Evaluation of a Digital Image-signal Approach on the Automatic Measurement of Cotton Yarn Snarls

Abstract

Residual torque or twist liveliness of a single-spun yarn, among other factors, is the most prominent and fundamental factor contributing to the spirality of plain-knitted fabrics and the skewness of denim fabrics made from the yarn. In textile processing, yarn snarling caused by the twist liveliness is also considered a serious problem that leads to yarn breakage, the deterioration of yarn properties, and equipment malfunction [1]. Over recent decades, various testing apparatus and procedures have been developed for the measurement of twist liveliness. Generally, all of these methods are based on manual operation and can be classified into the three categories of direct, semi-direct, and indirect methods [2]. 1 Recently, computer vision has attracted the increasing attention of researchers in textiles and apparel. Attempts have been made to replace the conventional observation method with computer vision to resolve the limitation of human vision. The notable applications include pilling assessment [3][4][5][6], fabric texture analysis [7-10], fabric defects identification [11][12][13][14], and fiber morphological measurement [15,16]. In our earlier work [1], the methodology of a digital image-signal approach was proposed for the automatic measurement of yarn wet snarls from an image of snarled yarn samples captured in a water bath. After the image acquisition and pre-processing, the two digital signal-processing methods of Fast Fourier Transform (FFT) and Adaptive-orientated Orthogonal-projecAbstract This paper is the second part of a series reporting the recent development of a computerized method for the automatic measurement and recognition of yarn wet snarls from an image of snarled yarn samples captured in a water bath. In our earlier work, a digital image-signal approach for fully computerized yarn-snarl measurement was developed and the effects of various influencing factors on the recognition algorithms were numerically examined. In this paper, the feasibility and accuracy of the fully computerized method on the measurement of actual yarn wet snarls are evaluated through laboratory experiments. One hundred percent cotton ring spun single yarns of 7, 10, 16, and 20 Ne are prepared and used for the evaluation. In addition to the number of snarl turns per unit length, the snarl height and width of the yarn samples are also objectively measured by using the computerized method. The measurement results obtained by the computerized method are analyzed and compared with those measured manually by using a twist tester and an interactive computer method.