Research On Location Of Multiple Plug Trays And The Acquisition Of Pot Seedlings Health Information In Greenhouse Seedling Bed

With the development of intelligent equipment and modern agriculture,intelligent mechanical equipment has received extensive attention in agricultural application,the development of agricultural machinery is also one of the research hotspots.On the basis of the previous study on fixed transplanting equipment,put forward the idea of mobile inspection and transplanting robot to reduce manually transfer the plug trays to the transplanter,which could perform pot seedling inspection and transplanting seedlings directly on the seedbed.In order to realize the automatic detection of pot seedlings by mobile inspection and transplanting robot,research image mosaic algorithm to obtain the images of long-span seedbed.A concentric rectangular box template matching algorithm was developed to accurately locate multiple plug trays in the image of the seedbed.Providing technical support for subsequent pot seedling inspection and transplanting.The main research contents and results are as follows Firstly,the sub-region of seedbed image was obtained and distortion correction.According to the requirements of greenhouse space,seedbed span and image mosaic on the proportion of overlapping area,the view field of the camera was planned and the appropriate camera and lens model was selected to calculate the installation height and shooting position of the camera.In view of the distortion in the image,the checkerboard image and TOOLBOX calib toolbox in MATLAB were used to calculate the camera parameters and distortion coefficient,realized the calibration of the cameraSecondly,in order to obtain the image of long-span seedbed,the image sequence of the seedbed was spliced and fused by the image mosaic algorithm.Based on block area matching,Harris corner detection and SURF feature point detection,the image sequence of the seedling bed was splicing.It was found that the image mosaic algorithm based on Harris corner detection ensures the accuracy of feature point matching through non-maximum suppression and other methods,and has a good effect in image fusion of seedling bed and algorithm efficiency.Finally,a concentric rectangular template matching aigorithm was designed and developed based on the layout of the plug trays and the growth area of the pot seedlings on the basis of the image of the seedbed in the large field of view to realize the scanning and positioning of the multiple plug trays.With the number of white pixels in the area between the inner and outer rectangular boxes(Ip)as the main reference,took the position with the minimum Ip value as the optimal solution for locating the center point of the plug tray.By means of regional scanning,threshold comparison and optimal value screening,the location of multiple plug trays in the image was realized.By comparing the positioning effects of template matching method and projection method,it was found that the plug tray positioning algorithm based on concentric rectangle template can avoid the influence of pot seedlings crossing the boundary on the positioning accuracy.On the basis of the location of eachplug tray,the area of the plug tray was divided,and the health of pot seedlings was detected according to the leaf area in each hole,the success rate of pot seedlings detection reached 98.65%,which met the requirements of actual operation.