Design And Implementation On Edge Defect Detection System Of Strip Based On Gaussian Kernel Function

The Iron and steel industry represents a country’s level of industrialization andaccounts for a non-negligible proportion in the country’s economic component. The stripis one of the important products in the steel industry which is useful and can be used in themanufacture of bicycle frame, wheels, blades and so on and its quality will affect thesecurity of many industries. The strip quality judgment standard, whether one is platequality. Shape quality is one of the criteria to judge the quality of strip steel. The camberscrap occupy very large proportion in all shape quality scrap and how to detect the edgedefect is a meaningful and challenging task.Detection of the camber scrap is to detect the strip having protrusions or depressionson their edge and you can regress the edge to determine whether it has a defect. The leastsquares curve fitting is easily affected by the over-fitting and slow convergence speed.Kernel regression is one of the nonparametric regression estimation method and don’tneed to assume the form of function which avoids the error caused by model assumes.Inorder to detect camber defects of strip accurately,we combine the disciplines ofmathematics, computer vision and artificial intelligence from the po int of view ofcomputer applications, propose and implement edge defect detection system of stripbased on gauss kernel function. Based on the selection of kernel function, parameterdetermination, the experimental results obtained, edge defect detection system of stripbased on the gauss kernel function has high recognition rate and strong robustness.