Abstract: In order to achieve real-time positioning of the workpiece in the industrial processes, a real-time location and matching method based on linear geometric hashing primitive is proposed. In the offline process, linear primitives are extracted from the image, one of the two linear primitives is selected to construct basement, quantify the remaining primitives and establish the geometric Hashing table; In the online process, a set of linear primitives is selected to be basement, quantify the remaining primitives, for the coordinates in the geometric Hashing table query corresponding substrate and vote them, in this way real-time locating and matching are achieved. Through the experimental analysis, the method for the positioning and matching of the workpiece, is good real-time, high accuracy.

Key words: machine vision, geometric primitives, geometric hashing, positioning and matching

摘要： 为实现工业过程中对工件的实时定位，提出一种基于直线基元的几何哈希法实时定位与匹配方法。离线学习模板过程，提取图像中的直线基元，选择其中两个直线基元构建基底，量化剩余基元并建立几何哈希表；在线实测图像中，选择一组直线基元构成基底，量化剩余基元，通过坐标在几何哈希表中查询对应的基底并投票，来实现实时定位与匹配。实验证明，该方法对于工件的定位与匹配，实时性好、准确性高。

关键词: 机器视觉, 几何基元, 几何哈希法, 匹配定位