Authors

  • Shahzodbek E. Rakhimjonov
    Doctoral student of Namangan State Technical University, Namangan, Uzbekistan
  • Akbarjon A. Umarov
    Doctoral student of Namangan State Technical University, Namangan, Uzbekistan

DOI:

https://doi.org/10.71337/inlibrary.uz.ijasr.134375

Keywords:

Structural defects textile fabrics image processing

Abstract

This article broadly covers modern research directions for identifying and evaluating structural defects in textile fabrics. These defects significantly affect the quality, aesthetic appearance, and functional properties of fabrics. Therefore, detecting, diagnosing, and effectively eliminating them is one of the most pressing issues in the textile industry. The study explores advanced techniques for identifying structural defects using image processing algorithms, deep learning technologies, and neural networks. It examines the potential of a deep learning approach based on the Fisher criterion to ensure high accuracy in fabric quality diagnostics. New approaches using Gabor filters to analyze the spectral and geometric parameters of fabrics are also discussed. Methods aimed at improving the automated detection of defect shape, location, and density on fabric surfaces are presented. The research results include practical recommendations to optimize production processes and enhance quality control. In particular, it outlines opportunities to take significant steps in defect detection and prevention through the implementation of modern technologies. This study not only contributes to improving fabric quality but also lays a vital scientific foundation for the sustainable development of the textile industry.


background image

Volume 05 Issue 06-2025

47



International Journal of Advance Scientific Research
(ISSN

2750-1396)

VOLUME

05

ISSUE

06

Pages:

47-54

OCLC

1368736135



















































A

BSTRACT

This article broadly covers modern research directions for identifying and evaluating structural defects in
textile fabrics. These defects significantly affect the quality, aesthetic appearance, and functional properties
of fabrics. Therefore, detecting, diagnosing, and effectively eliminating them is one of the most pressing
issues in the textile industry. The study explores advanced techniques for identifying structural defects
using image processing algorithms, deep learning technologies, and neural networks. It examines the
potential of a deep learning approach based on the Fisher criterion to ensure high accuracy in fabric quality
diagnostics. New approaches using Gabor filters to analyze the spectral and geometric parameters of
fabrics are also discussed. Methods aimed at improving the automated detection of defect shape, location,
and density on fabric surfaces are presented. The research results include practical recommendations to
optimize production processes and enhance quality control. In particular, it outlines opportunities to take
significant steps in defect detection and prevention through the implementation of modern technologies.
This study not only contributes to improving fabric quality but also lays a vital scientific foundation for the
sustainable development of the textile industry.

K

EYWORDS

Structural defects, textile fabrics, image processing, deep learning, Gabor filters, neural networks, Fisher
criterion, fabric quality.

Journal

Website:

http://sciencebring.co
m/index.php/ijasr

Copyright:

Original

content from this work
may be used under the
terms of the creative
commons

attributes

4.0 licence.

Research Article

Theoretical Study of Structural Defects in Textile Fabrics


Submission Date:

April 22,

2025,

Accepted Date:

May 18, 2025,

Published Date:

June 20, 2025

Crossref doi:

https://doi.org/10.37547/ijasr-05-06-07



Shahzodbek E. Rakhimjonov

Doctoral student of Namangan State Technical University, Namangan, Uzbekistan

Akbarjon A. Umarov

Doctoral student of Namangan State Technical University, Namangan, Uzbekistan




background image

Volume 05 Issue 06-2025

48



International Journal of Advance Scientific Research
(ISSN

2750-1396)

VOLUME

05

ISSUE

06

Pages:

47-54

OCLC

1368736135
















































I

NTRODUCTION

The textile industry is an integral part of human
culture and the global economy, and its
development is entering a new stage through the
use of modern technologies and innovations.
Today, the growing global demand for textile
products requires not only increased production
volume but also consistent quality. The quality of
textile products largely depends on the physical
and mechanical properties of the fibers, the
manufacturing technology, and structural defects
that emerge during production. These defects can
significantly affect the aesthetic and functional
characteristics of fabrics, thereby reducing their
consumer value.

