一种泡沫金属材料的裂缝缺陷检测方法  

A Cracks Detection Method of Foam Metal

作　　者：徐子雯[1] 何金镁 Xu Ziwen;He Jinmei(Central South University,Changsha Hunan 410000,China)

机构地区：[1]中南大学,湖南长沙410000

出　　处：《信息与电脑》2017年第2期63-66,69,共5页Information & Computer

摘　　要：针对泡沫金属材料裂缝缺陷会影响产品性能这一问题,笔者通过对泡沫金属图像进行两步定位的方法完成裂缝缺陷检测。在初步定位过程中,为减小光照影响,首先对待检测泡沫金属图像进行直方图均衡,然后运用形态学滤波提取疑似裂缝缺陷区域。在精确定位过程中,根据裂缝缺陷的自身特点,利用不同的模板提取疑似裂缝缺陷的形状和纹理特征,并输入Adaboost分类器进行检测。实验结果表明,该方法可以较为准确地定位泡沫金属图像中的裂缝缺陷。For the problem that foam metal cracks have a bad impact on the product performance,the auther proposed a two steps method for foam metal crack defect detection.In the preliminary detection process,for reducing the illumination effect,histogram equalization is applied to process foam metal image to be detected firstly,followed by a morphological filter to extract candidate crack defect region.In the precise detection process,considering the characteristics of crack defects,the shape and texture feature of candidate crack are computed based on different model and then input to Adaboost classifier for detecting precisely.The experimental results show that the method has a good performance in the foam metal crack defect detection.

关 键 词：泡沫金属 缺陷检测 形态学滤波 ADABOOST 

分 类 号：TB383.4[一般工业技术—材料科学与工程]