Design And Implementation Of Intelligent Grabbing System Based On Machine Vision

At present,robots and automation technology are constantly developing and have been widely used in the industrial field,which has improved production speed and efficiency,and at the same time liberated heavy manual labor.In the field of automated production,up-down material is one of the important links,and some enterprises such as bearings,valves,and magnetic materials companies still use manual or semi-automatic grasp methods for up-down material.There are the following problems in this way:(1)Workers' workload is heavy and boring.(2)Parts are easy to bump and damage.(3)Increasing production costs of enterprises.In order to achieve intelligent grasp,this subject has designed a intelligent grabbing system based on machine vision.The system uses a single industrial camera to obtain a rectangular part image,and uses the obtained image to identify and reconstruct the object,and then realize Intelligent grasp of the rectangular parts.Summarize the main work,including the following points:1.The design and construction of the intelligent grabbing system platform.This subject designs the three-dimensional model of the three-coordinate manipulator platform of the intelligent gripping system and complete the platform construction.And this subject also designs control,vision and human-computer interaction systems,and complete the construction.2.Identification and positioning of rectangular parts.This paper takes synthetic images and actual rectangular parts as research objects,and proposes a fast and highprecision rectangle detection algorithm.The algorithm completes rectangle detection and positioning by dividing the image,extracting the sub-pixel contours of the image,dividing the sub-pixel contours into straight lines,and the fuzzy fusion of straight line geometry and physical features.Experimental results show that the detection speed,accuracy of the algorithm meet the industrial requirements and have good robustness.However,this algorithm can only obtain the two-dimensional information of the parts.In the process of grabbing multiple parts palletizing,it is difficult to achieve true intelligence because it is impossible to obtain the threedimensional structure information of the parts and the grasping depth of the mechanical platform in the Z axis direction.In response to this problem,this article,then,proposes a three-dimensional reconstruction of a single image of the grab based on the palletizing task.3.Three-dimensional reconstruction of a single image of the captured object.Taking the rectangular parallelepiped parts in the palletizing process as the research object,this paper proposes a three-dimensional reconstruction method of single image grabbing based on the palletizing task.This method first extracts the part contour,and then uses the Hough line detection algorithm to extract the target straight line in the contour,groups the straight lines,optimizes the parallel straight line groups in the vertical,horizontal,and vertical directions,and then uses the parallel straight line groups and the minimum distance method to calculate the vanishing point.And this paper uses the vanishing point to obtain the internal and external parameters of the camera,and according to the internal and external parameters this paper finally obtains the three-dimensional information of the image and the grasping depth of the manipulator in the Z axis direction.Experimental results show that this method can accurately reconstruct a three-dimensional model that meets the accuracy requirements,and can accurately obtain the grip depth of the manipulator in the Z-axis direction.4.The 2D and 3D information of the parts have been obtained by 2,3.In order to make the whole intelligent grasping system platform more complete,it is necessary to integrate the acquired part information into the platform's humancomputer interaction system,And it finally realizes the grasping of rectangular parts in different scenes.Finally,the machine vision grabbing system platform is debugged,and the system's intelligent grasping ability,accuracy and robustness are tested.The experimental results show that the intelligent grasping system can achieve efficient and autonomous grasping of rectangular parts,and which is meeting certainrequirements.The grabbing accuracy is good robust.