Study On Yarn Hairiness Detection Technology Based On Single Angle Imaging

Yarn hairiness is an important indicatorin the production and processing of textiles,the length and number of hairiness not only affect the production process but also determinethe quality of the product.At present,there are many methods to detect hairiness,such as photoelectric method and manual blackboard visual method,which is easy to receive the influence of test environment,there are some limitations;artificial visual method can not be detected quantitatively,and subjectivity is too strong.However,the method based on digital image processing can serve the shortcomings of the above two methods,it is not subjected to objective and subjective conditions,it can be obtained scientific and accurate results.In this paper,a digital image processing method about yarn hairiness detection is proposed based on the traditional manual visual method.A high-speed dynamic image acquisition system is designedfor a sequence of yarn images collected by industrial cameras in this paper,which included LED white light source,yarn drive module,imaging module and image acquisition module.The abtained image sequences are processed and analyzed in order to extract the characteristic parameters of hairiness,therefore the yarn hairinessis digitally tested and characterized;According to the image sequence acquired by the scanner,the calculation of the yarn unevenness is mainly discussed,and the algorithm flow is as follows:1)through the gray-scale processing,the background separation of uniform gray yarn image;2)the gray uniform image of the yarn through the adaptive median filter processing,getting the noise-free yarn image;3)and morphological processing to obtain the main image of the yarn,finally,through the feature extraction algorithm,statistical yarn diameter,the uneven dryness CV.The images in this article are derived from two approaches,industrial camera and scanner acquisition.The algorithm flow is as follows: 1)Through the gray-scale processing,tilt correction,gray-scale correction,image denoising,image sharpening to get horizontal,gray-scale uniform andnoise-free images;2)It can be obtained the yarn hairiness trunk image sequence through making threshold segmentation and edge detection on denoised images,extracting the yarn hairiness characteristic parameters through multi-region contour tracing algorithm;3)Extracting the yarn trunk images by morphological processing,the detection of yarn unevenness of the stem.In order to improve the efficiency of the algorithm and the accuracy of experimental results,the experimental results are compared with the traditional image processing algorithms and other different algorithms to select the suitable experimental method for yarn hairiness extraction.The experimental results come from yarn image forming apparatus in this papershow that the yarn parameters are very similar to results of other traditional detection methods.The method has advantages of quickness,accuracy and simplicity.It can reduce the work intensity,improve the efficiency of hairiness detection and has an important practical value.