Accurate detection of tea buds and precise identification of plucking points are the essential prerequisites for intelligent tea-plucking robots. However, this task is challenging as the color of the targets exhibit color similarity with the background, and their shape, size, and morphology do not always coincide. Additionally, the tea buds are densely distributed and might undergo mutual occlusion within a visual horizon. In this paper, we draw inspiration from the analogy between plucking points and human keypoints and propose a framework, named PENet, that reformulates the plucking point localization task as a pose estimation problem. PENet splits the localization of the plucking points into two stages: tea bud detection and plucking point identification. In the first stage, we enhance a DINO-based detector by integrating spectrum attention module and multi-scale feature fusion, enhancing sensitivity to subtle color/texture contrasts. In the second stage, a modified High-Resolution Network with spatial coordinate encoding and bottlenecks transformer is used to accurately localize plucking points, even under occlusion or clustering. Owing to these two sequential stages, PENet can leverage the strengths of each stage to enhance flexibility on tea bud detection and plucking point identification in natural tea plantation environment. Through extensive experiments, PENet achieves competitive performance, with a mAP of 77.9% for tea bud detection and 86.2% for plucking point identification, outperforming several recent approaches under the same evaluation setting. These results demonstrate that PENet is a robust framework for intelligent tea harvesting—achieving high detection accuracy and plucking precision while reducing labor dependency and boosting operational efficiency in large-scale plantations.

Accurate detection of tea buds and identification of picking points are essential for automated tea harvesting. However, these tasks remain difficult in natural plantation environments due to dense clustering of buds, irregular spatial patterns, and frequent occlusions. To address these challenges, this study presents a two-stage perception framework that combines anchor-free dense tea bud detection with a graph reasoning approach to accurately identify picking points. In the first stage, an anchor-free dense tea bud detection strategy is adopted to avoid unstable anchor assignment in crowded scenes. It incorporates bounding box refinement with an Intersection over Union (IoU) class score to align detection confidence with geometric precision. In the second stage, the refined detections are exploited as structural cues for the occluded picking point identification module. A graph aware layer with relative position loss is employed to model spatial dependencies among picking points and auxiliary landmarks, enabling the inference of occluded targets based on learned structural cues. Experiments on a custom dataset of 5001 images demonstrate that the proposed framework achieves competitive performance compared with representative methods under the same evaluation protocol. Specifically, it achieves a mean Average Precision (mAP) of 60.9% for dense detection and 96.2% for identification of occluded picking points. The proposed framework has been deployed on embedded computing devices, achieving an inference speed of 28.50 frames per second (FPS) for detection and 8.88 ms per bud for picking point identification. These results demonstrate its feasibility for real-world tea harvesting applications.

To produce plug seedlings with uniform growth and which are suitable for high-speed transplanting operations, it is essential to sow seeds precisely at the center of each plug-tray hole. For accurately determining the position of the seed covered by the substrate within individual plug-tray holes, a novel method for detecting the growth points of plug seedlings has been proposed. It employs an adaptive grayscale processing algorithm based on the differential evolution extra-green algorithm to extract the contour features of seedlings during the early stages of cotyledon emergence. The pixel overlay curve peak points within the binary image of the plug-tray’s background are utilized to delineate the boundaries of the plug-tray holes. Each plug-tray hole containing a single seedling is identified by analyzing the area and perimeter of the seedling’s contour connectivity domains. The midpoint of the shortest line between these domains is designated as the growth point of the individual seedling. For laboratory-grown plug seedlings of tomato, pepper, and Chinese kale, the highest detection accuracy was achieved on the third-, fourth-, and second-days’ post-cotyledon emergence, respectively. The identification rate of missing seedlings and single seedlings exceeded 97.57% and 99.25%, respectively, with a growth-point detection error of less than 0.98 mm. For tomato and broccoli plug seedlings cultivated in a nursery greenhouse three days after cotyledon emergence, the detection accuracy for missing seedlings and single seedlings was greater than 95.78%, with a growth-point detection error of less than 2.06 mm. These results validated the high detection accuracy and broad applicability of the proposed method for various seedling types at the appropriate growth stages.

