基于小波变换的机织物高密度检测

High-Density Detection of Woven Fabric Based on Wavelet Transform

针对目前高密度机织物密度自动检测算法测量精度较低的现象,提出在三原组织织物密度自动检测过程中,利用小波变换的分解与重构特性分析预处理后的织物图像,对分解的子图像进行二值化、平滑处理来提取纱线周期性特征参数,统计二值图像中黑白循环更替次数来确定织物经纬密度。为提高密度检测精度,采用分区域平滑减小了斜纹纹路的影响;通过缎纹织物图像反面获取纬纱信息消除缎纹组织织物浮长的影响;运用形态学滤波去除因高密度织物上纱线毛羽导致平滑过程中出现的细条块。实验结果表明:该方法检测误约为1.00%,测量精度较高,具有一定的实用参考价值。

In view of the phenomenon that low measurement accuracy of automatic detection for high density fabric at present, a method was put forward in the process of Sanyuan group fabric density automatic detection, which used wavelet transform to analyze fabric pretreatment image taking its decomposition and reconstruction characteristics, and yarn periodic characteristic parameters were extracted by binarization and smoothing of the decomposed sub images from wavelet, then the thread count were determined by thresholding the sub-images. In order to improve the accuracy of the density test, smooth areas ways were used to reduce the influence of diagonal lines in the process of twill fabric image smoothing; eliminated floating long satin weave fabric by satin fabric image reversed weft information; morphological filtering method was adopted to solve the smoothing block strip problem caused by yarn hairiness. The experimental results showed the measurement error of the method was around 1%, and it had high accuracy and had a certain practical reference value.