一种基于Hough变换的大尺度线条提取算法
首发时间:2011-12-26
摘要:针对TFDS中对牵引梁弯曲故障的自动识别,本文在Hough变换的基础上提出了一种大尺度线条的提取方法,即先用Hough变换提取出图像中的感兴趣直线,通过编组和直线拟合定位牵引梁结构线,再用一种自定义的边缘扇形搜索算法将结构线补充完整从而得到我们需要的大尺度线条。该方法考虑了牵引梁弯曲的不确定性,可以兼顾直线提取和曲率较小的曲线提取,且对断裂的线条有填补和连接的作用。虽然传统的Hough变换法也能满足直线和曲线的提取,但提取曲线时需要其参数描述表达式,且在提取线条类型未知的情况下得不到很好的应用。反复实验表明,本文方法在提取牵引梁大尺度结构线时有较高的准确率,为故障判断提供了可靠的依据。
关键词: 图像处理 Hough变换 直线编组 边缘搜索 TFDS
For information in English, please click here
A Large Scale Line Extraction Algorithm Based on Hough Transform
Abstract:For the purpose of implementation of automatic recognition for draft sill warp, a method of large scale line extraction based on Hough Transform has been proposed in this article. Firstly, extract interested straight lines in the image through Hough Transform, then group and fitting them to locate structural edge of draft sill, finally make use of a fan-shaped edge search algorithm which is designed according to the practical problem to complete the large scale line extraction. The uncertainty of draft sill warp is considered, so that this method is able to extract straight lines and curve lines with small curvature at the same time, and is also capable of filling gaps between lines' discontinuity. Although traditional Hough Transform is within touch of both straight line extraction and curve line extraction, but the certain parametric equation which describes the curve line must be given before extraction, meanwhile, without aware of objects' line type, Hough Transform cannot perform well. Repeated experiments show that the algorithm proposed in this article can achieve high rate of accuracy when extracts large-scale structural lines of draft sill, which provides the reliable foundation for the fault judgement.
Keywords: Image Processing Hough Transform Straight Lines Grouping Edge Search TFDS
基金:
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
一种基于Hough变换的大尺度线条提取算法
评论
全部评论0/1000