Single vision based identification of yarn hairiness using adaptive threshold and image enhancement method. (November 2018)

• Record Type: Journal Article
• Authors: Wang, Wendi; Xin, Binjie; Deng, Na; Li, Jiaping; Liu, Ningjuan
• Is Part Of: Measurement. Volume 128(2018)
• Page Start: 220
• Page End: 230
• Publication Date: 2018-11
• Subjects: Yarn hairiness -- Adaptive grayscale enhancement -- Regional threshold segmentation -- Image analysis -- Single-view based; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2018.06.029
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: A single-view based yarn hairiness evaluation system is developed to digitalize the yarn image sequence. Adaptive grayscale enhancement algorithm is proposed to improve the contrast between yarn section and background. Linear region threshold segmentation algorithm avoids the effects of local background variance. Abstract: In order to measure the yarn parameter information more accurately with the image analysis method, an adaptive grayscale enhancement and linear region threshold segmentation algorithm are proposed systematically. The grayscale contrast of the yarn and background is enhanced to avoid the poor effect of single threshold based image segmentation method, thereby improving the recognition and measurement accuracy of the yarn hairiness. Using the self-developed image acquisition system to acquire the image sequence of yarn samples, the accuracy and effectiveness of the image analysis algorithm were validated accordingly. Experimental results show that the proposed two algorithms can significantly reduce the information loss of yarn image and good robustness could be achieved. The length and number of yarn hairiness detected by the image methods are highly correlated with those of the visual inspection method, and the CV of yarn evenness is also consistent with the traditional method.