形态骨架匹配算法及其在交通标志识别中的应用

交通标志是人为的、具有特殊形状的符号，就其结构可分为外层轮廓和内核两部分。应用计算机视觉可以对交通标志进行自动识别．其关健在于如何检测交通标志和识别它们的内核形状。本文应用数学形态学方法研究交通标志内核形状的形态特征，分析了内核形状的形态骨架，提出和讨论了内核形态骨架的相关函数和距离函数，并就交通标志的有效识别提供了理论分析和实验结果。

Matching Algorithms of Morphological Skeletons and Their Applications to Traffic Sign Recognition

Traffic signs on roads are of artificial and specific shape,and always contain meaningful inner shape within outer contours. The key to recognize them by computer vision is bow to detect them and identify their inner signs. In this Paper, mathematical morphology is used to research shape features of the inner shape, and their morphological skeletons are experimentally analyzed. A maximum normalized correlation and a minimum weighted distance are proposed and discussed. Experimental results show that the two proposed matching algorithms for the morphological skeletons serve as proper measures to identify the traffic signs.