检测不规则图形的改进广义Hough变换

Improved Generalized Hough Transform for the Recognition of Irregular Objects

广义Hough变换作为一种检测不规则图形的有效方法,具有抗噪、不怕遮挡等优点。由于R表的计算对断点和变形非常敏感,影响了广义Hough变换抗断点和变形的特性。针对这一问题,提出了一种计算图形上每个点的法线方向的新方法。利用法线方向作为R表索引项,对点进行分类,使得广义Hough变换不仅具有抗噪、不怕遮挡的优点,还具有抗断点、受变形影响小的特点。实验结果证明,该方法在检测不规则图形时具有很好的抗干扰能力,并在鞋样设计CAD系统中有效解决了鞋样匹配的问题。

The generalized Hough transform for the recognition of irregular objects is a powerful method, since it can detect overlapping objects under severe noise. The conventional computation method of the R-table＇s index is sensitive to discontinuity and distortion. So the GHT＇s robustness is degraded a lot. To solve the problem, this paper proposes a new procedure for calculation of the normal direction of each point. Characterizing the edge points by their normal directions can improve the GHT＇s robustness to discontinuity and distortion. Experimental results demonstrates that the improved method does work well for objects recognition. And the shoe sample could be detected accurately despite of noise, discontinuity or geometric distortion in the system of shoe sample CAD.