改进可拓理论的带钢表面缺陷图像分类方法

Classification of Steel Strip Surface Defect Images Based on Improved Extenics Theory

将可拓理论引入到带钢表面缺陷图像的分类中,提取分割前后缺陷图像的12个特征值,对可拓理论用于分类的关键步骤-关联度的计算方法进行改进,首先确定某缺陷特征值与某类缺陷对应特征值经典域的距绝对值,再计算出该特征值与各类缺陷对应特征值经典域的距绝对值之和,以二者的商作为关联度计算的权值.对该方法的可行性进行了论证,并对带钢表面缺陷图像分类进行了仿真,仿真结果显示改进后的方法分类准确性有较大的提高.

Extenics theory is introduced into steel strip defects＊ images classification. Twelve features are extracted from segmented and unsegmented images. The key step-relevancy values computering method is improved. The quotients of distance between feature values and classical domain and sum of these distances are used as weight coffecients in compute ring comprehensive relevancy values. This computering method enhances the influences of defects ’ self feature values on comprehensive relevancy. The effectiveness of the improved computering method is also demonstrated. Steel strip defects are selected to simulate the method, maximum relevancy value is used to group the unspecified defect image in one of preselected defect types, comparing to primary weight coefficient computering method, the improved theory is more effective in defects images classification.