基于像素组合的快速哈夫变换

直线参数检测是图像分析和计算机视觉中一个基本问题,哈夫变换能很好地解决这个问题,然而哈夫变换的巨大计算量使它难于实时应用。因为一个图像点对应参数空间的一条曲线,图像点与量化后的参数区间是一对多的关系,造成了标准哈夫变换的计算复杂性。而两点确定一条直线,两个图像点对应一个参数点,利用这个原理能大幅度减少哈夫变换的计算量。

A Kind of Fast Hough Transform Used for Line Detection

Hough Transform is a widely used algorithm in computer vision and pattern recognition.However,the tremendous computation required by Hough Transform impedes it used in real-time applications.The standard Hough Transform used only the information of a single non-zero pixel,which causes the huge computation.The relationship between two pixels is concerned in the presented scheme.A line parameter is obtained from the coordinates of two pixels,so the computation is deceased distinctly.The testing data shows that the Fast Algorithm is faster 10-100 times than the standard one in a same platform,while the precision is kept.