Abstract: For the purpose of high-precision measurement of yarn diameter, linear CCD was used to detect yarn diameter. Firstly, the average filtering algorithm was adopted to eliminate the influence of yarn hairiness. Secondly, the pixel coordinates of yarn edge were determined through the floating threshold method. Finally, the sub-pixel coordinates of yarn edge were obtained through the Gaussian integral fitting method. After formulating an experimental scheme, analyzing the deviation rate of the subdivided yarn diameter and the theoretical diameter, the feasibility of Gauss integral curve fitting method was verified. The experimental results showed that the average diameter deviation rate of the subdivided yarn segments was 3.19%, and the bar unevenness was reduced by 22%. The experimental results proved the accuracy of sub-pixel algorithm based on linear CCD, and provided a new method for online measurement of yarn diameter.

Key words: linear array, yarn diameter, floating threshold, Gaussian integral fitting, sub-pixel subdivision

摘要： 为实现高效高精度测量纱线直径,利用线阵CCD进行纱线直径的检测。首先通过平均滤波算法去除纱线毛羽的影响,其次利用浮动阈值法确定纱线边缘的像素坐标,最后基于高斯积分拟合的方法得到纱线边缘的亚像素坐标。通过制定实验方案,分析细分后的纱线直径与理论直径的偏差率,验证了高斯积分曲线拟合方法的可行性。结果表明:细分后纱线片段的平均直径偏差率为3.19%,条干不匀降低22%。实验结果证实了基于线阵CCD亚像素算法的准确性,为在线测量纱线直径提供了新方法。

关键词: 线性阵列, 纱线直径, 浮动阈值, 高斯积分拟合, 亚像素细分