|
Abstract: Wavelet transform and mathematic morphology algorithms are
used to process multi-pass seam image to get weld pass and groove’s
edges. Firstly, 2-D wavelet transform algorithm is used to process
multi-pass seam image and get its dark section and bright section
information, since the absolute peak values of dark section and bright
section are greater than other section, process the results of image’s
wavelet transform into binary image to get the location information
corrisponding peak values. Secondly, mathematic morphology algorithm is
used to process dark section and bright section images according to
their peak values’ location characters. At last the edges of weld pass
and groove are picked up. The edge information achieved via this
algorithm is in good agreement with that of original image.
Key words: Wavelet transform Mathematic morphology Dark section
Bright section
CLC No: TG409
国家863重大资助项目(2002AA602012-3).
Received 20040424, received in revised form 20041230
|