Defect detection of laminated surface in the automated fiber placement process based on improved CenterNet

Xuan Wang; Shuo Kang; Weidong Zhu

doi:10.3788/IRLA20210011

Infrared and Laser Engineering, Volume. 50, Issue 10, 20210011(2021)

Defect detection of laminated surface in the automated fiber placement process based on improved CenterNet

Xuan Wang1... Shuo Kang1, and Weidong Zhu1,23 |Show fewer author(s)

Author Affiliations

¹School of Mechanical and Engineering, Zhejiang University, Hangzhou 310027, China

²State Key Laboratory of Fluid Power and Mechatronic Systems, College of Mechanical Engineering, Zhejiang University, Hangzhou 310027, China

³Key Laboratory of Advanced Manufacturing Technology of Zhejiang Province, College of Mechanical Engineering, Zhejiang University, Hangzhou 310027, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(12)

Fig. 1. (a) Keypoint heatmap; (b) Keypoint offset; (c) Object size

Download full size

View in Article

Fig. 2. Schematic diagram of CenterNet network with ResNet-18 as backbone network

Download full size

View in Article

Fig. 3. Process of network learning fusion coefficient

Download full size

View in Article

Fig. 4. Structure diagram of AFP-CenterNet network

Download full size

View in Article

Fig. 5. (a) GT box and feasible boxes; (b) GT box heatmap obtained using a circular to filter feasible boxes; (c) GT box heatmap obtained using a ellipse to filter feasible boxes

Download full size

View in Article

Fig. 6. Improved bandwidth parameter solution method

Download full size

View in Article

Fig. 7. (a) Data acquisition platform; (b) Infrared images of six typical kinds of AFP laminated surface defects

Download full size

View in Article

Fig. 8. Graph of changes in loss value

Download full size

View in Article

Fig. 9. Detection results of single frame infrared image

Download full size

View in Article

Table 1. Detection results of CenterNet and AFP-CenterNet on AFP infrared data sets

View table

View in Article

Table 1. Detection results of CenterNet and AFP-CenterNet on AFP infrared data sets

Network model	AP	AP50	AP75	Time/ms	FPS	Memory/MB
CenterNet(DLA-34)	0.727	0.926	0.753	81	12.3	77.0
CenterNet(ResNet-101)	0.693	0.904	0.728	93	10.7	204.0
CenterNet(ResNet-18)	0.626	0.825	0.647	31	32.2	60.3
AFP-CenterNet(MobileNetV3+ASFF)	0.694	0.902	0.731	42	23.8	12.9

Table 2. Detection results of different network models on AFP infrared data sets
View table
View in Article
Table 2. Detection results of different network models on AFP infrared data sets
Network model AP50 Model memory/MB
SSD 0.806 86.5
YOLOv3 0.819 235.0
AFP-CenterNet 0.902 12.9

Table 3. Detection speed of different network models only using CPU
View table
View in Article
Table 3. Detection speed of different network models only using CPU
Network model Time/ms FPS
SSD 537 1.8
YOLOv3 451 2.2
CenterNet(ResNet-101) 425 2.3
AFP-CenterNet 235 4.2

Tools

Get Citation

Copy Citation Text

Xuan Wang, Shuo Kang, Weidong Zhu. Defect detection of laminated surface in the automated fiber placement process based on improved CenterNet[J]. Infrared and Laser Engineering, 2021, 50(10): 20210011

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Photoelectric measurement

Received: Jan. 12, 2021

Accepted: --

Published Online: Dec. 7, 2021

The Author Email:

DOI:10.3788/IRLA20210011

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Detection results of CenterNet and AFP-CenterNet on AFP infrared data sets

Table 1. Detection results of CenterNet and AFP-CenterNet on AFP infrared data sets

Table 2. Detection results of different network models on AFP infrared data sets

Table 2. Detection results of different network models on AFP infrared data sets

Table 3. Detection speed of different network models only using CPU

Table 3. Detection speed of different network models only using CPU

微信扫一扫：分享