The finite element simulation is a valid way for the rapid development of the root-cutting mechanism for hydroponic Chinese kale. The stem of the hydroponic Chinese kale was simplified as a transverse isotropic elastic body, and axial compression, three-point bending, and shear tests were performed. The ANSYS/LS-DYNA19.2 software was adopted for stem shear simulation, and the regression equation of the maximum simulated shear force was established. The optimized mechanical parameters were determined by minimizing the deviation between the maximum shear force obtained from the simulation and test. The three-dimensional scanning method was employed to establish the geometric model of the hydroponic Chinese kale stem. The cutting finite element simulation model and test platform were constructed. Displacement, deformation, and force measured from simulation and test were compared. Through measurement and simulation calibration, an axial elastic modulus of 6.22 MPa, axial Poisson's ratio of 0.46, radial elastic modulus of 3.56 MPa, radial Poisson's ratio of 0.44, radial shear modulus of 0.8 MPa, and a failure strain of 0.08 were determined. During the cutting simulation and test, the resulting maximum displacement deviations of the marking points on the end of the stem were 0.68 mm along the X-axis and 2.83 mm along the Y-axis, while the maximum deviations of the cutting and clamping force were 0.49 N and 0.77 N, respectively. The deformation and force variation laws of the kale stem in the cutting simulation and test process were basically consistent. It showed that the mechanical parameters calibrated by the simulation were accurate and effective, and the stem cutting simulation results with the finite element method were in good agreement with that of the cutting test. The study provided a reference for the rapid optimization design of the root-cutting mechanism for hydroponic Chinese kale harvest.

Temperature prediction is important for controlling the environment in the preservation of fresh products. The phase change materials for cold storage make the heat transfer process complex, and the use of physical models for characterization and temperature prediction can be challenging. In order to predict the variation of the thermal environment in a temperature-controlled container with a cold energy storage system, we propose an LSTM model based on historical temperature data in which the trends of temperature variations of the fresh-keeping area, the phase change material (PCM), and the fresh products can be predicted immediately without considering the complex heat transfer process. An experimental platform of a temperature-controlled container with a cold energy storage system is built to obtain the experimental data for the prediction model’s construction and validation. The prediction results based on the LSTM model are compared to the results of a physical model. In order to optimize the input data for better prediction performance, the proportion of input samples from the dataset is set to 80%, 50%, 20%, and 10%. The prediction results from different input groups are compared and analyzed. The results show that the LSTM model is able to accurately predict temperature variations of the fresh-keeping area and products, and the predicted values are in agreement with the actual values. The LSTM-based prediction model has a higher accuracy compared to the physical-based prediction model; the RMSE, MAE, and MAPE are 0.105, 0.103, and 0.010, respectively, and the relative error for the prediction of effective control hours of environmental temperature is 0.92%. It is suggested to use the initial 20% of the historical temperature data as the input to predict the future temperature variation for better prediction performance. The results of this paper offer valuable insights for accurate temperature prediction in the fresh-keeping environment with a cold energy storage system.

The mechanical properties of tobacco leaf are very important for the research and design of tobacco harvesting machines. A numerical model was developed to simulate the tension behavior of tobacco leaf samples using the Discrete Element Method (DEM). The model was constructed with spherical particles bonded together. The model outputs were tensile strength (sigma macro) and Young's modulus (Emacro) of the leaf. Tensile tests were conducted to measure the same tensile properties of tobacco mesophyll samples obtained from different locations of the plant and at different sampling orientations. The model was calibrated and verified by comparing simulation results with the test data. Testing results showed that sampling location and orientation had marginal effects on the tensile properties of tobacco leaf. The average tensile strength was 0.57 MPa and the average Young's modulus was 3.50 MPa. The relationships between model outputs (sigma macro and Emacro) and the two most sensitive model parameters, Young's modulus of particles (Emicro) and the critical tensile stress of bonds (sigma micro) were well described using two equations. The calibrated values of sigma micro and Emicro were 8.02 and 23.07 MPa respectively using the equations and measured values of tensile strength and Young's modulus. The calibrated model was able to predict the tensile strength and Young's modulus of tobacco leaf with relative errors of less than 2 %. The proposed DEM model and method can be applied to other plant leaves in simulating their tensile behaviors and properties.

