应用频域分析与距离匹配函数的织物纹理周期测量  

Periodicity Measurement for Fabric Texture by Using Frequency Domain Analysis and Distance Matching Function

作　　者：周建[1,2] 王静安[1] 潘如如[1,2] 高卫东[1,2] 

机构地区：[1]江南大学纺织服装学院,214122 [2]江南大学生态纺织教育部重点实验室,江苏无锡214122

出　　处：《东华大学学报（自然科学版）》2017年第5期629-633,共5页Journal of Donghua University(Natural Science)

基　　金：国家自然科学基金资助项目(61501209);江苏省博士后科研计划资助项目(1601017A);江苏高校优势学科建设工程资助项目(苏政办发2011-137号)

摘　　要：通过将频域分析与距离匹配函数(DMF)相结合,实现了机织物图像纹理周期的自动测量.首先按行和列方式分别将织物图像展开为行向量和列向量,然后应用一维快速傅里叶变换分别将所提取的行向量和列向量变换到频率域进行分析,并在此基础上提取两个最大峰值所对应的周期值,最后通过计算潜在周期的累加DMF实现织物纹理周期的准确测量.试验结果表明,该方法能有效地测量织物纹理周期,相比于累加DMF方法,其稳定性更好且计算效率高.Frequency domain analysis and distance matching function （DMF） were used to measure periodicity of fabric texture in automated manner. The proposed method is first to transform fabric image into two one-dimension vectors along row and column. Then, the one-dimension Fast Fourier Transform （FFT） is used to transform the extracted vectors into spectral domain, in which the periodicities of the first two largest amplitude are extracted. Lastly, the accumulative DMF for the two potential periodicities is calculated to determine final periodicity. The experimental results demonstrate that the proposed method can measure fabric periodicity effectively and accurately, and it is also more stable and efficient than the accumulative DMF method.

关 键 词：织物纹理 周期自动测量 傅里叶分析 距离匹配函数 频域分析 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]