Based On The Size Of The Mechanical Parts Of Machine Vision To Identify

The measurement based on machine vision is characteristic with rapid, high precision and flexibility. It has been studied and applied widely at home and abroad. The thesis aimed to measure the dimension of the circular features in machine parts. In this paper, the machine vision technology is studied for circular features detection and its diameter.The performance parameters of the machine vision hardware are analyzed. All the parts of the machine vision system, including the camera, the lens, the illumination and the frame grabber, are chose and assembled. The flow chart of the software for the machine vision system is given.There are inevitable aberration, noise and blur in the images that are acquired by the camera because of the non-ideal lens and camera. The image enhancement, image filtering and edge detection were studied and applied. An image preprocessing method was raised. The Sobel operator and the medium filtering were used in the machine parts image processing. The camera calibration based on the pin-hole camera model is completed with the help of the OpenCV library. Using the camera intrinsic parameter, the image rectification transforms the image to compensate lens distortion.The Hough Transform (HT) is studied and the Random Hough Transform (RHT) is proved. The proved method of RHT is used to circular feature identification in the image. The pixel precision parameters of the circular feature are calculated by the proved RHT. Sub-pixel edge location technology is a key method for improving measurement precision of image and decreasing the cost of the hardware. Spatial moment sub-pixel edge location technology and least-squares approximation sub-pixel edge location for the edge of circle are studied in this paper. The method to locate the edge in sub-pixel and calculate the circle.parameters in this...