Image matching for weld seam tracking based on nonlinear dimensionality reduction method Isomap

An approach of image matching for weld seam recognition is proposed based on the nonlinear dimensionality reduction algorithm Isomap. The target image to be recognized and the template images are considered as high dimensional data. The reduction to 2-D coordinate helps to estimate the relative location of the images. A method of image matching is developed, based on the method of Isomap.