利用梯度相位提取图像中的直线  

Detecting Lines of Image Based on Gradient Phase

作　　者：韩辉[1] 

机构地区：[1]国防科技大学机电工程与自动化学院,湖南长沙410073

出　　处：《数字技术与应用》2012年第11期198-199,201,共3页Digital Technology & Application

摘　　要：针对基于图像梯度幅值的直线提取算法计算量大、分辨率低的不足,本文利用梯度相位信息进行直线提取。算法首先计算图像梯度幅值和相位,按梯度相位编组直线支持区域,然后使用改进最小二乘法拟合直线,最后根据一种新的直线判定算子判定所提取出的直线是否对应直线模型。实验证明,本文方法能够从图像中快速准确地提取出大量直线。For the large amount of calculation and low resolution of line detection method based on gradient magnitude, this paper uses the gradient phase information for line detection. The algorithm first calculates the the image gradient magnitude and phase, organizes line support regions by gradient phase and then use improved least squares method fitting lines.Finally determine whether the detected lines corresponding to line model according to a new line operator. The experiments illustrate that the proposed method can detect a large number of lines quickly and accurately from the image.

关 键 词：梯度 相位编组 直线提取 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]