Citation: Jun, X.; Ruru, Pan.;
Weidong, G. Online Detection of
Fabric Defects Based on Improved
CenterNet with Deformable
Convolution. Sensors 2022, 22, 4718.
https://doi.org/10.3390/s22134718
Academic Editor: Kelvin K.L. Wong
Received: 10 May 2022
Accepted: 6 June 2022
Published: 22 June 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Online Detection of Fabric Defects Based on Improved
CenterNet with Deformable Convolution
Jun Xiang , Ruru Pan and Weidong Gao *
School of Textile Science & Engineering, Jiangnan University, No. 1800, Lihu Avenue, Wuxi 214122, China;
skyjun@163.com (J.X.); prrsw@163.com (R.P.)
* Correspondence: gaowd3@163.com
Abstract:
The traditional manual defect detection method has low efficiency and is time-consuming
and laborious. To address this issue, this paper proposed an automatic detection framework for
fabric defect detection, which consists of a hardware system and detection algorithm. For the
efficient and high-quality acquisition of fabric images, an image acquisition assembly equipped
with three sets of lights sources, eight cameras, and a mirror was developed. The image acquisition
speed of the developed device is up to 65 m per minute of fabric. This study treats the problem of
fabric defect detection as an object detection task in machine vision. Considering the real-time and
precision requirements of detection, we improved some components of CenterNet to achieve efficient
fabric defect detection, including the introduction of deformable convolution to adapt to different
defect shapes and the introduction of i-FPN to adapt to defects of different sizes. Ablation studies
demonstrate the effectiveness of our proposed improvements. The comparative experimental results
show that our method achieves a satisfactory balance of accuracy and speed, which demonstrate the
superiority of the proposed method. The maximum detection speed of the developed system can
reach 37.3 m per minute, which can meet the real-time requirements.
Keywords:
fabric defect detection; feature pyramid network; deformable convolution; object detec-
tion; online detection
1. Introduction
In the weaving process of fabrics, due to the influence of the technological process,
weaving equipment, or weaving environment, it is inevitable to cause various defects on
the surface of fabrics. The appearance of defects will not only affect the appearance of the
fabric, but also reduce the commercial value of the fabric. Relevant reports [
1
] show that if
there are obvious defects in the surface of the fabric, its price will be reduced by more than
50%; therefore, defect detection is an important step in fabric quality control; however, at
present, most textile enterprises still rely on manual cloth inspection, which not only has the
shortcomings of low efficiency and high cost, but is also prone to false detection or missed
inspection after visual fatigue. With the advancement of digitization and intelligence, the
development of fabric defect detection towards automation is an inevitable trend.
The automatic detection of fabric defects mainly includes two steps: firstly, images of
the fabric surface are captured by using an industrial camera, and then the existence and
type of defect in the image are judged by designing a recognition algorithm. The detection
methods based on computer vision have the advantages of high precision, high efficiency,
and strong stability; therefore, the automatic detection of fabric defects by machine vision
instead of human vision has become a research hotspot; however, as shown in Figure 1, the
main characteristics of the defects in the fabric are as follows: (1) rich types and different
shapes and (2) low visual significance, which makes the identification task very challenging.
Sensors 2022, 22, 4718. https://doi.org/10.3390/s22134718 https://www.mdpi.com/journal/sensors
