用于图象边缘检测的拉普拉斯-多项式算子
LAPLACIAN OF POLYNOMIAL OPERATOR FOR IMAGE EDGE DETECTION
-
摘要: 将用于求解二阶导数的▽2算子与多项式局部曲面拟合相结合,构造了用于边缘检测的拉普拉斯-多项式(Laplacian of Polynomial:▽2P)算子.图象物体边缘定义为▽2P与图象卷积的过零点.证明了:采用正交多项式作为拟合基函数时,可用卷积方法实现局部曲面拟合,使得计算效率大大提高.Abstract: An edge detection operator ▽2 p is developed by polynomial local surface fittingcombined with Laplacian operator.Edge point is defined as zero-crossing point ofthe output of ▽2 p.It is proved that,based on the orthogonal polynomiel function,the local surface fitting can be realized by convolution,thus the amount of operationis largely reduced.