摘要
该文提出了一种基于改进的多分辨率Hough变换的直线检测方法。对多分辨率图像由粗到细地循环处理,首先利用全局累加器序列对最小的图像做Hough变换得直线参数,利用该直线参数估计下一循环图像中的直线参数;建立了一种简单对应关系,将得到的直线参数的估计范围转换成图像空间范围;然后在图像空间范围内利用局部累加器对较大图像做Hough变换检测直线。最后还给出了线段端点检测、长度计算及连续性考察的方法。实验表明,该方法可以加快直线检测速度、节省存储空间,端点检测方法具有一定的鲁棒性。
A new method of line detection based on improved arithmetic of Muhiresolution Hough transforms is proposed. Multiresolution images reduce the number of the feature pixels which decides the computer power, The Hough transform is first applied to the smallest image using a large accumulator array and obtains parameters of lines. Computing the parameter ranges of the lines based on the lines parameter of previous iteration, then applying the Hough transform to the larger image using smaller accumulator arrays in the parameter ranges former mentioned. The methods of line segments＇end points detection and length computation and connectivity denotation are presented. The results of the experiment indicate that The line detection method can make the detection faster and reduce the storage capacity obviously. The end point detection method is robust.
出处
《计算机仿真》
CSCD
2006年第9期76-79,共4页
Computer Simulation
关键词
模式识别
多分辨率
直线检测
Pattern recognition
Multi - resolution
Line detection
