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　corresponding　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.