基于物理显微镜法的纤维含量AI检测与传统检测方式对比

Comparison Between AI Fiber Content Detection and Conventional Detection Methods Based on Physical Microscope Method

文章针对传统纤维含量显微镜测定法存在的弊端,提出了AI横截面自动检测仪的原理,即通过图像采集、语义分割等方法对单根纤维横截面进行分割生成蒙版图,根据蒙版图计算每种纤维的平均横截面面积,进而分析其含量;通过选择代表性试样开展实验,结果显示,AI横截面自动测试仪测得的纤维含量与传统人工测试相比最大误差在3%以内,测试时间由人工测试的50 min缩短到仪器测试的5 min。AI检测方法结果稳定,消除了人为误差,具有明显的优越性,可用于纺织品纤维含量的检测。

Aiming at the drawbacks of the traditional fiber content microscopy method,the paper proposes the principle of the AI cross-section automatic detector,that is,generating a mask map of a single fiber crosssection by image acquisition and semantic segmentation.On the basis of that,it can calculate the average cross-section of the fiber,and then analyze its content;by selecting representative samples to carry out the experiment,the results showed that compared with the traditional manual test,the maximum error of the fiber content measured by the AI cross-section automatic tester was within 3%,and the time was shortened from 50 min to 5 min.The AI detection method has stable results,eliminates human error,has obvious advantages,and can be used for the detection of textile fiber content.