目的 获取冷藏运输箱在保鲜运输过程中风速结合货物参数对运输箱内温度变化影响的一般规律.方法 以装载荔枝的冷藏运输箱为研究对象,采用计算流体动力学(Computational Fluid Dynamics)方法建立装载荔枝的考虑荔枝呼吸热和箱体热辐射的冷藏运输箱数值模型,获得箱内温度分布和变化情况.结果 随着箱内进风口风速的增大,荔枝的降温幅度较大,温度变化最显著的位置在每筐堆叠间隙处.当风机风速大于 8 m/s时,箱内降温速率变化不显著.在荔枝预冷温度低于 10℃时,可以提高箱内温度均匀性.在运输 2 h内,载质量越大,运输箱内荔枝的平均温度降幅低于 1℃,但温度均匀性变好.结论 在保鲜运输过程中,风速对运输箱内温度变化有着较大影响,增加货物堆叠时的间隙有利于提高传热效率.提前预冷货物,箱内温度的均匀性变好,增加载质量会增大箱内的热负荷,但是对箱内整体温度变化的影响较小.

为提高肥料利用率,降低肥料对水田的污染,该研究结合侧深施肥技术与液肥优点,研制一种水田滑刀开沟-气力引射式液肥雾化侧深施肥装置.该装置采用滑刀式开沟器开沟,利用气力引射式雾化施肥器雾化和引射液肥,将液肥侧深施于水稻根区附近土壤.设计了气液同轴气力引射式雾化施肥器内腔结构,以喉嘴距、混合室(喉部)直径、气体压力为因素进行全因子土槽试验,分析各因素对排肥量(液肥质量流率)和耗气量(气体流量)的影响.结果表明,影响液肥质量流率的主次因素顺序为混合室(喉部)直径、气体压力、喉嘴距;影响气体流量的主次因素顺序为气体压力、喉嘴距、混合室(喉部)直径.采用EDEM离散元仿真软件进行仿真优化,利用加权评分法综合评判仿真试验结果,结果表明,在不同工作速度下,滑切角为32.5°、刃口角为45°时,滑刀式开沟器可获得较优的工作性能.开展土槽试验验证仿真结果,滑刀式开沟器入土深度为30mm、前进速度为1.2m/s时,牵引阻力实测值为8.5 N,仿真结果为6.9 N,相对误差为18％,土壤扰动面积仿真结果为1 965.6 cm2;入土深度为50 mm、前进速度为0.6m/s时,牵引阻力实测值为14.4N,仿真结果为12.2N,相对误差为15％,土壤扰动面积仿真结果为2 137.2 cm2.土槽性能试验结果表明,该装置在入土深度为30mm,前进速度为1.2m/s时,排肥量标准差为0.2427 g/s,与最大排肥量的相对误差为1.42％,施肥深度与入土深度的相对误差为4.4％;在入土深度为50 mm,前进速度为0.6 m/s时,排肥量标准差为0.479 6 g/s,与最大排肥量的相对误差为2.13％,施肥深度与入土深度的相对误差为2.1％.研究结果可为水田液肥侧深施技术的应用提供参考.

