基于K-L算法的焊缝红外图像检测技术

Infrared Image Processing Technique for Welding Line Based on K-L Algorithm

提出计算机对所得图像进行滤波和图像分割,讨论了K-L算法的原理及其在红外焊缝图像检测中的应用;以红外光源照射焊缝,通过CCD摄取焊缝的图像,建立已焊和未焊的图像模板库,采用K-L算法对焊缝图像进行处理,并与模板库进行比较,得出焊缝是否焊接的结论;试验结果表明,该方法检测精度高,检测速度快,可以实时运用于对手工电弧焊的焊缝质量检测中。

A detection method for welding lines based on K-L algorithm is discussed. Infrared light source is applied to get high quality images of welding lines through CCD. Image template libraries for unwelded and welded lines are constructed and the welding line image is processed by K-L algorithm to match with the template libraries to judge if the welding line has been welded. The experiment shows that above method is effective with high precision and speed and it can be applied on real-time detection of welding lines.