二维图像谱空间矩特征的轮胎老化视觉测量  

Visual Measurement of Tire Aging Based on Two-Dimensional Image Spectral Space Moment Features

作　　者：郭宝英[1] GUO Baoying(School of Aviation,Liming Vocational University,Quanzhou 362000,China)

机构地区：[1]黎明职业大学航空学院,福建泉州362000

出　　处：《黎明职业大学学报》2023年第2期84-93,共10页Journal of LiMing Vocational University

基　　金：福建省中青年教师教育科研项目(JAT210889)。

摘　　要：在轮胎磨损及老化的工业非接触式的无损检测中,为定量衡量轮胎老化程度,提出一种图像谱空间距定量测量轮胎老化的算法。该算法采用高斯尺度滤波对轮胎胎面及轮轱侧面采集图像进行滤波,以去掉高频分量与噪声;然后采用Canny算法得到轮胎裂纹与自身的纹理轮廓图,由二维离散傅里叶变换得到反映裂纹密度数量的频谱图;最后提取该频谱图各谱线辅助频率u、v的一阶矩相对原点距离,来定量衡量轮胎裂纹数量与密度分布特征,从而完成对轮胎老化程度的定量测量。在轮胎经营部的现场实验表明,经过小批量样本的模型识别参数训练,该方法具有90%的老化程度识别准确率,具有一定的工业应用价值。In the industrial non-destructive testing of tire wear and aging,in order to quantify the aging degree of tires,an algorithm for quantitative measurement of tire aging by image spectral space distance was proposed.Gaussian filtering was applied for high scale space images of tire tread and gully,removing high frequency noise.The Canny algorithm was applied for the contours of the tire cracks and the tire texture.A spectrum diagram reflecting the number of crack density was obtained by two-dimensional discrete Fourier transform.The distances between the first moment of the auxiliary frequency u and v of each spectrum line and the origin was extracted.The steps above produced a quantitative representation of the number and density of cracks in the space domain and completed the quantitative measurement of the tire aging degree.The experiment in the tire shop showed that the method has a 90%accuracy rate of aging degree identification after the model identification parameter training of small batch samples,and that the algorithm has a broad application prospect in industrial automation.

关 键 词：轮胎老化 二维图像谱空间矩 非极大值抑制 轮胎裂纹 视觉测量 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]