摘要
针对标准Hough变换直线检测存在的问题,提出了一种基于Hough一维空间变换的直线检测算法。对图像进行边缘检测,将互不连通的边缘进行分组;利用Hough一维空间变换对边缘组进行直线检测并编组,降低了算法和空间复杂度;通过对直线组进行精确处理,解决了过连接、端点确定和检测精度存在的问题。实验结果表明:该算法有效地解决了标准Hough变换算法中存在的问题,降低了误检率和漏检率,提高了直线检测精度,能够实现对不同类型影像数据的直线检测处理。
Aiming at the existing problems in line detection by standard Hough transform, an improved algorithm based on Hough one- dimensional transform is proposed. Firstly, the edges of images are detected and the disconnected edges are grouped. Then, the Hough one-dimensional transform is adopted to detect in line and group the edges so as to reduce computation complexity and space complexity. At last, the problems on over-connecting, endpoints determining and detection precision are solved by added accurately processing. The experimental results prove that the improved algorithm can effectively solve problems in the standard Hough transform, reduce missing detection and false detection, improve the detection accuracy and achieve detection in line for different content images.
出处
《光学学报》
EI
CAS
CSCD
北大核心
2016年第4期158-165,共8页
Acta Optica Sinica
关键词
测量
直线检测
Hough一维空间变换
边缘编组
直线编组
meaurement
line detection
Hough one-dimensional space transform
edge grouping
line grouping
