基于数字图像处理的纱线毛羽检测

Detection of yarn hairiness based on digital image processing

为更准确地检测纱线毛羽长度及其根数,在结合视频显微镜和图像处理技术的基础上,提出一种新的毛羽检测方法。采用MOTIC SME-140视频显微镜采集纱线图像,经过灰度变换、图像分割、形态学开运算、图像细化处理,得到完整的纱线条干图像和细化后的毛羽图像,以纱线条干边缘为基准线,对毛羽分割点进行判断,最后得出不同长度的毛羽根数。图像法检测结果表明,纱线片段的毛羽根数值较为稳定,检测结果与目测图像计数的结果非常接近,因此,可认为所提出的毛羽检测方法较现有的光电检测方法更为准确、可靠。

In order to detect the length and root number of yarn hairiness more accurately, a new hairiness detection method was proposed based on video microscope in combination with image processing techniques. The yarn images were captured with a MOTIC SME-140 video microscope and then were processed with gray scale conversion, image segmentation, morphology opening and image thinning sequentially. The images of yarn core and hairiness thinning were obtained. And then with the yarn core edge as the measurement baseline to segment hairiness and the hairiness cut-point was judged, thus different lengths of hairiness root number were obtained directly. The result shows that the value of root number of every piece is more stable, and the results of this detection are very close to the visual counting values. Therefore, the present method for hairiness detection is more accurate and reliable than the photoelectric one.