摘要
为了能有效解决高危路段铁轨检测任务,必须准确、快速提取出数字图片中的铁轨。本文提出了一种基于多约束的Hough变换(HT)的直线提取算法。该算法采用局部算子法和全局直线拟合法提取铁轨,结合传统HT的思想,对基本算子进行约束,得到可能的铁轨点并标记方向。该算法计算时间短,具有较高的准确性。本文综合运用Hough变换在检测直线上的优越性和微分学的简单准确,来识别局部弯曲的铁轨,并在铁轨识别上进行仿真,取得了良好的效果。
In order to settle the problem of the rail detection in high risk railway section, the rail in the digital image must be extracted accurately and quickly. This paper brings forward an algorithm for straight-line extraction based on multipie constraints Hough transform technology （HT）. This algorithm extracts the rail by applying the local operator technology and overall straight line fitting technology. And then the possible rail points can be obtained and the direction can be marked by using the traditional HT imagine to restrict the basic operators. This algorithm is fast on calculation, and has higher accuracy. This paper applied the superiority of Hough transform technology on the straight line detection and accurate and simple of the differential calculus, to identify local bending of the rail and simulated on the rail. This algorithm can achieve good results.
出处
《新技术新工艺》
2009年第8期50-53,共4页
New Technology & New Process
