基于茎秆识别的辣椒穴盘苗健康状况判定与试验OA

In greenhouse seedling cultivation,to ensure that the seedlings in the trays to be transplanted are all healthy and that there are no missing seedlings,it is necessary to assess the health status of the seedlings and carry out the opera-tion of removing and supplementing seedlings.For this,focused on the parallel robot developed by the research team for seedling removal and supplementation in greenhouses,designing and equipping it with a vision system that employed an image processing method based on identifying the stems of pepper seedlings to achieve precise discrimination of tray seed-lings.Firstly,the captured images were cropped using the ndarray slicing principle to extract the target Region of Interest(ROI).Subsequently,a novel method based on the area and perimeter of the stem to determine the health status of pep-per seedlings was proposed.Specifically,the processing workflow included image grayscale conversion,filtering and de-noising,edge extraction and segmentation,as well as feature parameter extraction and statistical analysis.By calculating the stem feature parameters from multiple tray samples,a threshold range for healthy pepper seedlings within a fixed peri-od was obtained,which defined the health status of the seedlings,thereby enabling precise identification and localization of missing and diseased seedlings.Finally,a prototype of the parallel robot for seedling removal and supplementation in greenhouses was built,and experiments on image recognition of missing and weak seedlings were conducted.The experi-mental results indicated that among the recognition results of seedlings in a 128 cell tray,only two were misidentified,re-sulting in an overall recognition accuracy exceeding 98%,thus validating the feasibility and effectiveness of the method for determining the health status of pepper seedlings based on stem features.