【作者】 韩秋蕾； 朱明； 姚志军；

【Author】 Han Qiulei; Zhu Ming; Yao Zhijun (Image Processing Laboratory, Changchun Institute of Optics,Fine Mechanics and Physics, Chinese Academy of Sciences,Changchun 130033,China)

【机构】 中国科学院长春光学精密机械与物理研究所图像处理室；

【摘要】 为了从复杂背景中提取欲识别的目标,提出一种基于改进Hough变换的图像线段特征提取方法。针对传统Hough变换的局限性,在经典算法的基础上增加了全局阈值与局部阈值结合去除虚假尖峰,利用动态分组原则确定线段的端点,融合端点邻近的线段等改进算法。实验表明,该方法不但可以识别简单背景中的规则几何图形,还可以有效地提取复杂背景中真实目标的线段特征。更多还原

【Abstract】 In order to detect tragets in complex environment, a line segment feature extraction method is proposed. The algorithm is based on the analysis of limitations of traditional Hough transform. It consists of eliminating spurious peaks according to global and local thresholds, fixing on endpoints according to dynamic clustering rule, merging line segments with near endpoints. Experiment results show the approach not only can recognize regular geometric form in simple environment but also can extract the line segment features of real targets in complex environment.更多还原