Based on machine vision detecting system has the Characteristics of non-contact, real-time, high precision etc, in recent years it has became a hotspot in field of detection and has been used widely. Because the cannonball contour directly relate to the quality and performance of the cannonballs, this article mainly apply machine vision detection technology to detect the contour of the cannonballs real-time and on-line, in order to improve the rate of detection, inspection efficiency and the rate of correct judgment.This article that combined with application experience expounded machine vision inspection hardware system design, image processing software algorithm and using MATLAB simulation. This testing system has the advantages of high accuracy, strong function, simple hardware circuit, easy maintenance etc. And apart for it has realized automatic detection function, it also realized real-time detecting data storage, printing, display, good retrospective. Image processing is the core of this study, it is mainly contains image pretreatment, image segmentation, edge detection and sub-pixel edge detection algorithms, etc. Using these algorithms to the contour feature of cannonball and detect the error analysis. |