Performance Analysis of YOLO11 Models in PCB Defect Detection Tasks

Mehmet Dayıoğlu; Ali Kemal Eyüboğlu; Ridvan Unal

Article Details

Performance Analysis of YOLO11 Models in PCB Defect Detection Tasks

Authors:

Mehmet Dayıoğlu | Ali Kemal Eyüboğlu | Ridvan Unal

Publisher:

Dokuz Eylul University

Main Topics:

Elektronik

Journal:

Kuzey Ege Teknik Bilimler ve Teknoloji Dergisi Vol.2 Issue 1, pp.33-50

Document Type:

Article

Publication Date:

6/2025

Document Language:

English

Document ID (Org No.):

2001340839

pISSN:

3062-1038

Keywords:

Abstract:

 Printed Circuit Board PCB defect detection is critical in electronics manufacturing, as undetected faults can lead to severe quality control issues. Recent advancements in deep learning, particularly object detection models, have significantly improved inspection systems accuracy and speed. This study explores the performance of the YOLO11 You Only Look Once version 11 object detection architecture on a multi-class PCB defect dataset. Five YOLO11 variantsYOLO11n, YOLO11s, YOLO11m, YOLO11l, and YOLO11xwere trained and evaluated under identical conditions using high-resolution images containing six defect types. Metrics such as mAP50, mAP50-95, and FPS were used for evaluation. Results demonstrate that YOLO11l achieved the highest mAP50-95 of 0.551, while YOLO11n achieved up to 166 Frame Per Second FPS on an NVIDIA A100 GPU, confirming its real-time capability. Comparative analysis against state-of-the-art models confirms that YOLO11 variants offer an effective trade-off between accuracy and efficiency. This study positions YOLO11 as a strong candidate for real-time PCB inspection systems.

 Printed Circuit Board PCB defect detection is critical in electronics manufacturing, as undetected faults can lead to severe quality control issues. Recent advancements in deep learning, particularly object detection models, have significantly improved inspection systems accuracy and speed. This study explores the performance of the YOLO11 You Only Look Once version 11 object detection architecture on a multi-class PCB defect dataset. Five YOLO11 variantsYOLO11n, YOLO11s, YOLO11m, YOLO11l, and YOLO11xwere trained and evaluated under identical conditions using high-resolution images containing six defect types. Metrics such as mAP50, mAP50-95, and FPS were used for evaluation. Results demonstrate that YOLO11l achieved the highest mAP50-95 of 0.551, while YOLO11n achieved up to 166 Frame Per Second FPS on an NVIDIA A100 GPU, confirming its real-time capability. Comparative analysis against state-of-the-art models confirms that YOLO11 variants offer an effective trade-off between accuracy and efficiency. This study positions YOLO11 as a strong candidate for real-time PCB inspection systems.

 Printed Circuit Board PCB defect detection is critical in electronics manufacturing, as undetected faults can lead to severe quality control issues. Recent advancements in deep learning, particularly object detection models, have significantly improved inspection systems accuracy and speed. This study explores the performance of the YOLO11 You Only Look Once version 11 object detection architecture on a multi-class PCB defect dataset. Five YOLO11 variantsYOLO11n, YOLO11s, YOLO11m, YOLO11l, and YOLO11xwere trained and evaluated under identical conditions using high-resolution images containing six defect types. Metrics such as mAP50, mAP50-95, and FPS were used for evaluation. Results demonstrate that YOLO11l achieved the highest mAP50-95 of 0.551, while YOLO11n achieved up to 166 Frame Per Second FPS on an NVIDIA A100 GPU, confirming its real-time capability. Comparative analysis against state-of-the-art models confirms that YOLO11 variants offer an effective trade-off between accuracy and efficiency. This study positions YOLO11 as a strong candidate for real-time PCB inspection systems.

 Baskılı Devre Kartı PCB hata tespiti, elektronik üretiminde kritik bir süreçtir; tespit edilemeyen hatalar ciddi kalite kontrol problemlerine yol açabilir. Derin öğrenme ve özellikle nesne tespiti modellerindeki son gelişmeler, denetim sistemlerinin doğruluk ve hızında önemli iyileştirmeler sağlamıştır. Bu çalışma, çok sınıflı bir PCB arıza tespit veri kümesi üzerinde YOLO11 You Only Look Once versiyon 11 nesne tespiti mimarisinin performansını incelemektedir. YOLO11 ailesine ait beş farklı varyantYOLO11n, YOLO11s, YOLO11m, YOLO11l ve YOLO11xaltı farklı arıza türü içeren yüksek çözünürlüklü görüntüler kullanılarak, aynı eğitim koşulları altında değerlendirilmiştir. Değerlendirme için mAP50, mAP50-95 ve FPS gibi metrikler kullanılmıştır. Sonuçlar, YOLO11l modelinin 0.551 mAP50-95 ile en yüksek doğruluk değerine ulaştığını, YOLO11n modelinin ise NVIDIA A100 GPU üzerinde saniyede 166 kareye FPS kadar gerçek zamanlı performans sergilediğini göstermektedir. Yapılan karşılaştırmalar, YOLO11 ailesinin doğruluk ve verimlilik arasında etkili bir denge sunduğunu doğrulamaktadır. Bu çalışma, YOLO11 mimarisinin gerçek zamanlı PCB denetim sistemleri için güçlü bir aday olduğunu ortaya koymaktadır.

Article Details

Performance Analysis of YOLO11 Models in PCB Defect Detection Tasks

Authors:

Mehmet Dayıoğlu | Ali Kemal Eyüboğlu | Ridvan Unal

Publisher:

Dokuz Eylul University

Main Topics:

Elektronik

Journal:

Kuzey Ege Teknik Bilimler ve Teknoloji Dergisi Vol.2 Issue 1, pp.33-50

Document Type:

Article

Publication Date:

6/2025

Document Language:

English

Document ID (Org No.):

2001340839

pISSN:

3062-1038

PCB Defect Detection

Deep Learning

YOLO11PCB Hata Tespiti

Derin Öğrenme

YOLO11

Issues

Article Details

Performance Analysis of YOLO11 Models in PCB Defect Detection Tasks

Authors: Mehmet Dayıoğlu | Ali Kemal Eyüboğlu | Ridvan Unal

Publisher: Dokuz Eylul University

Main Topics: Elektronik

Journal: Kuzey Ege Teknik Bilimler ve Teknoloji Dergisi Vol.2 Issue 1, pp.33-50

Document Type: Article

Publication Date: 6/2025

Document Language: English

Document ID (Org No.): 2001340839

pISSN: 3062-1038

PCB Defect Detection

Deep Learning

YOLO11PCB Hata Tespiti

Derin Öğrenme

YOLO11

Issues

Authors:

Mehmet Dayıoğlu | Ali Kemal Eyüboğlu | Ridvan Unal

Publisher:

Dokuz Eylul University

Main Topics:

Elektronik

Journal:

Kuzey Ege Teknik Bilimler ve Teknoloji Dergisi Vol.2 Issue 1, pp.33-50

Document Type:

Article

Publication Date:

6/2025

Document Language:

English

Document ID (Org No.):

2001340839

pISSN:

3062-1038