基于小波变换的散斑纹理表面粗糙度信息提取  

Information extraction of surface roughness measurement based on wavelet transform of speckle pattern texture

作　　者：杨蕾[1] 卢荣胜[1] 雷丽巧[1] 

机构地区：[1]合肥工业大学仪器科学与光电工程学院,合肥230009

出　　处：《电子测量与仪器学报》2012年第12期1091-1095,共5页Journal of Electronic Measurement and Instrumentation

基　　金：国家自然科学基金(50875074)资助项目

摘　　要：基于计算机纹理分析法,研究了一种新的非接触式表面粗糙度测量方法。通过获取磨削金属表面的数字激光散斑纹理图像,采用小波分析对其进行变换,首先选取了小波基和分解级数,然后计算出变换后各子图的系数矩阵,并构造出新的系数共生矩阵,提取出该系数共生矩阵的均值、方差、能量和熵4个纹理特征参数,最后建立纹理特征参数与表面粗糙度Ra间的关系。研究结果表明,上述4个特征参数与表面粗糙度Ra均有良好的单调关系。若对同种材料和相同加工方式的物体进行标定,可以利用该方法实现表面粗糙度的测量。Based on computer texture analysis method,a new non-contact surface roughness measurement technique is presented.The digital speckle pattern texture images of grinding metallic surface are taken and decomposed by wavelet transform method.Wavelet base and decomposing degree are chosen at first.Coefficient matrices of sub-patterns are cal-culated and a new coefficient co-occurrence matrix is configured.The texture features-mean,variance,energy,and entropy of coefficient co-occurrence matrix are extracted.The relationships between texture features and surface roughness Ra are investigated.The research results show surface roughness has a good monotonic relationship with the four texture features.This technique can be used in surface roughness measurement with calibrating the same material and same machining method beforehand.

关 键 词：数字散斑 表面粗糙度 纹理分析 小波变换 

分 类 号：TG84[金属学及工艺—公差测量技术]