[目的]设计一种机电式流量调节阀,与已研制的气力引射式施肥器集成构建液体肥变量施用调节系统,实现水稻近根部微小流量液体肥精准施用.[方法]通过试验标定了系统质量流率理论模型,建立控制系统传递函数模型,设计了基于模糊推理的PID控制器结构、规则和初始参数;通过仿真试验,分析了PID和模糊PID控制的调控响应能力.[结果]仿真试验结果表明,模糊PID控制阶跃信号响应超调量、调节时间和稳态误差分别为0.12％、2.51 s和 0.007,与PID控制的对应值 42.90％、4.44 s和 0.010 相比均较低,表明模糊PID控制动态调节和稳定性更好;在幅值为 0.5、持续时间为 0.1 s的脉冲信号干扰下,模糊PID控制的调节时间为 0.61 s,比PID控制(1.67 s)更短,具有更强的抗干扰能力.性能试验结果表明,10 种目标质量流率条件下,模糊PID控制的质量流率绝对误差均低于PID控制,控制精度为 93.93％～96.88％,高于PID控制(90.00％～95.21％);在施肥量变化时,模糊PID控制的超调量为 12.2％,上升时间、调节时间和峰值时间分别为 1.5、10.7 和 1.7 s,均低于PID控制的 17.4％、2.1 s、13.3 s和 2.3 s.[结论]基于模糊PID控制的水稻液体肥变量施用调节系统具有较高的质量流率控制精度和跟踪性能,为研制水稻田液体肥变量施肥装备奠定了基础.

Automated sowing performance monitoring for pneumatic roller-type vegetable plug-tray seeders helps reduce labor and ensure high sowing quality. In this study, a row of optical fiber sensors was utilized, and an optoelectronic measurement system was constructed based on the row-by-row sowing principle of the seeder. Performance of the measurement system was evaluated by sowing ten types of vegetable seeds. The results demonstrated there were no significant differences in measurement accuracy at different sowing speeds. Further, the seed properties did not influence the missed-seeding consistency ratio (CRms). The average CRms of the developed system was 99.81%. Seed shape was an important factor that influenced the single-seeding (CRss) and multiple-seeding (CRmu) consistency ratios. Seeds with greater than 85% sphericity obtained an average CRss and CRmu of greater than 99%, which demonstrated that the developed measurement system could be applied to vegetable plug-tray sowing.

准确辨识水培芥蓝花蕾特征是区分其成熟度,实现及时采收的关键.该研究针对自然环境下不同品种与成熟度的水培芥蓝花蕾外形与尺度差异大、花蕾颜色与茎叶相近等问题,提出一种注意力与多尺度特征融合的Faster R-CNN水培芥蓝花蕾分类检测模型.采用InceptionV3的前37层作为基础特征提取网络,在其ReductionA、InceptionA和InceptionB模块后分别嵌入SENet模块,将基础特征提取网络的第2组至第4组卷积特征图通过FPN特征金字塔网络层分别进行叠加后作为特征图输出,依据花蕾目标框尺寸统计结果在各FPN特征图上设计不同锚点尺寸.对绿宝芥蓝、香港白花芥蓝及两个品种的混合数据集测试的平均精度均值mAP最高为96.5％,最低为95.9％,表明模型能实现不同品种水培芥蓝高准确率检测.消融试验结果表明,基础特征提取网络引入SENet或FPN模块对不同成熟度花蕾的检测准确率均有提升作用,同时融合SENet模块和FPN模块对未成熟花蕾检测的平均准确率AP为92.3％,对成熟花蕾检测的AP为98.2％,对过成熟花蕾检测的AP为97.9％,不同成熟度花蕾检测的平均准确率均值mAP为96.1％,表明模型设计合理,能充分发挥各模块的优势.相比VGG16、ResNet50、ResNet101和InceptionV3网络,模型对不同成熟度花蕾检测的mAP分别提高了10.8％、8.3％、6.9％和12.7％,检测性能具有较大提升.在召回率为80％时,模型对不同成熟度水培芥蓝花蕾检测的准确率均能保持在90％以上,具有较高的鲁棒性.该研究结果可为确定水培芥蓝采收期提供依据.

