摘要
对于产品内部零部件的漏装、错装及相互之间空间位置关系是否正确的快速自动无损检测,依旧是行业内没有很好解决的难题。研制了基于单相机的X射线视觉检测系统,获取产品在全周向视角下垂直于射线光路的图像序列,提取识别特征,完成快速自动识别。研究了系统成像的物理机理,提出了空间采样频率与先验已知的最小结构尺寸及系统分辨率之间关系的空间采样准则,保障空间结构信息的完整获取;剖析序列图像之间的有序相关性,实现在单一或有限方位下利用不完全数据的快速识别理论。统计结果表明,在平均误检率小于4.5%的情况下,系统对每个产品的平均识别时间小于5s。
It is still an unresolved issue to test the assembly structures inside products in industry applications. Spatial sampling criteria based on the single view imaging system are proposed for recognition of structural objects. There must be a maximum rotary step for an object within which the least structural size to be tested is ascertained. Rotating the object by the step and imaging it and so on until a 360~turn is completed, an image sequence is obtained that includes the full structural information for recognition. The maximum rotary step is restricted by the least structural size and the resolution of the imaging system. Based on the spatial sampling criteria for structural recognition, the theory of fast recognition of structural objects is proposed. It is verified that objects could be fast recognized at a single or some limited orientations with a single camera by analyzing the smooth correlations among the image sequence sampled from 0° to 360° orientation. The theory is applied to the X-ray digital radiography system designed for online automatic inspection of interior assembly structures of complex products. Experiments show that the average identification takes less than 5 s in test nroress at the wrong recognition rate of 4.5 %.
出处
《激光与光电子学进展》
CSCD
北大核心
2013年第12期99-106,共8页
Laser & Optoelectronics Progress
基金
国家自然科学基金(601171178)
山西省自然科学基金(2012011010-3)
2012年度山西省高等学校优秀青年学术带头人支持计划
关键词
成像系统
X射线
结构识别
快速检测
有限方位
imaging systems
X-ray structural recognition fast test limited orientation
