Research And Design Of Reseeding Machine About Sweet Corn Seeding Tray Based On Machine Vision

Sweet corn is an important economic crop.China is a big country of sweet corn consumption and planting.During the large-scale hole tray seedling raising in the factory,in the sowing link of seedling raising,due to the different shapes and sizes and uneven shapes of sweet corn seeds,the phenomenon of missing sowing holes will occur when the roller seed metering device is used to sow the hole tray.In the face of these missing sowing holes,additional manual identification and manual reseeding are required.Based on the machine vision technology,this thesis studies and designs a set of hole disc reseeding machine for sweet corn,which is used to liberate people’s eyes and hands and improve the reseeding efficiency.The main research and design contents are as follows:1.The background of sweet corn cave tray seedlings is introduced,the problem of low efficiency of artificial replanting is proposed,the research purpose and significance of the topic are expounded,and the research status of machine vision technology in agricultural applications is introduced;the research status of the replanting system at home and abroad is introduced;the research status of the replanting system is introduced,and the research content of this topic is finally determined.2.The hole image recognition system is designed.The image recognition environment is designed and built,and the hole disk image is processed by gray threshold,median filtering and morphological processing in RGB and HSV color spaces respectively,and then the seed contour is extracted.Finally,the seed area extracted in the two spaces is weighted to determine the final position of the seed,so as to obtain the position of the hole.3.Research and design a two-level cache reseeding system.The structure and control method of the first level cache unit in the array seed bin are studied and designed.The unit can quickly reseed multiple holes at the same time;The second level cache unit is designed,which can refresh the seeds for the first level cache unit and ensure that the first level cache unit has enough seeds to reseed holes.The path planning and speed planning of seed fetcher in L2 cache unit are studied and designed;Based on Modbus protocol,reliable interaction of hole data between reseeding system and machine vision system is realized;The reseeding strategy is studied based on the idea of page replacement algorithm;The opening time of seed bin door is planned to ensure orderly reseeding.4.Test the performance of the system,test the accuracy of sweet corn hole image recognition,test the accuracy and rapidity of the two-level cache reseeding system,and test the accuracy and rapidity of the overall reseeding system.Image experiments show that the average accuracy of hole image recognition is98.83%.The test of the whole machine showed that when 7 holes were replanted,the optimal speed range of the seed tray was 3～3.5 m/min.the system could replant 283～331plates per hour,and the precision range of replanting was 91～91.4%;When reseeding 15 holes,the best speed range of the hole disk is 1～2 m/min.the system can reseed 95～192disks per hour,and the precision range of reseeding is 90.7～92.2%;When 26 holes are replanted,the best speed range of the disc is 0.5～1 m/min.the system can replant 49～95discs per hour,and the precision range of replanting is 92.1～92.6%.The project design has achieved the expected goal,and the machine has certain rapidity and accuracy.