Vectorizaiton of lattice diagram based on straight line and circular arc fitting in variable-step

Vectorization of the lattice diagram after contour segmentation is a critical techniqure to keep tufted carpet from destortion in the process of manipulator-tufting. A new algorithm is put forward to realize the vectorization, and keep the fitting error in an allowable range. The process of curve vectorization includes several steps: Form a feature point set by extracting the Freeman chain coding feature points from the pixel points of the image, then run straight line fitting and circular arc fitting in variable-step, and finally obtain the contour curve represented by linear and arc equations. The algorithm in this paper is mainly used for carpet image recognition and transfer in a manipulator-tufting system. Through VC++ programming of the algorithm and then repeated experiments, it is proved that the algorithm is effective, feasible and quick to especially satisfy the requirement of high efficiency.