Font Size: a A A

Design And Experiment Of Greenhouse Seedling Picking Mechanism Based On Machine Vision

Posted on:2022-09-13Degree:MasterType:Thesis
Country:ChinaCandidate:H XuFull Text:PDF
GTID:2493306317453354Subject:Agricultural mechanization project
Abstract/Summary:PDF Full Text Request
In recent years,facility equipment and mechanized production have become the development trend of greenhouse facility horticulture.Currently,most of the transplanters for patching and transplanting seedling are semi-automatic.The trend of its research is to solve the problem of transplanting seedlings to soil,while studies on patching and transplanting seedlings on the plug are few.Patching and transplanting seedlings always rely on manual work with high labor intensity and low labor efficiency.There was a short period of time for patching and transplanting seedlings,so it may be easy to miss the best stages to decrease the economic benefits.Based on the above phenomenon,a patching and transplanting equipment for pepper seedling in greenhouse with image acquisition and processing,cavity tray transport,seedling patching and transplanting.This equipment can identify the quality of the seedlings on the plug and obtain the corresponding position information through the image detection system,and transmit it to the PLC controller,and give instructions to the executive elements,then the transplanting parts can be controlled to eliminate removed the bad seedlings and transplant healthy seedlings.The research covers the following areas:(1)Based on the size of the seedling structure of the cavity tray,we designed a a vision system to meet the operational requirements.By comparing common image processing methods,we selected the YOLOv3 deep learning model to process images of pepper seedlings relying on leaf size,number and stem characteristics,then it can identify their superiority and inferiority.The YOLOv3 model was analyzed by applying the loss factor Loss,the merging ratio IOU and MAP,and the ELU function was used as the activation function to optimize the image processing effect.The image processing performance was optimized.The results showed that the accuracy of unhealthy and healthy pepper seedlings was 0.981 and 0.95 respectively,the sample size and time consumption were reduced by half,which optimized the image processing performance.(2)The greenhouse pepper seedling transplanting equipment was designed based on transplanting demand and patching and transplanting operation principle,which was consist of the patching component,the lighting chamber,the initial conveying component,the classified conveying component,the transportation of transplanting seedlings component and the transportation of supplying seedling component.In order to achieve the expected function of transplanting equipment,the key structure design and the selection calculation of related components were carried out for its components.And then the feed precision of the transportation of transplanting seedlings component and the transportation of supplying seedling component was tested,the results showed that the feed precision was 3.0mm,which met the precision requirements of plug conveying.(3)Based on the function of pepper plug patching and transplanting seedlings,the transplanting circuit system and gas path system were designed.In order to coordinate the operation of various components stably,the selection of relevant components was calculated.Then,a control idea of plugging seedling test,plugging transplanting and stable patching and transplanting operation was designed.In order to improve the efficiency of transplanting equipment,the transplanting path of the transplanter was planned and analyzed,and a group of better transplanting paths are selected.The efficiency of transplanting equipment can reach 1300 plants/hour.(4)In order to improve patching and transplanting performance,seedling depth,horizontal running speed and matrix moisture content were studied by orthogonal test.The analysis of variance was used to obtain a set of optimal parameters for transplanting equipment.The results showed that the depth of seedling was 38mm,the horizontal running speed was 0.15m/s,the moisture content of substrate was 60.5%,the success rate of seedling picking by transplanting equipment was 100%,the success rate of seedling replacement was 95.8%,the success rate of transplanting was 95.6%,and the transplanting efficiency was 1300 plants/h.
Keywords/Search Tags:greenhouse transplanting, deep learning, pepper seedlings, weeding seedlings, path planning
PDF Full Text Request
Related items