The Research Of Defect Detection Method For Metal Cans Wall Based On Image Segmentation

Metal cans are a commonly used metal packaging container with the advantages of a good seal and anti-high internal pressure compared to paper containers.Due to the characteristics of their own materials and the impact of the production process,common defects of edge deformation,sand hole,pits and smudge are appeared in metal cans inside the wall,and its inner wall defect is an important factor in measuring the quality of metal cans.The surface defect detection based on machine vision is a fast non-contact non-destructive testing method,in order to the realization of the automatic detection method of the inner wall defects of the metal cans based on machine vision,under the premise of the existing metal cans sample,the image acquisition and image processing of metal cans were studied.The development of machine vision both at home and abroad were reviewed in this article,and focuses on the research progress and present situation of the defects detection of metal cans based on machine vision.The characteristics of the similar research and the research demand on the defect detection of metal cans wall based on machine vision were pointed out.Importance of image acquisition based on machine vision was analyzed.Theory foundation of the camera,the lens and the light source was elaborated.The selection basis and method of image acquisition device were given.A two-stage image acquisition scheme based on machine vision was proposed,and captured the images based on the scheme.Digital image processing theory was introduced,and the technical route and process flow of the inner wall defect detection of metal can was designed.The sample images captured were processed based on the visual software Halcon.The experiment results show that the images with some contrast in the defect can be captured with the designed image acquisition scheme,and defect detection of edge deformation,sand hole,pits and smudge can be realized with the used image processing algorithm.The research of this paper has some reference significance tothe realization of defect detection system of metal cans wall based on machine vision.