摘要
针对采用传统方法检测铸件缺陷存在检测效果不明显的问题,为提高检测效率,研究基于工业CT的铁路货车铸件内部缺陷自动检测算法。首先用工业CT扫描重建得到铸件的断面图像,然后对图像计算分形维数,利用铸件缺陷区域分形维数较高的特点自动定位出缺陷的大致区域,再用Facet模型对定位区域进行边缘检测,以便确定缺陷的准确形状。用工业CT对摇枕和侧架进行缺陷检测的结果表明,该方法不仅能获得铸件内部缺陷的位置和准确形状,而且处理自动化程度较高。
To detect the casting defects with traditional methods generally results in inconspicuous effects. In order to improve detection efficiency, the method for automatically detecting the inner casting defects of railway freight car based on industrial CT is presented. From the cross-section image of the castings obtained by industrial CT reconstruction, firstly, fractal dimension of the image is calculated and judged self-adaptively to locate the defect regions. The locating method is based on that the fractal dimension of a defect region is bigger than those without defects. Then, the edges of defects are detected by Facet model within those located regions in order to get the exact shapes of defects. The results of the actual data acquired by scanning the bolster and the side frame with industrial CT indicate that this method can not only automatically give the positions and exact shapes of inner defects in casting, but also can process with higher degree of automation.
出处
《中国铁道科学》
EI
CAS
CSCD
北大核心
2009年第4期76-80,共5页
China Railway Science
基金
国家自然科学基金资助项目(60672098)
国家"八六三"计划项目(2006AA04Z104)
