back to news
This innovative technology installed so far in 13 looms is based on a technology that integrates an automated quality control system based on computer vision, with the respective software supported by artificial intelligence models, which detects in real-time, at the time of production, defects in each centimetre of the fabric.
According to José Rosas, "the recognition of defects is carried out through parameterised styles for each article by adjusting the sensitivity of values reported in a previous set of photographs for each article to be produced", ensuring an optimisation of efficiency in the analysis and automatic recognition of "horizontal defects", "vertical defects" and "stains".
However, as the weaving manager explains, there is still a margin for evolution in the technology, "given that it does not respond to all types of articles, in particular articles with stripes or checked patterns, or even those that have irregular yarns in their composition, such as the so-called ‘fantasy yarns’".
Considering the promising results, Riopele is already testing the latest generation equipment to expand this quality control project based on artificial intelligence to articles in their most diverse variants, colours, types of yarn and fabric patterns.
"In the future we aim to achieve a technological solution involving artificial inspection for the vast majority of our fabrics, thus being able to optimise results and streamline inspection processes", José Rosas concluded.
This pioneering project is part of Riopele's digitalisation plan, under the Riopele Digital program, which has Industry 4.0 and Sustainability as drivers, opening up a new path in the implementation of innovative processes and technologies in the textile sector.