Object Recognition Based on Smoothing Segmented-contour
史思琦 SHI Si-qi, 石光明 SHI Guang-ming, 陈旭阳 CHEN Xu-yang, 李甫 LI Fu · ACTA PHOTONICA SINICA · 2010
An object recognition algorithm based on segmented-smoothing contour is proposed.The contour is divided into two kinds of zones by curvature:feature zones and non-feature zones.Then,the different kinds of zones are smoothed with Gaussian functions with different variance.Finally,the object recognition algorithm based on affine moment-invariant is applied to recognize the smoothed contours.The simulation result shows the better contour-smoothing results and the significant improvement to recognition rate.