Application And Research Of Cordyceps Automatic Grabing Technology Based On Machine Vision

Cordyceps sinensis is a valuable traditional Chinese medicine.The amount of natural growth is very small.With the improvement of economy and living standards,the demand for Cordyceps increases year by year.At present,most of the production of Cordyceps depends on artificial breeding,but the technology of artificial breeding is not mature enough,and the breeding mode is still unscientific,the survival rate is low,thus the overall output is far from enough to meet people’s needs.The Cordyceps automatic breeding system Based on machine vision integrates functions such as automatic grabbing,automatic transfer,measurement and screening.The use of machines instead of labor can save labor costs,and large-scale automated farming improves farming efficiency and yield.This article will study the automatic grasping function of the Cordyceps automatic breeding system.The main problem is how to realize the robot automatically and accurately grasp the insects and put them into the culture dish.In order to solve this problem,the main work of this paper is as follows :(1)In view of the incomplete problem of bug and background segmentation,this paper compares the global threshold segmentation and the MSER algorithm.The experimental data of multiple groups show that the MSER has more stable and more complete segmentation in the case of background reflection and impurity interference.(2)For the traditional calibration method with calibration plate has the problems of complicated calculation,complicated operation and high environmental requirements.Based on the analysis of the principle,advantages and disadvantages of the existing calibration algorithm,this paper decides to affine a calibration algorithm based on the affine transformation.The method is proved to be convenient and fast,and the calibration accuracy can meet the identification requirements of the system.(3)Because of the existence of skeleton line branching and difficulty in branch cutting in current skeleton line extraction algorithm,an improved algorithm for skeleton line extraction based on the largest disc center model is proposed in this paper.The experimental results show that the proposed algorithm effectively solves the problem of affecting the shape of the skeleton due to branch cutting,improves the accuracy to a certain extent,and is superior to the classical algorithm in execution efficiency.(4)Based on the completion of key technology research,a visual recognition system based on VC++6.0+Qt and Halcon library is designed,including image acquisition module and image processing module.The system software and hardware linkage was realized and the experimental platform was built.Many tests were carried out in the field.The optimization was debugged according to the results of each test.The final test results show that the whole system is stable and effective,and can accurately complete the bug capture.