Research On Mobile Phone Shell Feeding System Based On Vision Guidance

Vision guidance technology endows the robot with visual perception capabilities and improves the robot's autonomy and flexibility.Industrial robots based on vision guidance are widely used in industrial production sites because of their high accuracy,fast speed,good stability,strong adaptability,and uninterrupted work.The development of robots with perception and recognition capabilities is an important support for the realization of automated factories and a key force for the transformation and upgrading of China's manufacturing industry.At present,the electronic manufacturing industry relies heavily on manual labor,which has the problems of high cost,low speed,and low yield rate.Therefore,this article takes the precision electronic manufacturing production line as the experimental basis,and takes the metal mobile phone case as the research object.Aiming at the demand of automatic feeding of the production line,research and design a set of vision-guided robot system with the functions of detection,identification and grasping and feeding.The main research contents are as follows:First,according to the functional requirements and performance index requirements of the precision electronic production line,the overall design and hardware selection were carried out,and the vision-guided robot experimental platform was built.A dual-light source large-angle lighting solution for the 2.5D arc surface of the mobile phone case is proposed to solve the problem of the lack of the2.5D arc surface of the metal case and the reflection of the logo area,and obtain the image of the complete metal case.And compared the images obtained by other lighting schemes under the same conditions.Secondly,the method of noise reduction and edge extraction of mobile phone shell images is studied.The processing results of the four noise reduction methods,median filtering,Gaussian filtering,mean filtering,and bilateral filtering under dark light,normal light,and strong light conditions,are compared.Through the analysis of the results,this paper finally selects the one that can effectively retain the edge information of the image.Bilateral filtering is used as a noise reduction method;research on Otsu threshold segmentation,Canny edge extraction,Sobel edge extraction algorithm and morphological segmentation methods,and compare it to find that the morphological segmentation method can better ensure the continuity of the edge.Then study the detection and recognition method of mobile phone shell,and propose a recognition method that combines shape features and improved ORB matching algorithm to realize the recognition of mobile phone shell and greatly improve the recognition speed and accuracy.It includes two parts: coarse matching and fine matching.In the coarse matching stage,the invariant characteristics of the rotation and translation moment and the fast calculation speed of the HU invariant moment algorithm are used to quickly realize the pre-screening of the image and reduce the amount of calculation;in the fine matching stage,use The improved ORB matching algorithm improves the scale invariance of the algorithm and improves the matching accuracy.Third,study the camera calibration and robot hand-eye calibration of the Eye-to-Hand vision system,study the visual positioning method for the mobile phone case,obtain the center position of the mobile phone case and the rotation angle of the center axis of the mobile phone case,and calculate the grasping point and the mechanical claw angle.The center of the smallest circumscribed rectangle of the target shape is taken as the center position of the mobile phone case.Taking the longest side of the circumscribed rectangle as the reference for calculating the rotation angle of the mobile phone case,the slope of the straight line in the pixel coordinate system is calculated as the inclination angle of the mobile phone case.By converting the position of the mobile phone case in pixel coordinates to the position in the world coordinate system,the robot arm is guided to reach the predetermined grasping position,and the robot arm is rotated according to the inclination angle to complete the grasping work.Finally,a set of software control system is designed,and the method proposed in this paper is verified on the experimental platform.The experimental results show that the algorithm proposed in this paper has obvious advantages in recognition accuracy,recognition rate,and grasping accuracy.Verification from the point of view of the problem meets the design requirements.