Yuqori tezlikda trikotaj mato ishlab chiqarishda nuqsonlarni ishonchli va real vaqt rejimida aniqlash uchun ko‘p algoritmli sintez (integratsiya) tizimi

Admin bo'lib kirish (www.admin.slib.uz)

Tizim sinov (TEST) rejimida ishlamoqda! Murojat uchun @slib_support

Eski talqinga o'tish link

16 апрел 2026

Asosiy til:Ingliz

Yuqori tezlikda trikotaj mato ishlab chiqarishda nuqsonlarni ishonchli va real vaqt rejimida aniqlash uchun ko‘p algoritmli sintez (integratsiya) tizimi

Fan yo'nalishi:Muhandislik (turli xil)

JizPI Xabarnomasi 2026 ....pdf

PDF

JizPI xabarnomasi

Maqola chop etish see more

MAQOLA ANNOTATSIYASI

Avtomatlashtirilgan vizual nazorat (AVI) yuqori tezlikdagi trikotaj mato ishlab chiqarishda sifatni ta'minlash uchun hal qiluvchi ahamiyatga ega. YOLO kabi chuqur o'qitish (deep learning) modellari yuqori aniqlikni ta'minlasa-da, ularning samaradorligi harakatdagi xiralashish, yoritishning o'zgarishi va nuqsonlarning yangi turlari kabi real sharoitdagi o'zgaruvchanliklar ta'sirida pasayishi mumkin. Ushbu maqolada nuqsonlarni real vaqt rejimida ishonchli aniqlash uchun asosiy chuqur o'qitish detektori (YOLOv8), teksturadagi anomaliyalarni segmentatsiya qilish moduli (Mahalliy binar andozalar va Gauss aralashmasi modellariga asoslangan) hamda yuqori tezlikdagi an'anaviy klassifikatorning (Optimallashtirilgan Random Forest) afzalliklarini sinergik tarzda birlashtirgan yangi Ko'p algoritmli fuzion freymvork (MAFF) taklif etiladi. Tizim qaror qabul qilish darajasidagi fuzion mantiqdan foydalanadi: YOLOv8 dastlabki chegaralovchi ramkalarni (bounding boxes) taqdim etadi; tekstura moduli YOLO tomonidan o'tkazib yuborilgan nozik anomaliyalarni tahlil qiladi; Random Forest esa, ayniqsa past ishonchli YOLO bashoratlari uchun tezkor "tekshiruvchi" yoki nuqson kichik turlarining klassifikatori vazifasini bajaradi. Turli sanoat sharoitlarida 8 ta nuqson sinfini qamrab olgan 15 000 ta tasvirdan iborat xususiy ma'lumotlar to'plamida o'qitilgan va tekshirilgan MAFF tizimi 98,2% o'rtacha aniqlikka (mAP@0.5) erishdi. Bu ko'rsatkich mustaqil YOLOv8 (95,1%) va boshqa yagona modelli yondashuvlarda

MUALIFLAR

Teglar

# Detection# (AVI)# Fusion# YOLOv8# Forest# Fabric

Maqolani baholang

0

0 ta

Maqola idintifikatorlari

ROI:

Mavjud emas

DOI:

Mavjud emas

Foydalanilgan adabiyotlar

1. Zhang, L., et al. (2021). High-speed vision systems for web inspection: A review. IEEE Transactions on Industrial Informatics, 17(5), 3067-3078

2. Kumar, V., & Patel, R. (2020). Economic impact of automated vs. manual inspection in textile mills. International Journal of Production Economics, 230, 107829

3. Ozturk, S., & Koc, B. (2022). Challenges in knitted fabric defect detection: Elasticity, texture, and defect diversity. The Journal of The Textile Institute, 113(5), 789-798.

4. Redmon, J., & Farhadi, A. (2018). YOLOv3: An incremental improvement. arXiv preprint arXiv:1804.02767

5. Jocher, G., et al. (2023). Ultralytics YOLOv8: A state-of-the-art object detection model. https://docs.ultralytics.com

6. Korabayev, S., et al. (2025). Trikotaj mato ishlab chiqarishda avtomatlashtirilgan sifat nazorati uchun sun'iy intellekt asosi. JizPI Xabarnomasi, 2025(3), 95-99.

7. Wang, F., & Chen, Y. (2021). Domain shift and generalization in deep learning for industrial inspection. Engineering Applications of Artificial Intelligence, 104, 104384

8. Huang, C., et al. (2019). Detecting subtle barre defects in fabrics using deep learning and multi-scale analysis. Textile Research Journal, 89(19-20), 4015-4028.

9. Li, H., & Wang, J. (2022). Real-time defect detection using YOLOv5 on embedded systems. Journal of Manufacturing Systems, 62, 345-354.

10. Ngan, H.Y., et al. (2011). Fabric defect detection using morphological filters. Image and Vision Computing, 29(4), 278-289.

11. Zhang, R., et al. (2021). A hybrid SVM-CNN model for fabric defect classification with small datasets. IEEE Access, 9, 12345-12354.

12. Sagi, O., & Rokach, L. (2018). Ensemble learning: A survey. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 8(4), e1249.