Accurate weed detection is essential for the precise control of weeds in wheat fields, but weeds and wheat are sheltered from each other, and there is no clear size specification, making it difficult to accurately detect weeds in wheat. To achieve the precise identification of weeds, wheat weed datasets were constructed, and a wheat field weed detection model, YOLOv8-MBM, based on improved YOLOv8s, was proposed. In this study, a lightweight visual converter (MobileViTv3) was introduced into the C2f module to enhance the detection accuracy of the model by integrating input, local (CNN), and global (ViT) features. Secondly, a bidirectional feature pyramid network (BiFPN) was introduced to enhance the performance of multi-scale feature fusion. Furthermore, to address the weak generalization and slow convergence speed of the CIoU loss function for detection tasks, the bounding box regression loss function (MPDIOU) was used instead of the CIoU loss function to improve the convergence speed of the model and further enhance the detection performance. Finally, the model performance was tested on the wheat weed datasets. The experiments show that the YOLOv8-MBM proposed in this paper is superior to Fast R-CNN, YOLOv3, YOLOv4-tiny, YOLOv5s, YOLOv7, YOLOv9, and other mainstream models in regards to detection performance. The accuracy of the improved model reaches 92.7%. Compared with the original YOLOv8s model, the precision, recall, mAP1, and mAP2 are increased by 10.6%, 8.9%, 9.7%, and 9.3%, respectively. In summary, the YOLOv8-MBM model successfully meets the requirements for accurate weed detection in wheat fields.

The most important part of fruit cultivation is transplanting, which can be completed efficiently by an automatic transplanter. In this study, an automatic solanum lycopersicum transplanter was developed. It is primarily composed of the following devices: a seedling tray fixator, a mechanical transplanting arm, double horizontal driving modules, double vertical driving cylinders, a seedling separation device, double-crank five-bar planting devices, a power distribution system, and a PLC control system. An experimental test on an automatic transplanter was carried out on dry land. The experimental results showed that when the planting frequency was 80 plants/(min·row), the transplanting success rate was 93.89%, the missed planting rate was 1.58%, the replanting rate was 0.65%, the lodging rate was 1.94%, and the exposed seedling rate was 1.94%. Each device of the automatic transplanter was coordinated to complete the transplanting process. The automatic transplanter met the operation requirements, and it not only transplanted in dry land but also provided a theoretical basis for fruit cultivation in solar greenhouses in the future.

Collision is one of the main causes of mechanical damage to the seedling during transplanting. To reveal the impact damage characteristics of plug seedlings, the kinetics equations of seedling collision were established based on Hertz’s contact theory, and the kinematic characteristics, elastoplastic deformation, and collision damage during seedling collision were analyzed using high-speed photography. Using the Tekscan pressure distribution measurement system, the significant levels of various factors that affect impact peak force (IPF) and damage of seedling pot (DSP) were studied, the change rule of contact pressure distribution of seedlings under significant factors was measured, and a regression model between IPF and DSP was established. The results showed that collision material, drop height and seedling pot size had significant effects on IPF and DSP. The contact pressure area of different collision materials from large to small was soil block, steel, and ABS plastic. The contact pressure area of seedlings of different pot sizes was big, medium, and small in descending order. At a dropping height of 50~350 mm, a contact pressure > 10 kPa accounted for the major contact pressure area, which is the main reason for collision damage of the seedling pot. Linear regression models between IPF and DSP under different factors were established, and the determination coefficients (R²) were 0.98 and 0.94, respectively. The results provided a theoretical basis for understanding the collision damage mechanism of the plug seedling and how to reduce damage during transplanting.

There is a significant difference between the simulation effect and the actual effect in the design process of maize straw-breaking equipment due to the lack of accurate simulation model parameters in the breaking and processing of maize straw. This article used a combination of physical experiments, virtual simulation, and machine learning to calibrate the simulation parameters of maize straw. A bimodal-distribution discrete element model of maize straw was established based on the intrinsic and contact parameters measured via physical experiments. The significance analysis of the simulation parameters was conducted via the Plackett–Burman experiment. The Poisson ratio, shear modulus, and normal stiffness of the maize straw significantly impacted the peak compression force of the maize straw and steel plate. The steepest-climb test was carried out for the significance parameter, and the relative error between the peak compression force in the simulation test and the peak compression force in the physical test was used as the evaluation index. It was found that the optimal range intervals for the Poisson ratio, shear modulus, and normal stiffness of the maize straw were 0.32–0.36, 1.24 × 10⁸–1.72 × 10⁸ Pa, and 5.9 × 10⁶–6.7 × 10⁶ N/m³, respectively. Using the experimental data of the central composite design as the dataset, a GA–BP neural network prediction model for the peak compression force of maize straw was established, analyzed, and evaluated. The GA–BP prediction model’s accuracy was verified via experiments. It was found that the ideal combination of parameters was a Poisson ratio of 0.357, a shear modulus of 1.511 × 10⁸ Pa, and a normal stiffness of 6.285 × 10⁶ N/m³ for the maize straw. The results provide a basis for analyzing the damage mechanism of maize straw during the grinding process.

To improve the accuracy of simulation parameters used in discrete element simulation tests for the transplanting operation of the transplanting machine and to facilitate further optimization of crucial components of the transplanting machine, in this paper, the discrete element model of 50-hole plug seedling pots was calibrated and optimized based on the collision impact force between the plug seedling pot and the steel plate measured by a flexible film network tactile pressure sensor. Basic tests determined the contact parameters of the pot, and the initial parameters were screened for significance using the Plackett–Burman test. The pot-steel static friction coefficient, the pot-pot collision restitution coefficient, and the bond radius significantly affected the simulated collision impact force between the pot and the steel plate. According to the relative error value of the impact force between the pot and the steel plate as the evaluation index, the steepest climbing test was carried out on three significant parameters to optimize their value range. Based on the Box–Behnken test, a second-order regression model of the impact force and significant parameters regulating the interaction between the pot and the steel plate was established, where the target impact force between the pot and the steel plate was 11.78 N. The optimal parameter combination is obtained by optimizing the significance parameters: the static friction coefficient between the pot and steel is 0.790, the collision restitution coefficient between the pot and the pot is 0.325, and the bond radius is 1.542 mm. The test results show that the relative error between the actual and simulation tests is only 0.084%. The calibrated parameters of the discrete element model of plug seedling pots are accurate and reliable. The research results presented here can provide a reference for the subsequent transplanting operation simulation of the transplanter.