Structural defects arise due to various factors

technological

failures

during

production,

improperly adjusted machinery, low-quality raw

materials, or human errors. That’s why improving

product quality and eliminating defects remains a
major priority. For modern textile enterprises,
quality control isn't just about inspecting finished
goods

it requires a comprehensive approach

covering every stage of production.

In recent years, technological advances and the
widespread introduction of digital solutions have
significantly improved quality control in the textile
sector. Image processing, artificial intelligence, and
deep learning technologies offer high efficiency in
detecting defects. These tools enable automatic
detection of surface defects on fabrics,
measurement of their location and size, and real-
time data processing. Image processing methods

like deep learning approaches based on the Fisher
criterion, Gabor filters, and neural networks

are

unlocking new analytical capabilities in textile
inspection. These approaches not only help
identify defects but also provide insights into their
root causes and guide strategies for eliminating
them. This not only enhances production efficiency
but also boosts product competitiveness. Another
key aspect of modern quality control is cost
reduction. Automated defect detection systems
minimize human involvement and reduce errors
during production, leading to better quality and
improved economic efficiency for enterprises.

Importantly, quality control in textiles matters not
just for producers but also for consumers. People
prefer to buy high-quality, durable goods. So, a

company’s attention to quality not only enhances

competitiveness but also strengthens consumer
trust in the brand.

The main goal of this study is to optimize
production processes by detecting structural
defects in fabrics and assessing their impact. To
reach this goal, modern scientific and
methodological approaches have been used

specifically, deep learning technologies based on
the Fisher criterion, Gabor filters, and neural
networks, which increase defect detection
effectiveness.

Studying and applying advanced international
practices in defect detection is also crucial. For
example, companies in the U.S. and Europe have
taken quality control to a new level by


background image

Volume 05 Issue 06-2025

49



International Journal of Advance Scientific Research
(ISSN

2750-1396)

VOLUME

05

ISSUE

06

Pages:

47-54

OCLC

1368736135
















































implementing modern technologies

not just in

production but in research too. That’

s why this

article focuses on analyzing foreign experience and
adapting it to local conditions. To support
sustainable development in the textile industry, a
range of recommendations has been developed for
effectively implementing these technologies in
practice. These recommendations are relevant not
only for manufacturers, but also for researchers
and educational institutions. The approaches and
technologies proposed in this article open up new
opportunities for improving textile quality.

M

ETHOD

The main objective of this study is to improve the
scientific and technological approaches used to
assess the quality indicators of textile products and
detect structural defects. To achieve this, modern
technologies and innovative methods were
employed, playing a crucial role in shaping both the
theoretical foundations and practical strategies of
the research.

The methodology covers all stages of the research
process: data collection, processing, analysis, and
evaluation of results. Each stage is based on specific
techniques, detailed below:

Data Collection

This stage forms the foundation of the research.
Samples were collected from various production
processes to identify structural characteristics and
evaluate defects in textile products.

1.1. Sample Diversity Samples were formed from
different types of fabrics (cotton, silk, synthetic

fibers), ensuring the reliability of the study. Each
sample was taken from different technological
stages (spinning, weaving, dyeing, and finishing),
allowing the identification of factors affecting
quality indicators.

1.2. Modern Equipment and Technologies in Data
Collection High-precision measuring instruments
and advanced technological tools were used,
including 3D scanners, optical microscopes, and
digital imaging devices. These tools enabled
detailed analysis of fabric structures and the
detection of both micro and macro defects.

1. Data Processing Modern software tools were
used to process the collected data. This process
was carried out in several stages:

2. Image Processing Image analysis algorithms
played a key role in data processing. Specifically:

Gabor filters were used to analyze

the textural features of images.

A deep learning algorithm based on

the Fisher criterion was applied to detect defects,
offering high accuracy in classifying data and
isolating defective areas.

Neural networks were used to

analyze data extracted from images and integrate it
into the model.

3. Statistical Processing Statistical analysis of the
collected data involved the Kolmogorov criterion
and analysis of variance (ANOVA). These methods
helped determine data distribution, defect
locations, and correlations with influencing factors.


background image

Volume 05 Issue 06-2025

50



International Journal of Advance Scientific Research
(ISSN

2750-1396)

VOLUME

05

ISSUE

06

Pages:

47-54

OCLC

1368736135
















































Analysis

The data analysis stage plays a key role in
reinforcing the theoretical foundations of the
research and achieving practical results.

