Design And Implementation Of Practical Teaching Assistant System For The Geared Motor Assembly

At present,many vocational schools in China have opened the course "motor assembly training",which allows students to assemble a reducer motor by hand,so that students can experience the complete steps of assembling a reducer motor.However,some problems were also found when carrying out this course.The common problem is that there are few teachers and many students.The teacher cannot teach each student to assemble by hand.Because the students are beginners,they may not know each part and the assembly position,resulting in the feeling that students will have no way to start.With the continuous development of computer technology,it has affected all aspects of our life,so it is also an inevitable trend to introduce computer technology into curriculum teaching.This paper uses computer vision technology to realize a practical teaching assistant system for reducer motor assembly,which greatly improves the teaching efficiency of teachers.The main research contents are as follows:(1)This paper studies and analyzes the problems existing in the practical teaching of reducer motor assembly,defines the requirements of this auxiliary teaching system,designs the overall architecture and technical route of the system according to these requirements,and designs two functional modules that need to be equipped on this basis: the identification and positioning of parts and the visual interface.The basic theories and related technologies involved in the design and implementation of this system are also briefly described.(2)The part recognition and positioning function of this system is realized by using the classic target recognition algorithm based on threedimensional point cloud.The tool to obtain the point cloud uses the lidar carried on the iPad pro,because it is smaller,lighter and cheaper than the general lidar,which is suitable for this application scenario.With the help of arkit on iPad pro,an application program for obtaining point cloud data-pointcloudscanner is developed.It can upload the scanned point cloud,image frame and the corresponding camera internal and external parameter matrix to the host through LAN communication technology,and then carry out classical point cloud recognition on the scene point cloud scanned by pointcloudscanner application on the host to obtain the three-dimensional bounding box coordinates of each part of the reducer motor.Finally,through the camera internal and external parameter matrix uploaded by the application,the three-dimensional bounding box is projected onto the corresponding image,and the assembly sequence and name of the parts are marked,and then it is presented on the interface.The classical recognition algorithm based on point cloud is used in this paper,and the average position error RMSE of all parts’ 3D bounding box is 0.2671.(3)The visual system interface is developed using pyqt5.Students can easily use the functions of the system,display the recognition results of twodimensional images and the three-dimensional assembly positions of various parts through the interface.The 3D assembly drawing of the part is made and displayed in open3D,and then the display window is embedded in the interface of the visualization system.Students can quickly assemble a reduction motor according to the two-dimensional image recognition results and the three-dimensional assembly step diagram displayed in the interface.This paper designs and implements a practical teaching assistant system for the assembly of reduction motors.The system realizes the recognition of parts,the marking of the assembly sequence of each part and the visualization of three-dimensional assembly steps,which helps to cultivate students’ interest in learning and improve teachers’ teaching efficiency.