[Identification of fine wool and cashmere by using Vis/NIR spectroscopy technology]

Area of Science:

• Textile Science
• Spectroscopy
• Chemometrics

Background:

• Cashmere and fine wool fibers exhibit superficial similarities, posing challenges for accurate identification.
• Existing methods for differentiating these fibers often lack satisfactory resolution and accuracy.
• Developing a reliable and efficient identification technique is crucial for the textile industry.

Purpose of the Study:

• To develop an automatic recognition scheme for distinguishing cashmere and fine wool fibers.
• To apply Vis/NIR spectroscopy combined with a mixed algorithm for fiber identification.
• To enhance the accuracy and efficiency of cashmere and fine wool discrimination.

Main Methods:

• Vis/NIR diffuse reflectance spectroscopy was employed to collect spectral data from cashmere and fine wool.
• Data preprocessing involved Standard Normal Variate (SNV) for scatter correction and Savitzky-Golay smoothing.
• Principal Component Analysis (PCA) was used for dimensionality reduction, selecting 6 principal components (PCs).
• A three-layer back-propagation artificial neural network (BP-ANN) was trained using PCA scores for classification.

Main Results:

• The combined PCA-BP-ANN model achieved high accuracy in discriminating between cashmere and fine wool fibers.
• The system demonstrated rapid and effective performance, outperforming previous identification systems.
• Selected 6 principal components accounted for 99.8% of the spectral data variability.

Conclusions:

• Vis/NIR spectroscopy coupled with PCA-BP technology provides a robust and efficient method for cashmere and fine wool identification.
• The developed automatic recognition scheme offers significant advantages in speed and accuracy.
• This technique holds promise for practical applications in the textile industry for quality control and authentication.