Image Analysis Techniques for Fabric Pilling Evaluation

Fabric pilling, the undesirable balling of fibres on textile surfaces, has traditionally been assessed by visual comparison against standard reference cards. Recent advances in image analysis have permitted objective, reproducible and early-stage detection of pilling through a range of non-intrusive optical and computational methods. Three-dimensional imaging modalities such as optical coherence tomography reveal the submicrometre structure of fibre fuzz above the fabric plane. Two-dimensional camera-based systems exploit frequency-domain and spatial-domain transforms—such as Fourier, wavelet and Gabor filtering—to isolate pill clusters from background texture. Automated feature extraction pipelines quantify morphological descriptors including pill size, count, area fraction and textural homogeneity. Machine-learning classifiers, from support vector machines to neural networks augmented by principal component analysis, enable grading of pilling severity with accuracies approaching those of human experts. Integration of these image-driven methods into production lines and quality-control laboratories supports rapid feedback on material formulation, fabric finishing and machine settings, thereby enhancing global textile sustainability and consumer satisfaction.

Research from Nature Portfolio

Innovations in contactless surface metrology have extended roughness measurement systems to quantify fabric hairiness across successive production stages. Spatial parameters derived from a grid-based quadrat method capture variations in protruding fibre density before and after singeing, enabling precise monitoring of finishing treatments on woven shirts. Results show significant hairiness reduction post-singeing and stabilisation after easy-care finishing.

Three-dimensional imaging via optical coherence tomography has been applied to pilling prediction by reconstructing the fuzz layer above the fabric surface with micrometre-scale resolution. Textural features of the pilling layer are classified using both linear and non-linear discriminant analyses, achieving over 98 per cent accuracy for fabrics subjected to short-term abrasion. This method facilitates quantitative tracing of early pilling development and offers superior sensitivity compared with conventional visual tests.

Image Analysis Techniques for Fabric Pilling Evaluation publication trend

The graph below shows the total number of articles in image analysis techniques for fabric pilling evaluation across all publications each year (not limited to Nature Index journals).

Technical terms

Optical coherence tomography: A non-invasive imaging technique that captures depth-resolved cross-sections of fabric surfaces using low-coherence interferometry.

Wavelet transform: A mathematical decomposition that analyses image content at multiple scales, useful for separating texture and pill features.

Gabor filter: A linear filter used for edge detection and texture analysis, characterised by orientation and frequency selectivity.

Convolutional neural network: A deep learning architecture that automatically learns hierarchical image features through stacked convolutional layers.

Principal component analysis: A statistical method for reducing feature dimensionality by identifying orthogonal components that capture maximal variance.

Support vector machine: A supervised learning algorithm that constructs optimal hyperplanes to separate classes in a high-dimensional feature space.

References

1. Changes in hairiness of woven fabrics at the production and finishing stages. Scientific Reports (2025).
2. A Method for the Assessment of Textile Pilling Tendency Using Optical Coherence Tomography. Sensors (2020).
3. Applying Image Processing to the Textile Grading of Fleece Based on Pilling Assessment. Fibers (2018).
4. Using Deep Principal Components Analysis-Based Neural Networks for Fabric Pilling Classification. Electronics (2019).
5. An objective fabric pilling evaluation based on wavelet transform and Gabor filter. Journal of Engineered Fibers and Fabrics (2022).
6. Prediction of textile pilling resistance using optical coherence tomography. Scientific Reports (2022).