An ordinary fruit paper bag supplying device used for fruit bagging has been designed and preliminarily tested in our previous study. How to open the ordinary multilayer fruit paper bag fully without damage is a concerning problem for the supplying device. The fruit paper bag is extended by the open mechanism of the supplying device in the opening process. Research of contact action effects of rigid open mechanism on the flexible ordinary multilayer fruit paper bag is necessary and has not been reported to date. Based on kinematic analysis of the open mechanism, the slider stoke and slider speed were determined as important influence factors for the bag open effects. Mechanical characteristic parameters of a typical ordinary double-layer fruit paper bag were measured. A rigid-flexible coupled dynamic simulation model consist of rigid open mechanism and flexible double-layer fruit paper bag was constructed. Both simulation and experimental results indicated slider stroke of the open mechanism was an extremely important influence factor for equivalent stress distribution, overall deformation and opened size of the fruit paper bag. Larger slider stroke would cause high rate of broken bag. With increase of slider speed, there was no big difference of the maximum equivalent stress, deformation and opened size of the fruit paper bag. Larger slider speed was positive for bag opening efficiency. Simulation and experiment methods used in the study is helpful for improving application effects of the developed supplying device.

Taking out and opening the ordinary multilayer fruit paper bag for fruit bagging is labor intensive, costly, not efficient, and potentially dangerous to the operator's health. There is a high demand to develop a mechanical device for the operation in Chinese orchards. A novel supplying device based on the manual operated fruit bag case was proposed. The open hand of the supplying device operates like a farmer's hand that can continuously take out a multilayer fruit paper bag one by one, and open it fully from its inside. Mechanism configuration and dimension parameter of the open hand were designed based on preliminary tests. The operation functionality of the supplying device prototype at different driving trajectories and speed was investigated in the study. The laboratory experimental data indicated that driving trajectory was an extremely significant factor for efficiently taking out and opening the fruit bag without sliding off and damage. Driving speed had a beneficial effect on reducing supplying time. With the synchronous driving trajectory and allowable high moving speed, the developed supplying device could achieve more than 90% opening success rate and less than 2-second opening time. The study showed the potential of the developed mechanical supplying device for fruit bagging with ordinary multilayer fruit paper bags.

开展《机器3D测绘》实践训练有助于巩固零部件测量和绘制技能,提高学生分析与设计表达能力,增强学生工程应用能力.引入逆向工程测量技术,提出测绘实践训练目标、实施流程和具体要求.明确测绘对象选取原则,选定测量工具,制定零件三维特征建模和逆向重构方法.提出组长负责制的训练计划,组间互评、小组自评与教师评定结合的综合评价方法,制定详细评分标准和策略.应用实践表明,该训练方法提升了学生综合运用现代设计技术方法的能力,增强了工程素质,提高了就业能力.

[Objective] For stably picking papaya without damage.[Method] A clamping plan of papaya was designed by means of three-finger clamping symmetrically and wringing, and the equilibrium equation of contact force was constructed.On the basis of force screw theory, the clamping stability was analyzed.The picking contact mechanics model was formulated, and a picking test for papaya was conducted.[Result] There were no obvious deformation, crack or indentation on the surfaces of papaya samples.The pulp at the clamping position had no obvious color change or bruise after leaving for 24 hours at room temperature.The maximum clamping force was far less than the pressure limit on the transverse diameter of ripe papaya at the elastic deformation stage.Papaya mass and torque moment were correlated closely with transverse diameter, vertical diameter, stalk length and diameter of the twisted stalk.Mass multiple linear regression analysis achieved extremely significant level, and twist torque linear regression analysis reached significant level.There were good trend consistency between theoretical clamping forces and the measured clamping forces.The measured clamping forces were higher than theoretical values, but the maximum deviation was less than 20%.[Conclusion] The picking scheme can stably clamp papaya without mechanical damage.The papaya picking contact mechanics model is correct and practical.The research can provide a basis for designing papaya picking end-effector and controlling clamping force.

