Determining material composition using intelligent software and the NIRscan Nano EVM

If you conduct a web search with the phrase “How to find out what a fabric is made from,” then you're likely to find references to “burn tests.” Burn tests involve taking a small sample of the fabric and holding it over a naked flame to observe whether it shrinks, melts or burns and noting the resulting smell.


Now, there’s a much simpler and more accurate method to determine fabric and textile composition using the TI DLP® NIRscan™ Nano evaluation module (EVM) and the Sagitto system. Sagitto is dedicated to helping companies measure like never before by combining miniature near-infrared sensors with machine learning models. Each type of fabric has a unique near-infrared fingerprint, depending on its composition. Garments often comprise different types of fiber, and that precise composition mixture is important throughout a garment’s life.

Mean NIR absorbance of textiles

Figure 1: Near infrared absorbance spectra for textiles with different fiber content


Many countries require the fiber composition of textiles to be clearly labeled. Sometimes these labels are misleading. For example, in the image below we see a set of dish towels labelled as 100% cotton but – when tested by Sagitto – shown to be a mixture of 67% cotton and 33% polyester.


 Figure 2: Dish towels shown to be 67% cotton and 33% polyester, not 100% cotton as labeled

But why does fiber composition matter? It is estimated that 80 billion pieces of clothing are produced each year, of which 75% will end up in landfill or be incinerated. Consumers are pressuring large clothing retailers to find alternative ways of dealing with waste from high-turnover fashion retailing. Government entities are also starting to introduce regulations to encourage “the circular economy” and divert clothing from garbage piles.

Acrylic and polyester clothing place an especially high load on the environment, with each washing cycle releasing hundreds of thousands of microfibers into local wastewater treatment plants. As much as 40% of these microfibers may end up in rivers, lakes and oceans.

Figure 3: Textile waste is becoming a major problem worldwide

Thus, there is considerable pressure to develop new chemical recycling techniques for textiles. For example, these recycling techniques can reduce polyester and cotton garments to their constituent chemical components – cellulosic fibers and polyester monomers and oligomers. But first, recyclers using chemical recycling will need to accurately sort feedstock by fiber composition.

Traditionally, human workers have sorted waste textiles by feel and by sight, determining textile composition as they pick up each individual garment. Unfortunately, determining the exact composition of a textiles with mixtures of fibers to the level of accuracy needed by modern chemical recycling techniques is impossible for humans.

The incorporation of the TI DLP NIRscan Nano into a robotic hand, combined with sophisticated machine learning, makes it possible to develop accurate robotic sorting systems for chemical recycling plants.

Sagitto combines the DLP NIRscan Nano with cloud-based artificial intelligence. With Sagitto, you don't need to employ your own data scientists. You do not even need to gather your own data to train machine learning models. Sagitto removes the barriers of equipment cost, skills and data so that a whole new class of manufacturers and producers can optimize their production processes using the DLP NIRscan Nano EVM.

Utilizing Sagitto artifical intelligence software with a DLP NIRscan Nano EVM, you can experiment with unique demonstration models for fabric composition. Register on the Sagitto website and request access to the Sagitto demonstration account, which will enable an initial 50 predictions for free with the DLP NIRscan Nano EVM.

Additional resources