基于贝叶斯阈值的纱线毛羽检测方法研究

Study on Yarn Hairiness Measurement Method Based on Bayesian Threshold

研究基于贝叶斯阈值的纱线毛羽检测方法。首先利用图像增强以及保边递归滤波对纱线图像进行预处理,以增强纱线毛羽与背景之间的对比度;其次,利用贝叶斯阈值对预处理后的图像进行像素级分割,并去除条干,获得纱线毛羽;最后对获得的毛羽进行细化,并利用像素法对细化后的毛羽进行统计分析,计算出纱线毛羽长度、毛羽根数、毛羽面积指数及毛羽长度指数等指标。与等距线法及YG172A型纱线毛羽测试仪的检测结果相比较,本研究方法能够精确计算出纱线毛羽的各项指标。

The yarn hairiness measurement method based on Bayesian threshold was studied.Firstly,yarn image was pre-treated by image strengthening and edge preserving&recursive filtering to enhance the contrast ratio between yarn hairiness and background.Then,pixel-level division was performed on the image treated with Bayesian threshold.Evenness was excluded and yarn hairiness was obtained.In the end,the hairiness obtained was refined.Moreover,the pixel method was adopted for static analysis of the hairiness after refined.The indexes like yarn hairiness length,hairiness number,hairiness area index,hairiness length index and so on were calculated.Compared with the test results of equidistant line method and YG172A yarn hairiness tester,the method in the study could calculate each index of yarn hairiness accurately.