Akademik Məlumat İdarəetmə Sistemi
Izvestiya Vysshikh Uchebnykh Zavedenii, Seriya Teknologiya Tekstil'noi Promyshlennosti, vol.2025, no.2, pp.253-259, 2025 (Scopus)
The questions of obtaining initial data on the shape of cross-winding textile packages are considered. It has been established that geometric properties are characterized by a number of indicators, which in its turn can be subdivided into nominal and deviations from nominal ones. Deviations of these parameters from the specified values lead to a loss of pack quality. The winding conditions become worse. The packages are not triggered at the same time during warping, which leads to the necessity of using additional winding equipment for the rewinding of residues and more frequent changing of packages with decreasing winding density and volume. To obtain primary data on the shape of the package, it is proposed to use the shadow projection method in combination with automated pattern recognition tools. To analyze the shape of the winding in the process of its formation, it is proposed to obtain primary images in the form of video clips. Primary data is processed using the original software. The paper describes the structure of the algorithms underlying the present software. The developed processing algorithm involves splitting the video image into separate frames. It also involves conversion to a black-and-white palette, filtering, reading the edge coordinates of the bobbin image, scaling, and building a 3D image of the package. An algorithm has been created that allows the online processing of data on the shape of packages. This builds 3D models of real packages in the process of their winding. Comparison of the obtained models with the exemplary ones allows for identifying the causes of defects, as well as violations of the technological process that cause them. The information obtained in this way allows for optimizing the process of cross-winding package production in textile production.