Design And Implementation Of Robot Vision Control System Based On MFC

Science and technology are the primary productive forces.With the development of science and technology,more and more conveniences have been brought to human life.In the aging society is becoming more and more serious today,more and more intelligent robots to replace human,therefore,the robot has an indispensable importance for human and robot vision,occupy most of the robot's perception,and determines the degree of intelligent robot.In the end,we have realized a system that can control robot to move to the front and make robot catch air throwing objects by gesture.Based on the current domestic and foreign mainstream robot are discussed,the historic robot research,research on different application scenarios in the key technology of robot,robot at home and abroad are analyzed and discussed,and the importance of unfolds the digital image technology in robot vision and its application in a robot.Through the investigation of current market mainstream robot platform in mechanical design platform,using SolidWorks to design a new omni-directional wheel,relative to the Mecanum wheel has higher speed and accuracy.In the control system,combined with the open source Arduino platform for robot control part of the design,use the PID algorithm to the speed and number of laps of the motor accurately controlled by motor and photoelectric encoder,which makes the robot in the position control room is guaranteed.In throwing tracking using frame differential algorithm and KCF algorithm Kinect depth map frame based on the robot tracking,using the method of template matching to detect indoor robot in hand detection using Camshift algorithm to detect.The final world coordinates are converted to two-dimensional coordinates using OpenNI and OpenC V,and Matlab is used to fit the trajectory of the thrown object and calculate the pre landing coordinate.Finally,the robot control system is designed and implemented by combining MFC,and the interface is given with deeep learning as Caffe architecture to prepare for AI implementation.This paper summarizes t he research,and analyzes the inadequacies of the research,and follow-up research and development prospects.