Abstract: A novel image processing approach was proposed to reduce the noise in surface mount technology(SM7) solder joint image, based on wavelet packet transform and Wiener filter. At first, by the using wavelet packet transform, the approach not only decomposed the image into the low frequency part but also into the high frequency part of image in several scales so that the noise could be eliminated and the image useful information could be reserved. Then, after analyzing the wavclet packet tree coefficients, they were processed with the use of the Wiener filter, and leaved the wavelet packet tree low frequency coefficients without change. At last, the inverse wavelet packet transform was applied to reconstruct the SMT solder joint denoised image. The experimental results have illustrated that the proposed approach can accomplish a better result in image denoising compared with the conventional methods and can retain most image edges.

Key words: Surface mount technology solder joint, Image denoising, Wavelet packet transform, Wiener filter