Research And Development Of Visual System Based On Machine Vision Table Tennis Robot

Table tennis robot is made up of the visual system,mechanical system and control system with the combination of hand-eye system,of which the visual system is equivalent to the brain,eyes,through the trajectories of the visual system to forecast table tennis,calculate the shot point and time.the robot can return the ball effectively.Overseas has the method research of table tennis robot for many years,and of the existing research results can play multiple round with people,But in our country,only has carried on the related research of Zhe Jiang university^?4?,and designed a three generations of table tennis robot,and continues to study..According to the research,table tennis robot focuses on the visual system,calculates the 3d coordinates of the shot point and the time to hit the ball through the visual system,and controls the robot hand to finish the stroke.To complete the effective return of the machine hand,to obtain the accurate position and time,you will first need to obtain the accurate Fei HangDian coordinates and early forecast the trajectories of table tennis in the future and set aside enough time to make the machine hand strokes.This paper first introduces the status and development of table tennis robot and the research significance of table tennis manipulator.Then,the method of binocular calibration is introduced,and the relationship between the camera's internal reference and the camera is calculated,and the 3d coordinate information of table tennis is calculated using the calibration results.Then the algorithm of tracking image processing by comparing with the existing target is introduced,and a fast real-time image processing algorithm is developed based on the flight characteristics of table tennis.Based on the analysis of table tennis in flight,the flight path parameter model of table tennis was established,and the prediction of the flight model and the shot point were analyzed.The main research contents of table tennis robot are:first,multi-camera calibration:calibrating multiple cameras,obtaining the internal and external parameters of the camera,and drawing the relationship between multiple cameras.The second is the acquisition of three-dimensional coordinates:to obtain the image coordinates of table tennis flight under each camera,and to calculate the exact three-dimensional coordinates by using the calibration results.The third is to build a trajectory model,the trajectories of the table tennis is divided into three stages,respectively divided into stages of visual measurement,track prediction and placement and shot point prediction stage.The fourth is to analyze the parameters of the parameter model and analyze the sensitivity of the model,and the parameters can be optimized and adjusted by feasible methods.