Development of a model for friction stir weld quality assessment using machine vision and acoustic emission techniques

Abstract The surface texture of friction stir weld is related to internal structure of weld. The statistical image parameters along good and defect weld regions are quantitatively evaluated for quality from the weld bead images, which are processed and analyzed using machine vision technique and X-ray radiography. The weld strength obtained along the weld regions showing similar variations in Acoustic Emission (AE) data, which was acquired during welding and image data of the weld bead are analyzed to assess the weld quality. The combined model developed using limiting values of image data and AE data along different regions of weld evaluate the weld quality more reliably.