Automatic Fabric Defect Detection Method Using AC-YOLOv5
نویسندگان
چکیده
In the face of detection problems posed by complex textile texture backgrounds, different sizes, and types defects, commonly used object networks have limitations in handling target sizes. Furthermore, their stability anti-jamming capabilities are relatively weak. Therefore, when more diverse, false detections or missed likely to occur. order meet stringent requirements defect detection, we propose a novel AC-YOLOv5-based method. This method fully considers optical properties, distribution, imaging specific textiles. First, Atrous Spatial Pyramid Pooling (ASPP) module is introduced into YOLOv5 backbone network, feature map pooled using convolution cores with expansion rates. Multiscale information obtained from maps receptive fields, which improves defects sizes without changing resolution input image. Secondly, squeeze-and-excitation (CSE) channel attention proposed, CSE network. The weights each through self-learning further improve capability. Finally, large number fabric images were collected an inspection system built on circular knitting machine at industrial site, experiments conducted self-built dataset. experimental results showed that AC-YOLOv5 can achieve overall accuracy 99.1% for datasets, satisfying applications areas.
منابع مشابه
Fabric Defect Detection Using Homogeneity
Fabric defect detection algorithm based on local neighborhood is proposed to improve the accuracy and real-time of Fabric defect detection. A local neighborhood window moves over the entire inspection image. For homogeneity measurethe coefficient of variation is used. A defect-free region will generate a smaller value of Variation Coefficient than that of a defective region. To extract and segm...
متن کاملFabric defect detection using adaptive wavelet
This paper studies the adaptive wavelet design for fabric defect detection. In order to achieve translation invariance and more flexible design, the wavelet design focused on nonsubsampled wavelet transform. We design the wavelet filters under the constraints that the analysis filters are power complementary, and the wavelet has only one vanishing moment, which corresponds to a multiscale edge ...
متن کاملFabric Defect Detection Using Steerable Pyramid
In this paper, a novel idea is proposed for fabric defect detection. Defects are detected in the fabric using steerable pyramid along with a defect detection algorithm. Various steerable pyramid of four size 256*256, 128*128, 64*64, 32*32 and with four orientation bands 0,45, 90, 135 are used. Utilizing a Steerable pyramid proved adequate in the representation of fabric images in multi-scale an...
متن کاملDefect Detection in Textile Fabric Using Mrcsf
In textile industry, reliable and accurate quality control and inspection becomes an important element. Presently, this is still accomplished by human experience, which is very time consuming and prone to errors. Hence automated visual inspection systems become mandatory in textile industries. Here the automatic fabric defect detection using Multi Resolution Combined Statistical and spatial Fre...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Electronics
سال: 2023
ISSN: ['2079-9292']
DOI: https://doi.org/10.3390/electronics12132950