Portal de Periódicos da CAPES

A Ubiquitous Service-Oriented Automatic Optical Inspection Platform for Textile Industry

2022; Elsevier BV; Volume: 196; Linguagem: Inglês

10.1016/j.procs.2021.12.008

1877-0509

Dibet Garcia Gonzalez, Yusbel Chávez Castilla, Somayeh Shaharadaby, Ana Mackay, Lúcia Soares, Pedro Guimarães, Francisco Morais, Joel Puga, Filipe Meneses, Adriano Moreira, Miguel A. Machado, Telmo Adão, Luís Magalhães, Miguel Guevara,

Image and Object Detection Techniques

Within a highly competitive market context, quality standards are vital for the textile industry, in which related procedures to assess respective manufacture still mainly rely on human-based visual inspection. Thereby, factors such as ergonomics, analytical subjectivity, tiredness and error susceptibility affect the employee's performance and comfort in particular and impact the economic healthiness of each company operating in this industry, generally. In this paper, a defect detection-oriented platform for quality control in the textile industry is proposed to tackle these issues and respective impacts, combining computer vision, deep learning, geolocation and communication technologies. The system under development can integrate and improve the production ecosystem of a textile company through a properly adapted information technology setup and associated functionalities such as automatic defect detection and classification, real-time monitoring of operators, among others.