机器视觉在大型工件自动探伤检测中的应用

大型金属工件探伤一般用无损检测法，技术难度高、工程量大，往往难以获得精度的测量准确值。针对机器视觉产品将被摄取目标转换图像信号后，多数都还依赖于手工进行处理问题，文章提出了一种基于边缘撮的工业CT（断层）图像几何尺寸的自动测量方法。该方法采用Canny算子提取图像边缘信息，并在图像边缘上实现对大型工件CT图像的内部结构尺寸和缺陷尺寸的自动测量。该方法已成功地应用于工业CT图象处理软件中，实际应用表明，它具有较高的精确度和重复性。

Application of Machine Vision in Automatic Inspection of Objection for Large-Scale Work-piece

The scathe-less detect method is always used in the inspecting objection of large-scale metal work-piece, it is higher in technology difficuIty, more in engineering quantity, and always difficuh to obtain high accuracy measure value, Aimed at processing problem that is still mainly dependent of handwork operating after machine vision product transform incepted object into image signal, the paper presented a sort of automatic measure method that takes geometric dimension of Industrial Computerized tomography images based on image edge. This method adopts Canny operator to extract edges. And from this, the inner geometry parameters of the large-scale metal work-piece such as area, perimeter and pole can be measured automatically. This method has been successfully applied in the image processing software for industry computerized tomography images. The engineering application effect shows that it is higher in measurement accuracy, and better in measurement repeatability