| Potato is one of the important food crops,and the main cultivation mode is seed potato cutting.Artificial cutting seed production before sowing has high labor intensity,long cutting cycle and low operation efficiency.With the aggravation of population aging and the reduction of rural labor force,mechanized cutting has become an inevitable trend.However,the existing cutting machine has serious waste and low eye rate of potato tuber.In view of the problems faced by the above seed potato cutting,this study based on the agronomic requirements of potato cutting sowing,combined with image processing technology,designs a seed potato automatic cutting machine,which can achieve efficient and accurate cutting according to the distribution of seed potato buds.The main research contents and conclusions are as follows :(1)Investigate the agronomic requirements of seed potato cutting,determine the function and overall scheme of the machine,and use the measurement results of the geometric size of the seed potato as the basis for mechanical structure design and analysis.The feeding device,grading device,sorting device,image system and cutting device in the machine are designed in detail,and the working process of the machine is determined through the overall design.(2)Complete the structural design of the mechanical device based on the design scheme.The lift chain gap of the rod sleeve in the feeding device is 28 mm,and the lift angle is 20 °.The seed potato was divided into 5 grades by 13 grading rollers in the grading device.The diameter of the grading roller was 80 ～ 125 mm and the rotation speed was 90 ～ 150 r / min by taking the movement of the seed potato over the grading roller as the critical condition.The EDEM was used to do the discrete element single factor simulation test to determine the optimal value within the range with the grading accuracy as the standard.The whole sequence device structure can quickly gather and transport seed potatoes.The gantry threeaxis cutting device in the cutting device can accurately control the tool movement and avoid the potato “sticking knife”.The three types of tool structure design and finite element analysis by ANSYS show that the strength and total deformation meet the requirements.Solid Works is used to construct a three-dimensional model of the mechanical assembly.(3)The seed potato image system was selected and calibrated.The results showed that the image could accurately present the surface characteristics of the seed potato.Aiming at the problem of image interference caused by acquisition and transmission,a series of image preprocessing such as grayscale,filtering and noise reduction,and image segmentation are performed on the original image.Compared with the results of MATLAB simulation test,the average value method image graying,median filtering noise reduction,OSTU threshold segmentation,Canny operator edge detection seed potato image processing effect are remarkable.The image processing method of seed potato was integrated and designed by GUI.(4)Based on image processing,the extraction of seed potato centroid and the expansion of seed potato image data set are completed,and the motion path of the actuator is planned by tracking the centroid coordinate position of the tool center.The semantic segmentation method Deep Lab V3+ network model is applied to seed potato image processing,and the expanded seed potato image data set is simulated and extracted by python programming language.In the simplified model of cutting tool,the cutting mechanism of different cutting tools is determined by the number and position of seed potato bud eye,which can ensure that the potato block contains at least one bud eye. |
PDF Full Text Request