The Research On The Computer Vision Based Technology Of Automatic Orientation And On-Line Detection Used In CAM Grinding

The emphasis of this thesis is the mechanical application of computer vision. A computer vision based solution is proposed to solve the problem of automatic orientation and on-line detection during cam grinding. The development of computer vision, the basic theory of numerical image and the key algorithms of image processing are introduced firstly. According to the practical cam image, image processing results by each algorithm are compared and analysed. Finally, thresholding and median smoothing are choosen to get satisfactory result. In the theoretical research part of this thesis, the principles of computer vision based cam automatic orientation and detection are proposed. According to the principles, two means of system hardware set-up are designed. The comparison and analysis of advantages and disadvantages between these two set-ups are done from the aspect of precision and computing speed. Finally, the means of installing the laser and CCD camera facing to the cam is chosen. In addition, combining with the characters of laser stripe, image processing algorithm for laser strip is designed according to different requirements to speed and precision. The above algorithms are realized by Visual C++ program, and the computer vision based cam grinding automatic orientation and on-line detection system software is developed. Finally, system hardware and software are introduced and system test is done to automatic orient and on-line detect disk cam on NC cam grinder. The result certifies the correctness of orientation and detection principles as well as the efficiency of image processing algorithms.