Algorithm And Application Of Machine Vision Space Calibration And Target Locating For Automatic Packaging

With the development of modern intelligent technology and automatic control industrial robots are applied to improve the production efficiency.However,industrial robots cannot achieve locating,tracking and inspection for target objects.Therefore,cooperative external sensors are required.Machine vision is one of the most common approach that perform locating,tracking and inspection with industrial robots.It is widely used in kinds of industries due to its high precision,functionality and flexibility.Based on the technologies of machine vision and industrial robot,an automatic packaging systems of packaging paper is studied in this paper.Most of the industrial robots in the traditional automatic production line use off-line instruction to make the robot complete the action.The industrial robots move according to the preset trajectory to achieve placement operation.However,the off-line teaching method must be placed in strict accordance with the requirements of the position and Angle of the work piece,so its poor adaptability,can only be used in some occasions can not meet the expectations of intelligent industrial production.Therefore,in the process of identification and positioning of the work piece,machine vision technology is added to make use of visual perception of changes in the external environment,so that industrial robots can quickly and accurately obtain the position information of the work piece through visual perception of the external environment,and then use industrial robots to grab the work piece.Firstly,this article introduces the background and the state of art of machine vision and industrial robots,and then a linear camera model calibration method to establish the relationship of coordinate system transformation between the robot and the camera.Secondly,the images acquired by the camera are preprocessed,by binarization processing,image filtering,edge extraction and corner point extraction.After analyzing the edge detection operators,it is concluded that the edge features obtained by the Canny operator are the best,which lays a foundation for subsequent corner point detection.Finally,this paper proposes to use Hough line transform to detect corner points.This highly robust algorithm calculates center points of wrapping paper,achieves positions of wrapping paper in the robot coordinate system through coordinate transformation,and guides the industrial robots to perform the grabbing and placing of packaging boxes.According to the experimental results,the algorithm feasibility is proved.Therefore,the goal that industrial robots accurately grasp and place work pieces is achieved,which greatly improves the level of flexibility and intelligence of industrial robots.