Automatic 3D surface defects detection of fiber-reinforced epoxy resin composites

Bin Lin; Helin Li; Chen Zhang; Liang Xu; Tianyi Sui

doi:10.1117/12.2652472

2 November 2022 Automatic 3D surface defects detection of fiber-reinforced epoxy resin composites

Bin Lin, Helin Li, Chen Zhang, Liang Xu, Tianyi Sui

Proceedings Volume 12351, International Conference on Advanced Sensing and Smart Manufacturing (ASSM 2022); 123511F (2022) https://doi.org/10.1117/12.2652472
Event: International Conference on Advanced Sensing and Smart Manufacturing (ASSM 2022), 2022, Nanjing, China

Abstract

Fiber-reinforced epoxy resin composites (FRERCs) are widely used in the aerospace industry. The surface defects incurred during the manufacturing process of FRERCs have detrimental effects on the aircraft during high-speed flights. To accurately and efficiently detect FRERCs’ surface defects. A novel lightweight and two-stage 3D defect detection network is proposed in this paper, i.e., NGSP-Point-Net. The first stage of the network is 3D Mask proposers; the proposers generate multiple candidate 3D surface defect masks, allowing the second stage of the network to focus on potential defect regions. The second stage of the network is the 3D merger; it classifies the candidate surface 3D defect masks to determine whether the mask is a defect or not and keeps only the target masks to obtain the final detection results. In addition, a distributed semi-real-time detection system based on the NGSP-Point-Net is constructed in this paper to simulate the online and high-speed detection of FRERCs’ surface defects in the manufacturing process. Several NGSP-Point-Net-based experiments were conducted. The metrics of the precision, accuracy, Matthews correlation coefficient (MCC), and the mean intersection of union (mIoU) reach 0.9914, 0.9896, 0.9792, and 0.9243, respectively; and the inference speed reaches about 150,000 points/s with a single Nvidia RTX 3080. Experiments demonstrate that the NGSP-Point-Net provides a new possibility to accurately and efficiently detect surface defects of FRERCs.

Citation Download Citation

Bin Lin, Helin Li, Chen Zhang, Liang Xu, and Tianyi Sui "Automatic 3D surface defects detection of fiber-reinforced epoxy resin composites", Proc. SPIE 12351, International Conference on Advanced Sensing and Smart Manufacturing (ASSM 2022), 123511F (2 November 2022); https://doi.org/10.1117/12.2652472

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
6 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Defect detection

Clouds

3D acquisition

Composites

Manufacturing

Composite resins

Epoxies

Show All Keywords

Keywords/Phrases

Search In:

Publication Years