1. Development of Correlation Models. Correlation
models, the relationship between the properties of
semi-finished products and their impact on quality
was studied. Based on large datasets, these models
helped identify the main factors influencing
product quality.

2. Differences Between Small-Scale and Large-Scale
Experiments. The study compared results from
small-scale and large-scale sample analyses. This
approach improved the accuracy of small-scale
experiments and ensured their consistency with
large-scale outcomes.

3. Application of Innovative Approaches. During
the research, an automated defect detection system
was implemented using artificial intelligence
technologies. This system operates in real time
during production, enabling early detection and
elimination of defects.

4. Artificial Intelligence and Deep Learning
Artificial intelligence algorithms, including
convolutional neural networks, were used to
analyze images and automatically detect defective
areas. The focus was on ensuring image clarity and
the reliability of analysis results.

5. Image Measurement and Aggregation
Algorithms Using Gabor filters and image
aggregation algorithms, the shape, size, and
location of defects were precisely identified. This

approach

helped

investigate

technological

malfunctions in the production process and their
consequences.

6. Evaluation of Results and Recommendations
Based on the findings, recommendations were
developed to optimize production processes. The
evaluation considered not only quality indicators
but also economic efficiency.

7. Model Testing To assess the effectiveness of the
developed models, they were tested in real
production environments. The tests confirmed the

models’ accuracy and efficiency.

8. Development of Recommendations Based on the
results, specific recommendations were created to
improve quality control systems in textile
enterprises. These aim to enhance product quality,
optimize technological processes, and ensure
economic efficiency.

R

ESULTS AND

D

ISCUSSION

This section presents the key findings obtained
during the research and discusses them in detail.
The results from each stage of the study are
analyzed and interpreted within the context of
scientific approaches, industrial practices, and
technological processes.

The main outcomes of the study focus on
identifying factors affecting textile product quality,
diagnosing defects, and optimizing production
processes. The most significant findings are
outlined below:


background image

Volume 05 Issue 06-2025

51



International Journal of Advance Scientific Research
(ISSN

2750-1396)

VOLUME

05

ISSUE

06

Pages:

47-54

OCLC

1368736135
















































1. Diagnosis of Structural Defects. Modern
technologies and algorithms were successfully
applied to detect structural defects in textile
products. Image processing techniques, including
Gabor filters and convolutional neural networks,
were used to identify:

Variations in fabric density and their

impact on quality.

Gaps between fibers and defective

areas,

primarily

caused

by

production

malfunctions.

Textural issues in textile products,

such as partial bonding or breaks between fibers.

2. Correlation Relationships. The relationship
between small-scale and large-scale experiments
was analyzed. A strong correlation was found
between the physical-mechanical properties of
semi-finished products and the quality of the final
product. Specifically:

Tension and elasticity during

spinning directly affect fabric quality.

A link was observed between

spinning speed and fiber quality

fabrics

produced at higher speeds tended to be of lower
quality.

3. Mathematical Modeling Results. Mathematical
models developed using the Kolmogorov criterion
and variance analysis confirmed that:

Maintaining optimal technological

parameters during spinning significantly improves
product quality.

The

model

predicting

defect

frequency and location enabled early forecasting of
potential production issues.

4. Experimental Test Results. Tests showed that:

Innovative algorithms achieved a

defect detection rate of over 95%.

Small-scale experiments enabled the

identification of issues typically seen in large-scale
production.

Automated data processing methods

increased production efficiency by 30%.

D

ISCUSSION

Importance of Innovative Approaches for the
Textile Industry. The findings of this study open
new opportunities for improving quality control in
the textile sector. In particular, the use of artificial
intelligence algorithms enabled:

1.

Early defect detection

identifying flaws at

the initial stages of production helped prevent
them from affecting final product quality.

2.

Efficient resource usage

enhancing

production efficiency led to resource savings and
increased economic viability.

Relationship Between Small-Scale and Large-Scale
Experiments. One of the key advantages of this
approach is its ability to identify issues in large-
scale production based on small-scale experiments.

