Design And Implementation Of Baxter Robot Sorting System Based On Machine Vision

In industrial production,traditional manual sorting has many disadvantages,such as high labor intensity,low efficiency and so on.It is difficult to achieve the production goal.The robot teaching and sorting system has the disadvantages of poor stability and low accuracy.In view of these problems,adding machine vision technology to the traditional sorting system can make the sorting system more intelligent and flexible,improve production efficiency and reduce labor costs,which has important practical significance for the realization of robot intelligent sorting.In order to restore the industrial field environment and the real sorting process to the greatest extent,this paper proposes the Baxter robot intelligent sorting system based on machine vision based on the research object of several common workpieces and Baxter robot platform.This paper has carried out the corresponding research work from the following aspects:First,build an intelligent sorting system.Build an intelligent sorting platform based on Baxter robot hardware and supported by ROS software,study the classification,positioning,grabbing,placement and other related issues of the intelligent sorting system,and use the built sorting system for experimental verification.Secondly,the workpiece is classified and identified.Each kind of workpiece is collected as a training set.The surf algorithm is used to extract the feature points of all training pictures of each kind of workpiece,and K-means is used to cluster the feature points to make a visual word bag,which is represented by a word bag model for each training picture.For the small sample of the work piece in this paper,we choose the word bag model and the label of each class of work piece training picture to get the SVM model.After the target work piece to be detected is represented by the word bag model,we use the trained SVM to identify the work piece.Finally,the material basket and the target workpiece are located.In order to get the outline of the material basket and the center point of each sub material basket,Canny edge detection and target binarization are used.The image preprocessingmethods such as grayscale processing,Gauss filtering and target binarization are used to eliminate noise interference.The position information such as the center of mass of the workpiece is obtained by using the minimum external moment.Aiming at the error problem in the positioning process of the material basket and the target workpiece,the precise positioning method of hand eye is adopted to improve the positioning accuracy.The experimental results show that the robustness of the intelligent sorting system designed in this paper can meet the requirements of the actual workpiece sorting,and can realize the real-time detection of the sorting situation through the system platform.