Setting up comprehensive practice course is an important direction for practice course reform.In a comprehensive practice course,students are required to carry out innovative experiments or designs by using knowledge from several courses.Guiding the students to autonomic learning and self-managing is an effective method for conducting comprehensive practice courses.In the paper,carbon-free small car was used as the subject for a comprehensive practice course as a pilot reform.A managing organization consisted of students was founded to organize the course and a questionnaire survey was conducted after the course.The results indicate that the course improved the students' studying interests,theoretical knowledge and practical ability.As a consequence,the students show high degree of satisfaction to the course.

Carrying out the mechanical CAD/CAE/CAM integrated training can deepen students’ understanding and mastery of integrated design and manufacturing technology based on the unified product data information ,improve their abilities of solving problems in product development process ,and enhance their employability and competitiveness .Based on the actual enterprise product development process ,the integrated training objectives and implementation processes are brought forward , and the specific content and requirements for various aspects of development process are made .Solutions for selecting product development objects ,choosing development software and properly utilizing the short training time are put forward .Based on the condition of engineering training center of our university ,the integrated training plan is put forward . Through the assessment method of showing product development achievements in class and answering inquiries of teachers and classmates , establishing the scoring criteria ,the grade evaluation of the integrated training can be obtained .

Objective] To optimize chamber structure parameters of the pneumatic plate-type vegetable seed metering device , and to simplify the vacuum flow field in the air chamber as steady , regular, incom-pressible and turbulent fluid .[Method] The FLOTRAN module of ANSYS software was applied to simu-late and analyze the air chamber with different structure parameters .[Result and conclusion] Compared to the rectangular cavity air chamber structure , the rectangular groove connected air chamber structure could save more gas flow and improve the overall strength .Pressure loss of double outlet was less than single outlet, and flow field uniformity of double outlet was better .The optimized air source location was between the fourth and the fifth suction holes calculated from both sides of the metering device .Increas-ing channel depth could reduce pressure loss at transition region and ensure more uniform pressure and velocity distribution .But if channel depth increased , the size of metering device would increase corre-spondingly , and the time for forming steady flow would be longer .According to simulation results , 4 mm channel depth was optimized .Verifying tests of the air flow field in the vacuum chamber of seed metering device show that the simulation analysis results are comparatively consistent with the actual measurement results, with a consistent pressure distribution trend .This proves the feasibility of the numerical simula-tion method.

By analyzing comprehensive evaluation problem characteristics for fruit and vegetable picking mechanism configuration scheme, the grey relational analysis method was selected.Fruit and vegetable picking mechanism configuration scheme evaluation factor set was formulated. Scheme evaluation level structure was built. Evaluation factor weight sorting model was established based on fuzzy consistent matrix. Through comparison of geometrical similarity degree between picking mechanism configuration scheme evaluation column set and design requirement evaluation reference sequence,the grey correlation evaluation model was constructed. The application research on pineapple picking mechanism configuration scheme shows that the optimal scheme result is reasonable and the comprehensive evaluation model is scientific and practical.

为保证果蔬采摘机构设计质量,针对机构构型方案信息的模糊性和不确定性,选取模糊综合评价法进行比较优选.分析果蔬采摘机构工作性能指标的关联性和重要性,确定构型方案评价因素指标,构建方案评价层次结构.建立模糊综合评价数学模型,确定模糊量化方法和评价隶属度矩阵.选取层次分析方法,确定评价因素权重集.对菠萝采摘机构的评价优选表明,RRP串联手臂机构和RR并联手腕机构组成的构型方案综合评价分值最高,3RPS并联手臂机构和RR串联手腕机构组成构型方案评价分值最低.评价结果与理论分析结果一致,表明模糊综合评价模型具有合理性与实用性.