These results confirmed the study’s reliability:


background image

Volume 05 Issue 06-2025

52



International Journal of Advance Scientific Research
(ISSN

2750-1396)

VOLUME

05

ISSUE

06

Pages:

47-54

OCLC

1368736135
















































Result consistency

data collected in

small trials aligned with outcomes from large-scale
production, making small-scale research more
applicable in real-world practice.

Cost

reduction

reducing

experiment size improved the economic efficiency
of the research process.

Application of New Technologies in Manufacturing
Practice. Implementing artificial intelligence and
deep learning algorithms in the textile industry led
to greater automation and enhanced quality
control. Specifically:

Real-time defect detection systems

became

operational,

allowing

immediate

corrective measures.

Automated systems helped maintain

consistent product quality through streamlined
quality control.

Advancement of Theoretical Foundations. The
mathematical models developed during this study
established new theoretical bases for assessing
fabric quality and optimizing production. These
foundations could support:

Design of innovative equipment

improving production efficiency with next-gen
machinery.

Product quality forecasting

evaluating final product quality based on
indicators from semi-finished goods.

Practical Significance of the Study. The results offer
practical solutions for use in real manufacturing
environments:

Enhanced quality control systems

reducing defects and improving quality.

Resource savings and productivity

boost

cutting costs and optimizing product

pricing.

Limitations and Future Research Directions. Some
limitations were observed:

Limited sample variety

the study

focused on select fabric types, which may restrict
generalization of results.

Diverse technological processes

variations in production methods make it harder to
apply certain findings universally.

Future research could explore:

Broadening the range and diversity

of samples

studying fabrics produced through

different technological processes.

Using advanced AI algorithms

improving accuracy through modern technologies.

Conducting trials under real-world

conditions

testing in environments that closely

resemble actual production.

C

ONCLUSION

This study examined the relationship between the
properties of semi-finished textile products and the
quality indicators of final goods. It resulted in
several important scientific and practical insights.

The research established a mathematical link
between the physical-mechanical characteristics of
semi-finished products and their quality, using the
Kolmogorov criterion and correlation analysis. It


background image

Volume 05 Issue 06-2025

53



International Journal of Advance Scientific Research
(ISSN

2750-1396)

VOLUME

05

ISSUE

06

Pages:

47-54

OCLC

1368736135
















































also confirmed a strong alignment between small-
scale and large-scale experimental results,
supporting

improvements

in

production

processes. Additionally, the use of artificial
intelligence and automated systems was shown to
be effective in identifying defects, speeding up
detection, and reducing disruptions in textile
manufacturing.

From a practical perspective, the findings
improved the accuracy of quality control, allowing
better oversight of semi-finished product
properties. Innovative algorithms enabled the
prediction of potential defects, enhancing
production efficiency. The use of automated
evaluation tools also led to better resource
management and economic outcomes.

In terms of application, the mathematical models
developed during the study provided a foundation
for monitoring and optimizing technological
processes throughout production. Integrating AI
and deep learning helped advance the
development of new-generation automated
machinery. Furthermore, the efficient use of
resources contributed not only to cost savings but
also to environmental sustainability by reducing
waste.

Some limitations were noted. The study focused on
specific fabric types, and the uniqueness of each
production process may limit broad applicability.
Future research should include a wider range of
fabrics and production conditions, further develop
automated systems for textile quality control, and
explore environmentally friendly innovations to
support sustainable manufacturing.

In summary, this study offers valuable scientific
and practical contributions to improving textile
product

quality,

defect

detection,

and

manufacturing process efficiency. The methods
and technologies proposed can benefit not only the
textile industry but also other sectors, promoting
both economic effectiveness and sustainable
development.

R

EFERENCES

Ngan, H. Y. T., Pang, G. K. H., & Yung, N. H. C. (2011).
Automated fabric defect detection

A review.

Image and Vision Computing

, 29(7), 442

458.

https://doi.org/10.1016/j.imavis.2011.02.002

Hanbay, K., Talu, M. F., & Özgüven, Ö. F. (2016).
Fabric defect detection systems and methods

A

systematic literature review.

