Research And Application Of Visual Image Localization Method Based On CCD Sensor

With the development and application of information technology such as computer technology and image processing technology, computer vision is widely used in industries and fields such as industrial inspection, precision measurement and control, automatic production lines, postal automatization, medical microscopic operation, even in some special fields, for example scientific research works in abominable or extreme environment, computer vision is now a necessity. This thesis tries to achieve positioning, measurement, calibration and verification of the multi-disc assembly components assembling process, based on COGNEX Insight and its industrial robot automation control system, using vision positioning technology to make balancing ring components to solve the imbalance during assembling process of Winchester Model hard disc drive disk-pile modules.The author has accomplished the following work:(1) Analysis of the current status of Industrial robot vision in domestic and abroad, analysis and description of the key technologies based on visual robot image positioning method, including image segmentation, grayscale morphological processing, edge extraction via edge examination algorithm, etc.(2) Crate two stage decomposition method which based on image binarization algorithms, to complete the acquisition and processing of accurate imaging sample, gather statistics of its optimal threshold based on image greyscale distribution, this image positioning method can obtain dynamic threshold from image gray level, complete these image preprocessing process such as separating invalid area, filtration, enhancement, binaryzation, etc, thereby complete the positioning system image.(3) Test and get the positioning and measuring data of image sample and the Reliability&Repeatability of the experimental model, guide the designing and manufacturing of gage and balancing ring module, apply to the assembly station of hard disc drive disk-pile modules, verified this image positioning method function in calibrating the balancing module’s parameter such as position, height, brightness, color aberration, etc.In this thesis, robot vision positioning method based on CCD image sensors is analyzed and designed, using gray-level morphological operation, edge detection, image binarization algorithms, it has realized the balancing ring module live adjustion in online detection and monitoring of the producing and assembling process, controlling the quality and eligible rate, achieved the aim of improving the products’ reliability reducing production cost and scrap rate.