The need for an objective test system for identifying fine natural fibres such as cashmere and vicuna remains a goal to be sought by all producers of quality fashion garments, for both quality control and legal defence of labelling. A common contaminant of “100%” cashmere is fine lambswool, but because both of these classes of fibre are animal in origin and therefore mainly keratin, the chemical differences are difficult to detect. Although there have been claims for the potential of chemical methods such as electrophoresis and DNA cloning, the inherent complexities involved in these approaches would appear to make them unlikely components of a practicable laboratory test for natural fibres materials and their blends.
FTIR spectroscopy provides a powerful but relatively rapid chemical analysis method, and with modern technology and ATR sampling textile fibres and fabrics can supply usable spectral data. With regard to the analysis, there are a number of artificial intelligence tools available to textile scientists. One of these, is the neural network, which emulates brain learning to develop internal functions which can recognise and classify patterns of data, such as spectral values. Another tool, is fuzzy logic, which sets out to develop an explicit set of rules based on probabilities and memberships of sets within the data.
This paper describes a project to evaluate the potential of combining FTIR-ATR and artificial intelligence as the basis for a commercial test for natural fibres, blends, and their products. The development paths of the neural network and fuzzy logic systems are compared, and their respective contributions to the data analysis is evaluated.
It is shown that preprocessing of the spectral data and optimisation of parameters and models result in a fibres classifier which performs with good accuracy in differentiating Merino wool and Chinese cashmere samples. It is also shown that the approach can be adapted to provide quan
By: Gordon Allan School of Textiles, Heriot-Watt University, Galashiels, Scotland, UK
Submit Date: 6/14/2010 18:00