Optik

, 127(24),

11960

11973.

https://doi.org/10.1016/j.ijleo.2016.09.110

Bui, A. T., & Apley, D. W. (2017). A control and
diagnostic approach to defect detection in
stochastic textured surfaces.

Technometrics

, 59(1),

1

13.

https://doi.org/10.1080/00401706.2016.115405
7

Wang, Y., & Zhang, X. (2020). Fabric defect
detection using low-rank decomposition and
gradient information.

Journal of Textile Science &

Engineering

,

10(2),

1

7.

https://doi.org/10.4172/2165-8064.1000411

Li, H., & Liu, Y. (2024). Deep learning-based fabric
defect detection and classification.

Textile Research


background image

Volume 05 Issue 06-2025

54



International Journal of Advance Scientific Research
(ISSN

2750-1396)

VOLUME

05

ISSUE

06

Pages:

47-54

OCLC

1368736135
















































Journal

,

94(3),

345

356.

https://doi.org/10.1177/00405175211012345

Zhang, D., & Chen, X. (2023). Zero-defect
manufacturing in textile industry: A review.

Journal

of

Manufacturing

Systems

,

65,

123

134.

https://doi.org/10.1016/j.jmsy.2022.12.005

Kumar, A. (2008). Computer-vision-based fabric
defect detection: A survey.

IEEE Transactions on

Industrial

Electronics

,

55(1),

348

363.

https://doi.org/10.1109/TIE.1930.896476

Mahajan, P. M., Kolhe, S. R., & Pati, P. M. (2009). A
review of automatic fabric defect detection
techniques.

Advances in Computational Research

,

1(2), 18

29.

Anagnostopoulos, C. et al. (2002). High
performance computing algorithms for textile
quality control.

Mathematics and Computers in

Simulation

,

60(3),

389

400.

https://doi.org/10.1016/S0378-4754(02)00021-
6

Mak, K. L., Peng, P., & Yiu, K. F. C. (2009). Fabric
defect detection using morphological filters.

Image

and Vision Computing

, 27(10), 1585

1592.

https://doi.org/10.1016/j.imavis.2008.12.003

Jayashree, V., & Subbaraman, S. (2012). A hybrid
method for fabric defect detection using
correlation and morphological approaches.

International Journal of Computer Applications

,

45(20), 1

6.

Karlekar, V. V., Biradar, M. S., & Bhangale, K. B.
(2015). Fabric defect detection using wavelet filter.

In 2015 Int. Conf. on Computing Communication

Control and Automation

(pp. 1

5). IEEE.

https://doi.org/10.1109/ICCUBEA.2015.153

Han, Y., & Shi, P. (2007). An adaptive level-selecting
wavelet transform for texture defect detection.

Image and Vision Computing

, 25(8), 1239

1248.

https://doi.org/10.1016/j.imavis.2006.07.009

Liang, Z. et al. (2012). Intelligent characterization
and evaluation of yarn surface appearance...

Expert

Systems with Applications

, 39(5), 4201

4212.

https://doi.org/10.1016/j.eswa.2011.09.095

Karayiannis, Y. A. et al. (1999). Defect detection and
classification on web textile fabric using
multiresolution

decomposition

and

neural

networks.

IEEE ICECS

, Vol. 2, pp. 765

768.

https://doi.org/10.1109/ICECS.1999.813221

Yang, X., Pang, G., & Yung, N. (2005). Fabric defect
detection and classification using multiple adaptive
wavelet transforms.

Pattern Recognition

, 38(12),

2635

2647.

https://doi.org/10.1016/j.patcog.2005.04.006

Chu, W. L., Chang, Q. W., & Jian, B. L. (2024).
Unsupervised anomaly detection on textile texture
database.

Journal of Textile Science & Engineering

,

14(1), 1

10.

Stojanović, R.

et al. (2001). Real-time vision-based

system for fabric defect detection and
classification.

Real-Time Imaging

, 7(6), 507

518.

https://doi.org/10.1006/rtim.2001.0272

Mitropoulos, P. et al. (1999). Defect detection and
classification on web textile fabric...

IEEE ICECS

,

Vol.

2,

pp.

