作　　者： ; ;

机构地区： 华南理工大学电子与信息学院

出　　处： 《数据采集与处理》 2007年第4期475-478,共4页

摘　　要： 提出了虚拟铸件缺陷X光图像的模拟算法,以方便对检测过程中所用的算法及在线系统进行测试。该算法主要针对轮毂缺陷中常见的疏松缩孔进行模拟。先建立通用模板限定待处理图像区域,在选定区域内综合利用图像灰度特征和空间位置特征生成缺陷的形状,调整缺陷参数,将该缺陷图形经灰度变换后叠加至缺陷的期望位置。实验结果对比表明,所生成缺陷在形状和灰度方面逼近真实缺陷,且形状不单一,具有随机性。 To evaluate the sensitivity of the online inspection algorithm and the system, a new approach for creating virtual radioscopic images of casting defects is presented. The most common defects of the aluminum casting-the shrinkage cavities and porosities, are simulated. Firstly, the normal template is constructed to restrict the size of the image area. Then the shape of simulation flaw is taken using gray features and spatial location information of the data, and the flaw parameter ia adjusted to the expected scale. By the grey-level adjustment, the simulation flaw is overlapped to the real radioscopic images of the aluminum casting. Experimental results show that the proposed method is effective in making virtual flaws, and their shapes and the grey level are more stochastic and natural.

关 键 词： 检测 光图像 轮毂 缩孔和疏松 模拟

领　　域： [自动化与计算机技术] [自动化与计算机技术]