The Visual Inspection Of Strip Suraface Defects And Implementation Of Algorithm

With the rapid development of modern industry, strip steel products is using more and more widely in the industry of aviation, ship, automobile and other manufacturing, and the quality of the products is increasingly being attention, therefore the control of quality in the production process has the increasingly significant meaning. As one of the common quality problem in the strip products, the surface defect not only affects the normal use of products, but also produces a bad influence to the image of the product and enterprise. So the effective detection and identification to the strip surface defects become an important direction in the quality testing field.In this paper, it uses the strip surface detection system of machine vision to detect and identify the surface defects of strip products. As one of the mainstream detection system, this detection system has many advantages of light, high precision, the strong anti-jamming ability and easy configuration, etc. It can produce quite intuitive detection effect to the surface of strip products and is easy to do the comprehensive assessment of the surface quality of strip products.The hardware of surface inspection system of machine vision include industrial CCD cameras, lenses, vision lighting system, image acquisition card, movement contained units and computer, etc. In this paper, it chooses and analyzes the various hardware modules, and built the strip surface inspection platform of machine vision on this basis. At the same time, it also does the experiment and analysis about some key problems in visual inspection system, such as the design of visual lighting system and the camera calibration, and finally gets the corresponding conclusion.For the section of image processing in the surface inspection system of machine vision for strip products, it does the design and research on the image processing algorithms for the strip surface defects which are scratches, scarring and cracks, and its main contents include the related algorithms which are about image preprocessing, image segmentation, grayscale morphology and the identification of defect feature. Based on the selected and designed image processing algorithms, it uses MATLAB and C++ to achieve the measurement of the aspect ratio, area, perimeter, and the smallest circumscribed rectangle and other morphological characteristics for different defects, and ultimately attains the effective detection and identification for the strip surface defects.