Abstract: Based on the similarity theory, the characteristics of tire pattern structures were analyzed. Using image processing method such as gray histogram, aHash algorithm and co-occurrence matrix method, a 2D similarity detection system for tire patterns was built in MATLAB software. The similarities of the existing tire patterns were detected by the system, and the similarities of the specified patterns with local change and the original patterns were detected. Results show the stability and the reliability of the system. It may provide a basis for similarity quantization of different tire patterns.

Key words: tire pattern, similarity theory, aHash algorithm, co-occurrence matrix method

摘要： 基于相似性理论分析了轮胎花纹结构的特点。运用灰度直方图、均值哈希算法、共生矩阵法等图像处理方法，在MATLAB软件中搭建了一个轮胎花纹2D相似度检测系统。利用该系统对现有的轮胎花纹进行了相似度检测与分析。将特定花纹的局部改变与原有花纹进行相似度检测，结果验证了系统的稳定性和可靠性。该系统可为不同花纹之间相似度的量化提供依据。

关键词: 轮胎花纹, 相似性理论, 均值哈希算法, 共生矩阵法