and bioactive compound accumulation. In this study, Dendrobium officinale Kimura & Migo (D. officinale) was used as a model medicinal plant to investigate the effects of Cd exposure on plant quality and rhizosphere microbial community composition. The results show that different D. officinale strains exhibited significant differences in growth traits and Cd accumulation under Cd exposure. Under the 0.9 mg·kg⁻¹ Cd treatment, mean Cd concentrations in the strain with the highest Cd accumulation (Strain 2) reached 6.42, 1.08, and 1.70 mg·kg⁻¹ DW in roots, stems, and leaves, respectively, compared with 2.13, 0.73, and 0.66 mg·kg⁻¹ DW in the strain with the lowest Cd accumulation (Strain 1), corresponding to approximately 3.0-, 1.5-, and 2.6-fold higher accumulation. Cd exposure was associated with strain-specific changes in metal transport-related gene expression, antioxidant responses, bioactive constituent accumulation, and rhizosphere microbial community composition. Together, these changes highlight the complex relationship between Cd accumulation, plant physiological responses, and medicinal quality. These findings provide a theoretical reference for the selection of elite D. officinale strains and for the safe cultivation of this medicinal plant.

Accurate reconstruction of the multilayer structure and progressive fracture of high-moisture garlic stems and leaves underpins prediction of rotary-shredding quality and operating loads. Stable transfer of tissue-scale parameters to intact-stem and machine-scale responses is critical. Existing discrete element models cannot describe contact indentation, post-peak softening, and sustained load bearing before complete separation. A heterogeneous Hysteretic Spring-Elastic-Progressive Damage Bond (HS-EPDB) model was therefore developed. Within a hierarchical calibration framework for base contact, intratissue bonds, and tissue-interface bonds, compression and tensile tests sequentially determined the base-contact parameters and bond parameters between tissues; all errors were within 9.01%. Intact-stem three-point bending validation yielded errors of 1.86% and 3.66% in peak force and corresponding displacement, and 4.65%, 2.75%, and 2.98% in pre-peak, early post-peak, and common-boundary absorbed work. It reproduced morphological progression from midspan indentation through continuous bending to progressive failure. Rotary-shredding experiments and simulations evaluated machine-scale transfer. A controlled comparison of four configurations with shared calibrated parameters and equivalent deletion thresholds distinguished the effects of base-contact formulation and bond-failure path. Compared with elastic recovery and abrupt, threshold-triggered bond deletion, HS-EPDB reduced nonphysical rebound in high-moisture tissues and preserved load transfer through damaged connections until complete failure. It unified macroscopic load bearing, energy dissipation, and progressive failure, while jointly representing shredding quality, fragment-length composition, and dynamic cutter-roll loads. The approach enables reliable cross-scale transfer from tissue tests to machine-scale shredding simulations for layered high-moisture crop materials and supports equipment optimization.

Targeting the insufficient attention given to the real-time prediction of digging resistance in traditional research, which makes it difficult to achieve drag and consumption reduction during the mechanized harvesting of underground crops, this study proposes a real-time digging resistance prediction method integrating Butterworth low-pass filtering, wavelet denoising, and a Long Short-Term Memory (LSTM) network. First, a Butterworth low-pass filter is employed to extract the low-frequency trend component of the resistance signal, while wavelet denoising is applied to the high-frequency component that characterizes the excitation and cutting properties of vibration digging. Subsequently, the denoised high-frequency component and the low-frequency component are reconstructed to form a new resistance signal, based on which a residual LSTM prediction model is established for training and prediction. Finally, the coefficient of determination (R²) and the root mean square error (RMSE) are selected as evaluation indicators to assess the performance of the proposed method, and ablation experiments are conducted to further verify the necessity of the hybrid denoising strategy in enhancing prediction performance. The results indicate that the R² and RMSE of the proposed method reach 0.9010 and 442.76 N, respectively. Compared with existing time-series prediction models, this method exhibits superior prediction performance. The hybrid denoising mechanism increases the R² by 21.9% and reduces the RMSE by 16.6%.

