| In recent years,with the greater complexity,scale and accuracy of the micromanipulation process,it is difficult for manual manipulation to meet the actual needs of micromanipulation,so people urgently need a set of automated micromanipulation system to reduce the difficulty of work.With the increasing improvement of vision-based target detection and positioning technology and visual servo control theory,microscopic visual servo system has become a new trend in the research of automated micromanipulation system.In the paper,by microcopper wire as the manipulation object and pipette as the manipulation tool,the microscopic visual servo system based on the manipulation process of electrochemical microcomponents is studied,and the main contents are as follows:In order to the problem of low automation of micromanipulation system,the micromanipulation process based on electrochemistry is analyzed,to realize a binocular microvision micromanipulation system and built a micromanipulation experimental platfrom.The hardware of micromanipulation system is select based on the performance indicators of system.To improve the interactivity of the actual manipulation process,Lab VIEW is used to design a PC system with functions such as visual acquisition,camera setting,intelligent recognition and positioning,motion control.In order to solve the problem of the target recognition and positioning under nonocclusion in micromanipulation system,a contour-based manipulation point positioning algorithm is proposed.Aiming at confusion between unfocused shadow and target caused by incomplete focus of microscopic vision target,the threshold selection of traditional OTSU threshold segmentation algorithm is modified.The shape matching algorithm based on contour features is used to identify pipettes and microcopper wires.The pipette nozzle profile polygon fitting algorithm is used to determine the coordinates of the two end points of the pipette nozzle.The microcopper wire contour external rectangle fitting algorithm is used to determine the position coordinates of the microcopper wire.The coordinates of pipette and microcopper wire manipulation points are determined according to the corresponding coordinate relationship selected by cameras with different visions.In order to solve the problem of manipulation tool tracking and positioning under occlusion in micromanipulation system,a contour-based particle filtering algorithm is proposed.Aiming at the pipette nozzle manipulation point cannot be directly position under occlusion,the tracking and predictive positioning method is selected to realize the positioning of the pipette manipulation point.Aiming at the classical particle filter algorithm is sensitive to lighting,target profile feature matching is used instead of histogram matching.A comparative experiment of tracking algorithm is designed to verify the advantages and disadvantages of contour-based particle filter algorithm in tracking accuracy and algorithm time.In order to improve the control accuracy of the micromanipulation system,an image-based binocular vision servo control scheme is adopted.Through the binocular microvision micromanipulation platform,the automatic picking experiment of microcopper wire is carried out to verify the effectiveness of the micromanipulation system based on visual servo.Through the positioning experiment of different manipulation points of main vision,the applicability of the proposed micromanipulation system positioning algorithm is verified. |
