Improved YOLOv8-based defect detection for furniture artificial panels under complex texture backgrounds.
Saved in:
| Title: | Improved YOLOv8-based defect detection for furniture artificial panels under complex texture backgrounds. |
|---|---|
| Authors: | Yang, Junhang1,2 (AUTHOR), Luo, Lufeng1,2 (AUTHOR) luolufeng@fosu.edu.cn, Chen, Mingyou1,2 (AUTHOR), Su, Shijie1,2 (AUTHOR), Yan, Jianhua3 (AUTHOR) |
| Source: | Nondestructive Testing & Evaluation. Aug2026, Vol. 41 Issue 8, p4932-4960. 29p. |
| Subjects: | Nondestructive testing, Object recognition (Computer vision), Engineering inspection, Engineered wood, Loss functions (Statistics) |
| Abstract: | In panel furniture production, furniture artificial panels are prone to defects, while manual inspection remains subjective and inconsistent. Automated non-destructive testing provides objective detection but faces challenges from severe feature distortion in complex texture backgrounds, reducing accuracy. This study proposes improved YOLOv8-based defect detection for furniture artificial panels under complex texture backgrounds. Initially, the C2f-MFM module is introduced to enhance defect–background separation through multi-dimensional feature mixing, effectively distinguishing edge defects. Subsequently, the MFS-Net module improves the detection of defects with diverse morphologies and random distributions across large IoU ranges. Ultimately, Wise-IoU v3 is employed to mitigate the impact of low-quality samples during training. Compared with the original YOLOv8s, the proposed model reduces parameters by 21.27% and FLOPs by 17.71%. Experiments on a self-constructed dataset demonstrate improvements of 1.9% in mAP@0.5 and 2.84% in mAP50–95, achieving 97.5% and 63.32%, respectively. Compared with other defect detection methods, the proposed approach shows clear advantages. Heatmap visualization using Grad-CAM validates its effectiveness, offering an efficient and robust solution for non-destructive defect detection in complex texture backgrounds. [ABSTRACT FROM AUTHOR] |
| Copyright of Nondestructive Testing & Evaluation is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
Be the first to leave a comment!
