基于高效Hough变换的圆柱特征检测方法  

Efficient Hough transform based cylinder feature detection algorithm

作　　者：杨向东[1] 芮晓飞[1] 谢颖[1] 

机构地区：[1]清华大学机械工程系,北京100084

出　　处：《清华大学学报（自然科学版）》2015年第8期921-926,共6页Journal of Tsinghua University(Science and Technology)

摘　　要：椭圆Hough变换方法是特征提取领域重要的研究课题之一。为提高算法的精确度和时效性,提出了高效Hough变换(EHT)圆柱特征检测方法。该方法适用于检测圆柱特征中的椭圆轮廓,利用单个参数的近似值,估计其他参数的近似值。为控制算法误差,提出了EHT的失效模型,给出了预置参数的安全误差范围。对程序生成的椭圆图像进行检测实验,证实了EHT方法的有效性。Ellipse Hough transform method is a significant research topic in the field of feature extraction.Efficient Hough transform（EHT）is proposed to improve the algorithm＇s accuracy and timeliness.This method is suitable for detecting the ellipse contour in cylindrical feature.The algorithm mainly consists of estimating other parameters＇approximation by single parameter＇s approximation.EHT＇s failure model is proposed to give the security error range of preset parameter.Synthetic images of ellipse contour are tested to demonstrate EHT＇s effectiveness.

关 键 词：HOUGH变换 特征提取 椭圆检测 圆柱 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]