基于焊缝裂纹检测的模糊缺陷图像恢复研究

Research on Fuzzy Defect Image Restoration Based on Weld Crack Detection

磁粉检测是碳钢原件表面和近表面最有效的无损检测手段,在复杂工况下,自动化磁粉检测技术仍然存在缺陷图像模糊、缺陷识别率低和误报等问题。针对裂纹图像采集过程中出现的模糊问题,建立匀速直线运动图像退化模型,求解运动模糊方向和尺度,采用逆滤波、维纳滤波和Richardson-Lucy 3种恢复方法进行恢复,结果证明采用Richardson-Lucy恢复方法可以较好的改善磁粉检测图像采集过程中存在的运动模糊现象。

Magnetic particle testing is the most effective non-destructive testing method for the surface and near surface of carbon steel components.Under complex working conditions,automated magnetic particle testing technology still faces problems such as blurry defect images,low defect recognition rates,and false alarms.A uniform linear motion image degradation model was established to address the blurring issue during crack image acquisition.The direction and scale of motion blur were solved,and three restoration methods,namely inverse filtering,Wiener filtering,and Richardson Lucy,were used for restoration.The results showed that the Richardson Lucy restoration method could effectively improve the motion blur phenomenon in the image acquisition process of magnetic particle testing.