Fuzzy detection of welding image based on wavelet and morphology

Image detection is an important step of welding process automatic control. In view of the welding image feature of strong noise and poor stability, through comparing the two algorithms between wavelet transform and morphology, a fuzzy detection algorithm of welding image based on wavelet transforms and morphology is presented. An original image is denoised by wavelet operator and morphology operator respectively, and then the two denoised images are merged and fuzzy processed, at last, the morphology edge should be detected. The experimental results showed the approach could acquire the clear and precise weld pool edge, and the detection result could be used in welding penetration control.