Recognition of magneto-optical image of micro-gap weld using fractal dimension method

Magneto-optical imaging method was used to detect the position of micro butt joint, which is non-grooved and almost indistinguishable by naked eyes. The problem of interference in weld magneto-optical images was solved by using the fractal dimension method. Magneto-optical images were processed based on relative information in a large field of vision, which can avoid the influence of tiny detail disturbance among images on measurement accuracy of weld position. In comparison of conventional image processing method, the fractal dimension method is more capable in anti-interference. The weld magnetooptical images were denoised and divided into blocks, and then the fractal dimension of each image block was calculated. An appropriate threshold was selected to segment the images, and the weld center position can be extracted accurately. Experimental results show that the weld edge region characteristics can be accurately extracted and the weld position can be detected by using fractal dimension method, providing an important basis for seam tracking.