Early-Stage Defect Detection in Manufacturing: Smartex Streamlining the Supply Chain

Though there are various causes of defects in the textile manufacturing process, the majority occur during the knitting stage. Traditionally, detecting these defects has relied on manual inspection, making early-stage identification challenging. Addressing this issue, Portugal-originated Smartex was founded to provide a solution using advanced AI and machine learning algorithms to detect defects during the manufacturing process.

The Genesis and Journey of Smartex: Pioneering Defect Detection in Textile Manufacturing

Gilberto Loureiro, one of the founders of Smartex, had a consciousness of textile defects and their associated costs long before the inception of Smartex. Hailing from Portugal, Gilberto, the son of parents employed in a textile factory, spent his high school summer vacations working as an inspector of textile fabrics. Gilberto, the first in his family to pursue higher education, met Antonio Rocha, and Paulo Ribeiro and together they laid the foundation for Smartex.

A turning point came in 2015 when Gilberto and Antonio's mutual friend, Carlos Lei Santos, launched the Silicon Valley startup "UpLink." Inspired by their friend's success, they envisioned launching their own startup and identified the textile industry as an ideal target. The textile industry relied heavily on advanced machinery, yet its processes were prone to errors in manual tasks, becoming a significant source of defects.

Drawing from Gilberto's experience as a textile inspector, he understood that reducing defective products could bring substantial benefits to the industry. Especially for circular knitting machines, he envisioned a system that would detect defects, making them visible to inspectors before becoming apparent. Automatic production halts for adjustments would not only make these defects visible but also reduce energy, water, dye, and yarn wastage in factories.

At that moment, the three co-founders, Gilberto, Antonio, and Paolo, each with expertise in the textile industry, electronic engineering, and software, pursued the concept of Smartex while attending graduate school, dedicating nights and weekends to their vision.

In early 2018, they discovered a breakthrough. At a textile factory owned by Gilberto's friend, the team modified existing circular knitting machines, equipping them with high-resolution cameras and lights. Captured images of the fabric were sent to machine learning algorithms trained to recognize visual clues related to defects. When defects, such as the elusive 'Lycra Dash' appearing as fine lines or streaks on the fabric's surface, were detected in the images, the knitting machine would automatically stop. This mechanism aimed to halt production upon detecting defects not easily visible to the human eye.

Encouraged by the success of the system, they founded Smartex with support from the HAX Accelerator Program in November 2018, participating in the program in Shenzhen to further expand the reach of Smartex's technology.