Welding Defects Feature Recognition Based On The Bath Visual Mag Welding

In allusion to the characteristic of in MAG welding, a passive direct visual sensing system is designed and re-established. Typical and clear welding pool images can be shoot for different droplet transfers in MAG. The weld pool vision geometry, the shape and the dreg characteristic are gained and the gray characteristic is further described by using the false colored enhancement and 3-D gray image. With the synchronization experiments, connection between typical welding defects and the pool image characters has been systematically studied and the image characters have been gained corresponding to the weld defects such as surface pore, internal pore (dreg in weld) ,weld burn through , weld crinkle and so on. Changes and the strange characters of weld pool images have been studied by the gray average value and the standard variance in the weld defect generating process. The experiments indicated that, it is feasible to judge pore and other weld defects from weld pool characters, which have provided the technical basis for weld defect automatic diagnosis which is based on visual sensing; The experiment and the analysis also indicated that, one kind of image character possibly indicated that many kinds of weld defect produce and one kind of weld defect possibly demonstrated many kinds of image characters.