Design Of Control System Of Colony Picking Robot Based On Machine Vision

With rapid growth of the number of genomic sequences,researchers are faced with the task of managing huge clone libraries.A single experiment may involve classifying tens of thousands of bacterial,yeast or phage colonies.Traditionally,co lony picking and inoculating is accomplished by hand.While,repetitive operations,which are followed by a lower efficiency and accuracy of colony picking and inoculating,will lead to mistakes of colony picking and have an impact on downstream experiments.There is no doubt that manual picking cannot meet the demand of high throughput in biological experiments,due to its inefficiency and low accuracy.Therefore,efficient colony picking machine comes into being.This paper introduces a colony picking control system based on machine vision.For vision system,that colony to pick have small diameters and distribute intensively has a restriction on efficiency of recognition and positioning of colony.Meanwhile,the stability of control system has an influence on picking speed and accuracy.Therefore,colony recognition and positioning method applied in vision system and design and development of control system are the key technology topics to solve.The study carried on above topics has a great meaning.The thesis work is mainly carried out from the following aspects:(1)Colony recognition and positioning procedure based on visual software is introduced.This procedure differentiates from general colony recognition and positioning procedure,which simplify recognition steps and use Locator modules and blob analyzer provided by visual software to exploit applications directly.(2)A modular method is adopted to design the control system framework.The function of modules and hardware implementation are introduced and hardware selection of CCD camera module is described in detail.In the final,control system achieves the control of actuator,picking and inoculating of colony and cleaning and heating of needles.(3)The control system is developed with modules and the application software interface is designed.Through analyzing tasks needed to fulfill,three difficulties encountered during development of control system and those source codes are put forward.Piking colony is achieved.By analyzing the specific requirements of the control system,the software interface of the application is designed.(4)Colony recognition and positioning method is proved to be feasible and software parameters during the recognition process are optimized through the test.And also,camera calibration method is proved to be feasible and visual positioning accuracy is calculated.Finally,the mechanical error matrix of the system was measured,and the feasibility of the software compensation method was verified.