Fast polynomial fits sub-pixel edge detection algorithms

Since traditional edge detection algorithms have the disadvantages of low precision and efficiency in measuring the edge,a sub-pixel edge detection algorithm based on the polynomial fittings is proposed.According to the grey level distribution of the picture,this algorithm uses cubic polynomial to fit the edges to realize sub-pixel localization.Traditional sub-pixel edge detection algorithms detect coarse position at first,and then carry out the sub-divide,therefore the running time is relatively long.This method first fetches some points near the edge,and then carries out the sub-pixel edge detection algorithm,so the running time is reduced.Finally,the experimental results show that the proposed algorithm is very reliable and efficient.