Akademik Məlumat İdarəetmə Sistemi
Izvestiya Vysshikh Uchebnykh Zavedenii, Seriya Teknologiya Tekstil'noi Promyshlennosti, vol.2024, no.1, pp.151-157, 2024 (Scopus)
It is shown that the known indicators used to assess the twist are not suitable for self-twisted yarn. This is due to the peculiarities of its structure. The presence in the yarn of zones with left and right twist, separated by a section of yarn that does not have a twist, requires the use of specific indicators. The following single indicators are proposed: twist unevenness, average length of transition sections, length of twist zones and maximum length of zones with zero twist, relative difference of twists in the zones. It is shown that the existing methods for determining the indicators of yarn twist are ineffective for determining the newly proposed indicators. It is possible to increase the efficiency of their determination by using computer vision methods. The substantiation of the computer analysis of the image has been carried out. An analysis of the structure of the image processing algorithm was carried out. The requirements for the original image and methods for obtaining it are formulated. The main parameters of the original image are resolution, background width, displayed yarn length.