基于线激光截面成像的棉花白色异性纤维检测方法

Linear Laser Detecting Method of White Foreign Fibers in Cotton Based on Sample Cross-section Imaging

摘要:

针对棉层表面散布绒毛或游离的纤维细丝，而异性纤维表面光洁无毛的特点，提出了一种基于线激光的样本截面图像获取方法，获取并显示了其微观结构在激光照射下的图像差异。在此基础上，根据激光照射下二者周围有无高亮度的星点，以及星点的分布特征，提出了一种棉花中白色异性纤维的识别算法。实验结果表明，该方法对棉花中典型白色异性纤维的识别率达到了86.9%。

Abstract:

Machine vision inspection of white foreign fibers in cotton is a research hotspot in China and aboard. Based on the fact that there are flosses and free filaments on the surface of a cotton layer，and most of the white foreign fiber surfaces are smooth, without any floss， an image acquisition method of the cross-section of the samples under the illumination of linear laser was proposed, which showed the differences of the microstructure characteristic of their surface from the reflection and transmission of the linear laser. Then, according to the features of the distribution and density of the flosses around the white spots in the images, an algorithm for identification of cotton and white foreign fibers was suggested. The experimental results indicated that, by the new method, the detecting rate of the white foreign fibers in cotton was up to 86.9%.