| Component parts with micro internal bore such as pin valve take up the very large proportion in the modern manufacturing industry. At present, traditional man-detection method is used widely in our country of the aperture's precise detection. This method is not only measuring inefficiency, and difficult to control the precision. In order to make fast and accurate automatic detection, and quality control, the micro internal bore surface quality is characterized by the technique, which is based on expert system precise detecting in this paper.By tools of image processing combining pattern recognizing technique and expert system, this paper mainly object is how to realize the precise judgment on surface quality of the component parts, and make an intensive study on micro internal bore's detection. To make sure the micro internal bore image reliability in the hardware, the system uses CCD camera, straight rod optical fiber endoscope and acquisition card to obtain the image, and to raise the detection accuracy. Based on pattern recognition, expert system and image processing such as the gray, the value of two, the median filtering, edge detection, image threshold processing, contour extraction, characteristic and statistical parameter extraction of the micro internal bore and so on, to achieve the component precise detection. This thesis studies the detecting technique, with the needle valve surface quality of the micro internal bore. Using computer programming and experiment concludes a micro internal bore surface quality precise detecting technique, which is based on expert system, it provided technical support to such component surface quality detection, digital image acquisition and processing, expert system and pattern recognition technique.This system is the embodiment of mechanical-photoelectric integration. It combines image processing technique, expert system and pattern recognition technique, satisfying the micro internal bore precise detection, achieving better identification effect. So studying on micro internal bore of the non-traditional detecting method, has an important application value. |