摘要
提出了一种基于Sobel算子的亚像素边缘检测方法。此种方法先由Sobel算子确定图像边缘的大致位置,然后用三次样条插值函数对灰度边缘图进行内插计算,使目标边缘定位达到亚像素级。对插值后的灰度边缘图,利用最大类间方差法确定阈值提取边缘。实验表明,该方法能精确定位目标边缘,优于传统的边缘检测方法。
A sub-pixel edge detection method based on Sobel operator is proposed in this paper. This method determines the probable position of image edge by Sobel operator and then takes three-order spline interpolation function to interpolate the gray edge image, which makes object＇s edge located at sub-pixel level. It adopts maximum between-cluster variance method to determine threshold and abstract edge in the interpolated gray edge image. Our experiment proves that this method can aecurately locate object＇s edge, which is better than the traditional edge detection method.
出处
《南昌航空工业学院学报》
CAS
2005年第2期100-102,共3页
Journal of Nanchang Institute of Aeronautical Technology（Natural Science Edition）
基金
国家自然科学基金资助项目(60141002)
江西省测控中心基金资助项目(KG200204011)。
关键词
边缘检测
亚像素
三次样条插值
edge detection
sub-pixel
three-order spline interpolation
作者简介
朱颖（1982-），女，南昌航空工业学院计算机系硕士研究生，研究方向为图像处理，模式识别。
