EfficientDet for fabric defect detection based on edge computing
Source abstract
Abstract
The productivity of textile industry is positively correlated with the efficiency of fabric defect detection. Traditional manual detection methods have gradually been replaced by deep learning algorithms based on cloud computing due to the low accuracy and high cost of manual methods. Nonetheless, these cloud computing-based methods are still suboptimal due to the data transmission latency between the end devices and the cloud. To facilitate defect detection with more efficiency, a low-latency, low power consumption, easy upgrade, and automatical visual inspection system with the help of edge computing are proposed in this work. Firstly, the method uses EfficientDet-D0 as the detection algorithm, integrating the advantages of lightweight and scalable and can suit the resource-constrained edge device. Secondly, we performed data augmentations on five fabric datasets and verified the adaptability of the model in different types of fabrics. Finally, we transplanted the trained model to the edge device NVIDIA Jetson TX2 and optimized the model with TensorRT to make it detection faster. The performance of the proposed method is evaluated in five fabric datasets. The detection speed is up to 22.7 frame per second (FPS) on the edge device Jetson TX2. Compared with the cloud-based method, the response time is reduced by 2.5 times, with the capability of real-time industrial defect detection.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Industrial Vision Systems and Defect Detection · Surface Roughness and Optical Measurements · Image and Object Detection Techniques
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
College of Electronics and Information, Xi’an Polytechnic University, Xi’an, China
Funding
Shaanxi Provincial Education Department
References (35)
Show source-provided references
1. 10.1109/TASE.2017.269674810.1109/TASE.2017.2696748 ↗
2. 10.1016/j.procs.2018.07.05810.1016/j.procs.2018.07.058 ↗
3. 10.1177/004051751984063610.1177/0040517519840636 ↗
4. 10.3788/AOS201838.081500210.3788/AOS201838.0815002 ↗
5. 10.1109/JIOT.2020.298305010.1109/JIOT.2020.2983050 ↗
6. 10.1016/S0262-8856(99)00062-110.1016/S0262-8856(99)00062-1 ↗
7. Tsai DM · Measurement · 2014 · 44 · 40
8. 10.1177/152808371349000210.1177/1528083713490002 ↗
9. 10.1016/j.neucom.2017.01.03910.1016/j.neucom.2017.01.039 ↗
10. 10.1080/00405000.2015.110245810.1080/00405000.2015.1102458 ↗
11. Xiaobo Y · J Text Res · 2013 · 4 · 26
12. 10.1109/TMM.2014.229883210.1109/TMM.2014.2298832 ↗
13. 10.1111/mice.1226310.1111/mice.12263 ↗
14. 10.1109/TPAMI.1986.476785110.1109/TPAMI.1986.4767851 ↗
15. Jing J · J Eng Fiber Fabr · 2020 · 15
16. Redmon J · arXiv preprint · 2018
17. 10.1109/ACCESS.2019.291362010.1109/ACCESS.2019.2913620 ↗
18. 10.1016/j.neucom.2018.10.07010.1016/j.neucom.2018.10.070 ↗
19. 10.1109/TIP.2019.295974110.1109/TIP.2019.2959741 ↗
20. 10.1109/ACCESS.2020.300965410.1109/ACCESS.2020.3009654 ↗
21. 10.1109/TWC.2019.294614010.1109/TWC.2019.2946140 ↗
22. Shi W · J Comput Res Dev · 2017 · 54 · 907
23. 10.1016/j.future.2019.02.05010.1016/j.future.2019.02.050 ↗
24. 10.1109/MNET.001.180052610.1109/MNET.001.1800526 ↗
25. 10.1109/TII.2019.290821010.1109/TII.2019.2908210 ↗
26. 10.1109/TSC.2017.266200810.1109/TSC.2017.2662008 ↗
27. 10.1109/ACCESS.2018.283491610.1109/ACCESS.2018.2834916 ↗
28. Tan M · 10781
29. 10.1186/s40537-019-0197-010.1186/s40537-019-0197-0 ↗
30. Tan M · arXiv preprint · 2019
31. Chollet F · 1251
32. Liu S · 8759
33. Redmon J · 7263
34. Liu W · 21
35. Goyal P · IEEE Trans Pattern Anal Mach Intell · 2018 · 39 · 2999