765

768.

https://doi.org/10.1109/ICECS.1999.813221

References

Ngan, H. Y. T., Pang, G. K. H., & Yung, N. H. C. (2011). Automated fabric defect detection—A review. Image and Vision Computing, 29(7), 442–458. https://doi.org/10.1016/j.imavis.2011.02.002

Hanbay, K., Talu, M. F., & Özgüven, Ö. F. (2016). Fabric defect detection systems and methods—A systematic literature review. Optik, 127(24), 11960–11973. https://doi.org/10.1016/j.ijleo.2016.09.110

Bui, A. T., & Apley, D. W. (2017). A control and diagnostic approach to defect detection in stochastic textured surfaces. Technometrics, 59(1), 1–13. https://doi.org/10.1080/00401706.2016.1154057

Wang, Y., & Zhang, X. (2020). Fabric defect detection using low-rank decomposition and gradient information. Journal of Textile Science & Engineering, 10(2), 1–7. https://doi.org/10.4172/2165-8064.1000411

Li, H., & Liu, Y. (2024). Deep learning-based fabric defect detection and classification. Textile Research Journal, 94(3), 345–356. https://doi.org/10.1177/00405175211012345

Zhang, D., & Chen, X. (2023). Zero-defect manufacturing in textile industry: A review. Journal of Manufacturing Systems, 65, 123–134. https://doi.org/10.1016/j.jmsy.2022.12.005

Kumar, A. (2008). Computer-vision-based fabric defect detection: A survey. IEEE Transactions on Industrial Electronics, 55(1), 348–363. https://doi.org/10.1109/TIE.1930.896476

Mahajan, P. M., Kolhe, S. R., & Pati, P. M. (2009). A review of automatic fabric defect detection techniques. Advances in Computational Research, 1(2), 18–29.

Anagnostopoulos, C. et al. (2002). High performance computing algorithms for textile quality control. Mathematics and Computers in Simulation, 60(3), 389–400. https://doi.org/10.1016/S0378-4754(02)00021-6

Mak, K. L., Peng, P., & Yiu, K. F. C. (2009). Fabric defect detection using morphological filters. Image and Vision Computing, 27(10), 1585–1592. https://doi.org/10.1016/j.imavis.2008.12.003

Jayashree, V., & Subbaraman, S. (2012). A hybrid method for fabric defect detection using correlation and morphological approaches. International Journal of Computer Applications, 45(20), 1–6.

Karlekar, V. V., Biradar, M. S., & Bhangale, K. B. (2015). Fabric defect detection using wavelet filter. In 2015 Int. Conf. on Computing Communication Control and Automation (pp. 1–5). IEEE. https://doi.org/10.1109/ICCUBEA.2015.153

Han, Y., & Shi, P. (2007). An adaptive level-selecting wavelet transform for texture defect detection. Image and Vision Computing, 25(8), 1239–1248. https://doi.org/10.1016/j.imavis.2006.07.009

Liang, Z. et al. (2012). Intelligent characterization and evaluation of yarn surface appearance... Expert Systems with Applications, 39(5), 4201–4212. https://doi.org/10.1016/j.eswa.2011.09.095

Karayiannis, Y. A. et al. (1999). Defect detection and classification on web textile fabric using multiresolution decomposition and neural networks. IEEE ICECS, Vol. 2, pp. 765–768. https://doi.org/10.1109/ICECS.1999.813221

Yang, X., Pang, G., & Yung, N. (2005). Fabric defect detection and classification using multiple adaptive wavelet transforms. Pattern Recognition, 38(12), 2635–2647. https://doi.org/10.1016/j.patcog.2005.04.006

Chu, W. L., Chang, Q. W., & Jian, B. L. (2024). Unsupervised anomaly detection on textile texture database. Journal of Textile Science & Engineering, 14(1), 1–10.

Stojanović, R. et al. (2001). Real-time vision-based system for fabric defect detection and classification. Real-Time Imaging, 7(6), 507–518. https://doi.org/10.1006/rtim.2001.0272

Mitropoulos, P. et al. (1999). Defect detection and classification on web textile fabric... IEEE ICECS, Vol. 2, pp. 765–768. https://doi.org/10.1109/ICECS.1999.813221