Research On Spray Painting System Based On Kinect Calibration And Lance Trajectory Planning

At present,the automobile painting operation is mainly done by hand,The quality of the spray painting is greatly influenced by the technical level of the workers’ spray painting.This leads to the high labor cost,and the paint operation environment is usually very bad.All kinds of paint are very harmful to the human body,and the long term paint work is easy to cause various diseases.Therefore,speeding up the automation process of painting is very important.In China,most spray painting robots generate spray painting path through manual teaching to achieve fixed type and fixed area car painting operation,the efficiency is relatively low.In this thesis,we analyzes the posture calibration and spray trajectory planning of spray painting robot and focus on the accuracy optimization of posture calibration for painting robot and the problem of paint trajectory planning in scratched area of automobile paint.We use a new visual sensing device to design and implement an intelligent spray painting system based on Kinect calibration and lance trajectory planning.The system can automatically generate the trajectory of the spray gun for complex car paint scratching arearay and then we pass it to the spray painting robot through the high-precision calibration algorithm to achieve automatic painting work in the scratched area of the car.The main contents of this paper are as follows:(1)Calibration based on KinectCalibration is a key problem in the painting robot,if the calibration accuracy is not enough,the painting trajectory generation will not accurately transfer to the actual paint spray robot.In this thesis,we based on Kinect and singular value decomposition(SVD)method for obtaining high-precision pose calibrationand then we have verified the results of the calibration.(2)Preprocessing of three dimensional point cloudDue to the existence of interference factors,the 3D point cloud data obtained by Kinect usually have some noise.In order to obtain accurate and complete 3D point cloud data of the car scratch area,in this thesis,we have improve the original feature point registration algorithm based on the Sampson weighted operator,and verifiy the improved algorithm through the actual test.(3)The formation of the trajectory of a painting robotThe current trajectory planning algorithm is mostly aimed at the regular shape plane.For complex vehicle surface,the method of manual teaching is used to create spray painting trajectory.In this thesis,a method of automatically generating paint spray track in automobile scratch area is designed by using point cloud sectioning method,and the feasibility is verified by experiments.In addition,we also optimize the search algorithm used in point cloud slicing according to the time factor,and verify the algorithm.(4)The establishment of an intelligent spray painting systemA kind of intelligent spray painting system,which is built by Kinect,robot body,industrial control machine,spray gun,air pump and so on,is studied to realize the repair function of automobile paint surface.The system mainly includes the following functions:3D point cloud acquisition function,fault area detection function,success energy of spray gun trajectory,and motion control function of spray painting robot Kinect and the position function relationship of spray painting robot.Finally,the effectiveness of the system is verified by experiments.