Derinlemesine Özellik Piramit Ağı Kullanarak Yüzey Hata Tespiti

Öz

Yüzey hata tespiti, imalat sistemlerindeki en önemli kalite kontrol bileşenlerinden biridir. Üretim sistemlerinde otomatik yüzey hata algılama yöntemlerinin uygulanması, yüksek kaliteli ürünlerin sağlanmasında önemli bir etkendir. Bu çalışmada, otomatik yüzey hata tespiti için derinlemesine ayrılabilir evrişim tabanlı Derin Özellikli Piramit Ağ (DÖPA) mimarisi geliştirilmiştir. Bu ağ mimarisinde, önceden eğitilmiş VGG19 ağ mimarisinin öğrenilmiş parametreleri kullanılmıştır. Önerilen modelin performansını test etmek için hata tespit görüntüleri içeren MT veri seti kullanılmıştır. Deneysel çalışmalarda, önerilen DÖPA mimarisi kullanılarak %86,86 F1-skor elde edilmiştir. Bu sonuçlar, önerilen modelin var olan çalışmalardan daha başarılı olduğunu göstermiştir.

Anahtar Kelimeler

• Yüzey Hata Tespiti
• Derinlemesine Ayrılabilir Evrişim
• Özellikli Piramit Ağı
• Derin Öğrenme.

Surface Defect Detection Using Depthwise Feature Pyramid Network

Öz

Surface defect detection is one of the most important quality control components in manufacturing systems. The application of automatic surface defect detection methods in production systems is an important factor in ensuring high-quality products. In this study, depthwise separable convolution-based Deep Feature Pyramid Network (DÖPA) architecture was developed for automatic surface defect detection. In this network architecture, the learned parameters of the pre-trained VGG19 network architecture were used. MT dataset with defect detection images was used to test the performance of the proposed model. In experimental studies, 86.86% F1-score was obtained using the proposed DOPA architecture. These results showed that the proposed model was more successful than the existing studies.

Anahtar Kelimeler

• Surface Defect Detection
• Depthwise Separable Convolution
• Feature Pyramid Network
• Deep Learning

Kaynakça

1. Balzategui, J., Eciolaza, L., & Arana-Arexolaleiba, N. (2020). Defect detection on Polycrystalline solar cells using Electroluminescence and Fully Convolutional Neural Networks. Proceedings of the 2020 IEEE/SICE International Symposium on System Integration, SII 2020, 949–953. https://doi.org/10.1109/SII46433.2020.9026211
2. Bergmann, P., Fauser, M., Sattlegger, D., & Steger, C. (2019). MVTEC ad-A comprehensive real-world dataset for unsupervised anomaly detection. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2019-June, 9584–9592. https://doi.org/10.1109/CVPR.2019.00982
3. Cao, J., Yang, G., & Yang, X. (2021). A Pixel-Level Segmentation Convolutional Neural Network Based on Deep Feature Fusion for Surface Defect Detection. IEEE Transactions on Instrumentation and Measurement, 70. https://doi.org/10.1109/TIM.2020.3033726
4. Computer Vision Group, F. (1996). TILDA Textile Texture-Database. https://lmb.informatik.uni-freiburg.de/resources/datasets/tilda.en.html
5. Dong, H., Song, K., He, Y., Xu, J., Yan, Y., & Meng, Q. (2020). PGA-Net: Pyramid Feature Fusion and Global Context Attention Network for Automated Surface Defect Detection. IEEE Transactions on Industrial Informatics, 16(12), 7448–7458. https://doi.org/10.1109/TII.2019.2958826
6. Firat, H., & Hanbay, D. (2021). Classification of Hyperspectral Images Using 3D CNN Based ResNet50. 2021 29th Signal Processing and Communications Applications Conference (SIU), 1–4. https://doi.org/10.1109/SIU53274.2021.9477899
7. Fırat, H., & Alpaslan, N. (2020). An effective approach to the two-dimensional rectangular packing problem in the manufacturing industry. Computers and Industrial Engineering, 148, 106687. https://doi.org/10.1016/j.cie.2020.106687
8. Hanbay, K., Talu, M. F., & Özgüven, Ö. F. (2016). Fabric defect detection systems and methods—A systematic literature review. Optik, 127(24), 11960–11973. https://doi.org/10.1016/j.ijleo.2016.09.110

9. He, Y., Song, K., Meng, Q., & Yan, Y. (2020). An End-to-End Steel Surface Defect Detection Approach via Fusing Multiple Hierarchical Features. IEEE Transactions on Instrumentation and Measurement, 69(4), 1493–1504. https://doi.org/10.1109