基于Beamlet的图像线特征提取方法  

Beamlet-based Method of Line Feature Extraction from Images

作　　者：杨玲香[1] 李军侠[2] 水鹏朗[2] 

机构地区：[1]西安电子科技大学理学院数学科学系,西安710071 [2]西安电子科技大学雷达信号处理国家重点实验室,西安710071

出　　处：《计算机工程》2008年第9期203-204,229,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(60472086);教育部高校青年教师专项基金资助项目

摘　　要：提出一种基于Beamlet变换图像线特征提取方法。用基于方向窗的方向滤波突显图像中不同方向的线特征,通过阈值处理将灰度图像转换为一组二值图像,利用逻辑或运算融合成一幅图像,用Beamlet变换从融合图像中提取图像的线特征和曲线特征。实验结果表明,在信噪比很低的情况下该方法能够得到满意的结果。This paper proposes a Beamlet-based method for line feature extraction from noisy images. The directional filtering using a set of directional windows is used to highlight the line and curve features along different directions. The thresholding process is used to transfer the obtained gray-level images to a set of binary images and these binary images are fused into a single binary image by the logical ＇or＇ operator. The Beamlet transform is applied to the fused binary image to extract the line and curve features in the original noisy image. Experiments show that the proposed algorithm can achieve satisfactory performance in the cases of low signal-to-noise ratios.

关 键 词：BEAMLET变换 特征提取 方向窗 方向滤波 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]