| Since the laser technology was applied to welding area in 1970, laser welding was widely used in various industry fields. With its high precision, high quality, high efficiency, low distortion and high speed, and many other features, laser technology is the most watched and promising welding technology in the 21st century. In shipbuilding, metal, processing industry, aerospace industry, nuclear industry, electrical industry, semiconductor industry, biomedical and other industries, laser welding technology plays a significant role.In the process of laser welding, it’s very important to accurately detect the weld seam position automatically, because the precise seam tracking is the premise of high quality for welding. Especially with the increasing demand of the precision welding and reducing the cost, the welding problem of no groove and tight butt-weld micro gap is ugly to be solved., Weld wire should not be filled when butt-weld micro gap is welded, therefore, much small welding deformation and much even weld joint can be obtained, and joint defect detection after welding and repair work can also be reduced. At the same time, the welding efficiency and economic benefits can be significantly increased. Especially for the laser welding, with the high laser power density, small light spot diameter and allowed weld path deviation, usually when the deviation between the laser beam and the weld is more than 0.2 mm, welding failure will be caused. So the technology of automatic detection the butt-weld gap and seam tracking is very important.Faraday magneto-optic effect method is used to detect the position of micro butt joint(O-O.lmm), which is butt-weld tight and non-grooved. First, put the test specimen in the magnetic field which strengthen is defined in advance. Then, find out the Magneto-optic images of weld through magneto-optic sensor. The distortion of magnetic field would appear around the weld, so there is a weld transition zone in Magneto-optic images. When grasp Magneto-optic images with magneto-optic sensor, the quality of images would be affected by the interference produced by nearby environment. Therefore, the fractal dimension is used to solve the problem of interference in weld magneto-optic images. Magneto-optic images are processed based on relative information in a larger field of vision, which can avoid the influence of small disturbance among images on measurement accuracy of weld position, In comparison of other traditional image processing methods, the fractal dimension method has better ability of anti-interference.Capture the fractal dimension with differential box-counting, double blanket and Fractal Brown Motion, and then extract the result after the comparison of weld position. After that, denoise and divide the weld magneto-optic images into blocks and calculate each image block’s fractal dimension. An appropriate threshold is selected to segment the images, and the weld seam center position can be extracted accurately. So that, the system state equation and measuring equation would be build based on the weld seam center position. In the end, estimate the optimal position under Kalman filter and calculate the optimal value of welding seam deviation with minimum mean squared error. Experimental results show that box-counting can accurately extract the weld seam position of Magneto-optic images and Kalman filter can provide significant foundation for seam tracking control with optimization results. |
PDF Full Text Request