Research On Exceed Field Workpieces Recognition Location And Grabbing System Based On Machine Vision

In the current workpiece visual inspection applications of machine vision system,for the exceed field workpiece which volume is lager than the single camera field of view,due to can not shoot the complete picture of the workpiece,which leads to the difficulties of identification and positioning.In view of these problems,this paper proposes and study a exceed field workpieces recognition locating and grabbing system(WRLGS)based on machine vision.The main contents of the research include the following aspects:(1)Research on recognition of exceed field workpieces based on machine vision.A method to recognition based on extract the feature points and the contour of the ROI intercepted from the selected centre feature points is used as a template to match the shapes is designed by comprehensive theory and practice.This method can recognition the exceed field workpieces validly.(2)The method of locating the exceed field workpieces based on machine vision is studied and discussed.A method point at the characteristics of exceed filed workpieces does not require complex calibration process have been designed The method is to define the standard position of each kind of work piece,and to locate the work piece by calculating the relative standard position offset under each coordinate axis..(3)The grasping of exceed field workpieces based on machine vision have been studied.The whole mobile area of the robot is divided into three blocks.The system chang the traditional way which use single layer control network,Through the design of coarse to fine grabbing network The whole process of grasping is divided into the process of combining the rough adjustment of fuzzy control and the fine adjustment of fine calculation.In order to adapt to the modern production demand of multi variety,flexibility and intelligence,the system can adapt itself to a variety of workpiece applications by self-learning,Realize in the case of no manual operation,self-learning the recognition and grasping process of various kinds of workpieces.(4)The system model is constructed and the system example is built successfully.The system software platform and hardware examples are used to verify the design of the system.Experiments show that grabbing time of the system is less than 6 s,and the errors between the grab position and the target in horizontal direction,height and angle are no more than 2mm,0.2mm and 0.3 degree respectively.The feasibility,accuracy,flexibility and stability of the system are proved.It is hoped that a new reference pattern and an implementation approach can be obtained for the identification and positioning of exceed field workpieces under unmanned operation.It is hoped that a new reference pattern and an implementation approach can be obtained for theWRLGS under unmanned operation.