Surface defect detection for capsule based on improved YOLOv3 algorithm

Chunfeng Yang; Huiyu Chen; Jiajia Lu

doi:10.1117/12.2660841

2 February 2023 Surface defect detection for capsule based on improved YOLOv3 algorithm

Chunfeng Yang, Huiyu Chen, Jiajia Lu

Proceedings Volume 12462, Third International Symposium on Computer Engineering and Intelligent Communications (ISCEIC 2022); 124622Z (2023) https://doi.org/10.1117/12.2660841
Event: International Symposium on Computer Engineering and Intelligent Communications (ISCEIC 2022), 2022, Xi'an, China

Abstract

Aiming at the low detection speed and poor detection accuracy of traditional image processing algorithms, this paper introduces a surface defect detection method for capsule based on an improved YOLOv3 algorithm. Firstly, to obtain important information, the key parts of the image with more refined features are extracted by introducing a dual attention mechanism into the network. Secondly, considering the feature variability of different defect scales of the capsule, the feature detection network is optimized by using adaptive convolution to improve the detection ability of targets. Thirdly, we use the K-means++ algorithm to cluster the anchor box of the target samples to obtain more suitable anchor boxes for the detection task. The experimental results show that although the algorithm loses a little detection speed, it improves the accuracy of capsule detection and meets the demand for real-time detection. Compared with YOLOv3, the average detection accuracy increases from 84.3% to 88.5%.

Citation Download Citation

Chunfeng Yang, Huiyu Chen, and Jiajia Lu "Surface defect detection for capsule based on improved YOLOv3 algorithm", Proc. SPIE 12462, Third International Symposium on Computer Engineering and Intelligent Communications (ISCEIC 2022), 124622Z (2 February 2023); https://doi.org/10.1117/12.2660841

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