Cutting through the complexity

One of the first ways in which artificial intelligence (AI) and machine learning are poised to have a positive and dynamic impact on the textile industry is through the widespread adoption of automated fabric inspection systems.

This is currently a trending topic in the industry, with the majority of leading mills around the world still relying on human inspection within their operations, according to Calvin Wong, CEO of Hong Kong-based AiDLab, the Laboratory for Artificial Intelligence in Design.

“In the next 10 years at least 50% of these manufacturers will be employing AI-based systems,” he said during the recent ITMAconnect Innovative Technologies webinar, which is now available to view in full at goto.itmaconnect.com. The latest camera-based vision systems for textile material inspection can eliminate defects, in addition to precisely verifying colour shades and achieve enormous savings.” He added that quality “will significantly reduce returns which are currently extremely costly to the industry and the environment”.

WiseEye

AiDLab is the developer of WiseEye, an AI-based textile material inspection technology that automatically and accurately detects fabric defects and colour shading at high speeds and in real-time inspection environments.

With the integration of machine vision algorithms and deep learning techniques, the system achieves an accuracy of over 90% and provides high-speed inspection at up to 60 metres per minute. The system also offers automatic defect labelling to meet industry requirements. It can inspect various fabric types and identify over 40 common fabric defects.

Qualitative data

Both Florian Pohlmeyer, head of digitalisation at RWTH Aachen University’s Institute for Textile Technology (ITA) and Sverker Evefall, senior application manager for Swedish vision systems developer ACG Eyetech, stressed the need for good data during the ITMAconnect webinar.

To prepare companies for the digital future, ITA opened the Innovation and Learning Center (ILC) in Aachen in 2017. In its Model Factory 4.0, digital solutions for production are being developed and via workshops and seminars, basic knowledge on the topic of Industry 4.0, condition monitoring, sensor technology and automation is provided to companies.

ACG EyeTech exploits 2D, 3D and deep learning techniques for inspection, measuring and robotic guidance. At ITMA 2023 in Milan, sister companies ACG Kinna and ACG Nowo demonstrated a complete process for robotic pillow filling with the ability to fill and finish some 3,840 pillows in an eight-hour shift. ACG EyeTech is now adding machine vision systems to these systems for fault-free and seamless fibre handling, filling, sewing and packing.

Evolutionary process

“Without qualitative data the most advanced machines will not achieve what they are fully capable of,” Pohlmeyer said. “The textile industry has a complex value chain and many processes within it, and data which is collected at every stage is often unstructured or distributed across different software systems, databases and machines. Each has its own format and it’s very important that they are standardised to enable systems to communicate.”

Evefall added that any vision system has to be shown what to do, like teaching a child.

“Don’t underestimate the time involved, because the algorithm is not an intelligent being that can just find defects, you actually have to teach it what to look for,” he said. “Many manufacturers have the perception that AI can do everything but implementing it is an evolutionary process which takes time. The system has to be shown vast amounts of data.”

People power

Both agreed that good people were critical to success, and panel chair Dr Andre West, associate professor and director of the Zeis Textiles Extension (ZTE) at the Wilson College of Textiles at North Carolina State University, added that collaborating with AI experts was a key to getting knowledge into a company.

“The training of people who are working with the AI in production is paramount because this is not just about technology, but the development of skill sets,” Pohlmeyer said. “Companies need people in many fields such as data analytics and cyber security in order to combine these new tools with textile domain knowledge.”

Other areas where AI and machine learning will have an impact were said to include the use of robots for materials handling and transportation inside mills, exploiting generative AI to create new fabric patterns, colour matching and the optimisation of dye formulations.

Most immediately, the sorting of post-consumer textile waste for successful recycling will be a priority as the European Union introduces a ban on it going to landfill or incineration at the start of 2025.

Successfully sorting post-consumer textile waste by type, colour and structure is simply not possible without AI and industrial cameras, the panellists agreed.

“Industrialists must keep their eye on these developments every single day because it’s moving so fast in the manufacturing sector,” Dr West said in conclusion. “I think that at the next ITMA in Hannover three years from now we’re going to see massive developments.”

A recording of the webinar is available on ITMAconnect on-demand for subscribers. If you are not a subscriber, register for an ITMAconnect visitor account now for just €25 (plus applicable VAT).

Watch On-demand Now Subscribe to ITMAconnect