Edge detection of metal transfer image based on wavelet transform

Wavelet transform and improved Prewitt operator were used to detect the edge of the droplet image in GMAW.After preprocessing, wavelet transform was used to perform decomposition to different detailed levels, and enhance diagonal detail coefficients by multiplying it with the proper parameters, then reconstruct image. Improved Prewitt operation was used to detect the edge of reconstructed image.The experiment results demonstrate that the method is effective to detect the real outline of droplet and real edge of weld pool, and the levels of decomposition and the multiples are vital factors, which depends on the real image.