Recent advances in nickel-based catalysts for plastic pyrolysis, emphasizing reaction mechanisms, catalyst design, and emerging innovations that enhance product selectivity, yield efficiency, and sustainable polymer recycling.

With the rapid expansion of industrial activities, the accumulation of heavy metals in the environment has become a serious threat to ecological security and public health. Rhizosphere microorganisms play a crucial role in supporting the growth and quality of medicinal plants by facilitating nutrient uptake and regulating hormonal balance. However, medicinal plants can absorb heavy metals from contaminated soils during growth, resulting in toxic metal accumulation in plant tissues and reduced efficacy of active compounds. At the same time, excessive heavy metal levels suppress rhizosphere microbial growth and activity, disrupt community structure and function, and weaken their beneficial interactions with plants. These processes collectively lead to soil fertility decline, hindered plant development, and compromised safety and quality of medicinal materials. This review systematically summarizes the mechanisms by which heavy metals affect medicinal plants and their rhizosphere microbiota, and highlights that future research should focus on elucidating these interactions, developing advanced remediation technologies, and establishing comprehensive monitoring systems for the quality and safety of medicinal plants, thereby providing a scientific basis for their safe utilization and quality improvement.

In conventional mechanized garlic seeding process, seed remains a persistent challenge that is difficult to avoid. This study proposes a solution by designing and testing a garlic seeding device based on a rigid–flexible coupling mechanism, aimed at minimizing seed damage during sowing. The seeding pocket was constructed from a flexible metal sheet, which served as its structural foundation. A slider moving along a fixed track enabled the retraction and release of the pocket, thereby facilitating seed collection and discharge. The effects of pocket radius, rotational speed of seed discharge disc, and thickness of metal sheet on the stress of garlic seeds were investigated through the finite element method. Subsequently, an experimental bench was set up to analyze the effects of influence of these parameters on seed damage rate, single-seed rate, and leakage rate. Results demonstrated that under optimal parameters—a pocket radius of 12 mm, a seed discharge disc rotational speed of 0.21 rad/s, and a metal sheet thickness of 0.15 mm—the mechanism achieved a single-seed rate of 78.4%, a leakage rate of 11.4%, and a maximum stress on garlic seeds of only 0.535 MPa. Notably, this stress level was well below the damage threshold of garlic seeds, resulting in zero damage that outperformed conventional rigid seeding devices. These findings demonstrate the mechanism’s strong potential to preserve seed integrity, although the overall seeding performance remains modest and warrants further optimization in future designs.

In the context of the conventional mechanized garlic seeding process, the issue of damage to garlic seeds is a pervasive challenge that is challenging to circumvent. This paper proposes a solution to this problem by designing and testing a garlic seeding device that incorporates a rigid-flexible coupling mechanism. The purpose of this innovation is to mitigate the damage to garlic seeds during the sowing process. Firstly, the seeding pocket is constructed from a flexible metal sheet, which functions as its structural foundation. A slider along the fixed track enables the retraction and release of the pocket, thereby facilitating the process of collecting and dropping garlic seeds. Secondly, the effects of pocket radius, rotational speed of seed discharge disc and thickness of metal sheet on the stress of garlic seeds were investigated through finite element method. Finally, the experimental bench was set up to analyze the effects of pocket radius, rotational speed of seed discharge disc and thickness of metal sheet on the damage rate, single-seed rate and leakage rate. Results demonstrated that under optimal parameters—a pocket radius of 12 mm, a seed discharge disc rotational speed of 0.21 rad/s, and a metal sheet thickness of 0.15 mm—the proposed mechanism achieved a single-seed rate of 78.4%, a leakage rate of 11.4%, and a maximum stress on garlic seeds of only 0.535 MPa. Significantly, this stress level is well below the damage threshold of garlic seeds, resulting in a remarkably zero damage that outperforms conventional rigid seeding devices. This demonstrates the mechanism's remarkable capacity to safeguard the integrity of the seeds, even though the effect of the seeding process is comparatively low and shows considerable scope for enhancement in future iterations.