The Welding Robot Has No Teaching Mode Three-dimensional Weld Information Extraction

In order to improve the degree of automation of welding robot,the problem of 3D information detection of weld seam in the case of no teaching format is studied.Digital Structure Light (DSL) 3D scanner is added into the robot system.And the eye-to-hand visual system is formed.The point cloud data of weld is obtained by DSL 3D scanner.The weld feature is detected by using two kinds of algorithms.The math expression of weld seam is obtained and the path of the welding robot is planned.The primary research works are as follows:First,the problem of calibration of robot vision system is studied.The cameras' internal and external parameters are obtained and the relationship of the cameras and the robot is also obtained with the help of DSL 3D scanner.The three-dimensional point cloud data of weld seam is obtained by dealing with the pictures of the camera.Second,two kinds of algorithms to detect the weld seam are studied.One is basing on the change of the normal vector between points.After evaluating the normal vector according to the local data of the weld surface,the weld seam could be detected by the angle between normal vectors.The other algorithm is basing on the coordinate information.After calculating the distance between each point and the upper surface,the weld seam could be detected by the distance.And the difference between the results through the two algorithms and the results through Imageware is compared.Finally,the feature point of weld seam are fitting by the method of polyfit and the math expression of weld seam are obtained.The approaching vector,directing vector and normal vector of welding gun are calculated.The path of the welding robot is planned and showed.The error of the math expression is analyzed by comparing the detecting result of Imageware.And the precision can meet the engineering practical needs.