摘要
针对传统的边缘检测算法精度不高,提出了基于不同核函数的最小二乘拟合的亚像素边缘检测算法。采用Canny算子对图像边缘进行粗定位,再以基于梯度幅值拟合的高斯函数与基于灰度值拟合的反正切函数和双曲正切函数作为拟合核函数,对一幅拍有标准量块图像的边缘进行检测,结果表明该算法能够有效的进行亚像素边缘检测。
In view of the inaccuracy edge detection of traditional algorithm,the sub-pixel edge detection algorithm based on least square fitting different kernel functions is proposed.It adopts Canny operator for the coarse positioning of the image edge,and combines Gauss’ s function based on the gradient amplitude fitting and inverse tangent function and hyperbolic tangent function based on gray value fitting as kernel fitting function.By detecting an image edge with standard gauge,the results show that the sub-pixel edge detection algorithm can perform a sub-pixel edge detection effectively.
出处
《软件导刊》
2018年第2期60-62,67,共4页
Software Guide