Flaws Detection For The Tubular Bottle

Machine visual inspection is one of important research areas in information processing fields. This dissertation mainly uses the technologies and methods of digital image processing to develop a system of flaws detection for the tubular bottle, and researches the algorithm of software implement and the method of hardware implement.Firstly, in the dissertation, some basic conceptions of machine vision and digital image processing are introduced, including their histories, characteristics, application fields and future developing territories.Then the structure of the inspection system is introduced, especially the principle of image collection.This dissertation expatiates on some digital image processing technologies that the inspecting system may apply to the research and analysis of the collected bottle-flaw image. For example, using neighborhood averaging method and median filter method to restrain the image noises and make the image edge smooth, splitting image by threshold mode and image segmentation method based on the edge detection, calculating the area and perimeter of the bottle-flaw image, and using the circularity to judge whether this image is a bottle-flaw image.During the course of using the digital image processing to research and analyze the bottle-flaw image, Microsoft Visual C++6.0 program language is applied to develop this inspecting system. The modularization program method is used in the storage of image, image segmentation, feature selection and flaws judgment.