商空间理论的Hough变换直线提取

Hough Transformation Based on Quotient Space Theory for Line Extraction

Hough变换以其对图像噪声的鲁棒性,一直是参数化图形检测提取的经典工具,尤其是在直线提取方面,但传统的Hough变换存在的一些缺点,限制了Hough变换的应用场合,粒度计算为这些缺点的改进提供思路,文中在全面分析传统Hough变换及其存在缺点的基础上,结合商空间的粒度计算理论模型,提出一种基于商空间理论的Hough变换直线提取算法,实验证明,该算法在算法复杂度、实用性和准确度等各方面都有一定的改善。

Hough Transformation （HT） is famous for its robustness to image noise, and becomes a classical tool for parameterized shape extraction, especially for line extraction. However, some defects in conventional HT limit its application. But, granular computing provides the ways for the improvement of these defects. With complete analysis of conventional HT and its defects, the paper combined HT and quotient space theory, which is a granular computing theory model, presented a line extraction computing method for HT based on quotient space theory. Experiments show that this algorithm made an improvement in algorithm complexity, usefulness and accuracy in practice.