Detection of Military Aircraft Using YOLO and Transformer-Based Object Detection Models in Complex Environments

Fatih Şengül; Kemal Adem

Article Details

Detection of Military Aircraft Using YOLO and Transformer-Based Object Detection Models in Complex Environments

Authors:

Fatih Şengül | Kemal Adem

Publisher:

Gazi University

Main Topics:

Derin Öğrenme | Yapay Zeka (Diğer)

Journal:

Bilişim Teknolojileri Dergisi Vol.18 Issue 1, pp.85-97

Document Type:

Article

Publication Date:

1/2025

Document Language:

English

Document ID (Org No.):

2001339773

pISSN:

1307-9697

Keywords:

Abstract:

 Computer vision and deep learning techniques are widely applied in object detection tasks across various domains, including defense technologies. Accurate and efficient detection of military aircraft plays a critical role in strengthening air defense systems and enabling effective strategic decision-making. This study evaluates the performance of YOLOv7, YOLOv8, and RT-DETR models in detecting military aircraft using a dataset consisting of 19.514 images spanning 43 aircraft models. The dataset incorporates images captured from various angles and diverse backgrounds, such as urban, rural, and coastal areas, ensuring realistic testing conditions. However, class imbalance is observed, with certain aircraft models, such as the F14 and F16, being more represented than others, which may affect model generalization. To address these challenges, hyperparameters were optimized, and performance metrics, including mean Average Precision mAP and recall, were analyzed. Experimental results show that YOLOv8 achieved 94 mAP and 88.1 recall, YOLOv7 reached 90.2 mAP and 82.7 recall, while RT-DETR demonstrated consistent performance with 92.7 mAP and 90.4 recall. These findings highlight the strengths and limitations of the evaluated models and provide inferences for improving detection systems in defense applications.

 Computer vision and deep learning techniques are widely applied in object detection tasks across various domains, including defense technologies. Accurate and efficient detection of military aircraft plays a critical role in strengthening air defense systems and enabling effective strategic decision-making. This study evaluates the performance of YOLOv7, YOLOv8, and RT-DETR models in detecting military aircraft using a dataset consisting of 19.514 images spanning 43 aircraft models. The dataset incorporates images captured from various angles and diverse backgrounds, such as urban, rural, and coastal areas, ensuring realistic testing conditions. However, class imbalance is observed, with certain aircraft models, such as the F14 and F16, being more represented than others, which may affect model generalization. To address these challenges, hyperparameters were optimized, and performance metrics, including mean Average Precision mAP and recall, were analyzed. Experimental results show that YOLOv8 achieved 94 mAP and 88.1 recall, YOLOv7 reached 90.2 mAP and 82.7 recall, while RT-DETR demonstrated consistent performance with 92.7 mAP and 90.4 recall. These findings highlight the strengths and limitations of the evaluated models and provide inferences for improving detection systems in defense applications.

 Computer vision and deep learning techniques are widely applied in object detection tasks across various domains, including defense technologies. Accurate and efficient detection of military aircraft plays a critical role in strengthening air defense systems and enabling effective strategic decision-making. This study evaluates the performance of YOLOv7, YOLOv8, and RT-DETR models in detecting military aircraft using a dataset consisting of 19.514 images spanning 43 aircraft models. The dataset incorporates images captured from various angles and diverse backgrounds, such as urban, rural, and coastal areas, ensuring realistic testing conditions. However, class imbalance is observed, with certain aircraft models, such as the F14 and F16, being more represented than others, which may affect model generalization. To address these challenges, hyperparameters were optimized, and performance metrics, including mean Average Precision mAP and recall, were analyzed. Experimental results show that YOLOv8 achieved 94 mAP and 88.1 recall, YOLOv7 reached 90.2 mAP and 82.7 recall, while RT-DETR demonstrated consistent performance with 92.7 mAP and 90.4 recall. These findings highlight the strengths and limitations of the evaluated models and provide inferences for improving detection systems in defense applications.

 Bilgisayarla görme ve derin öğrenme teknikleri, savunma teknolojileri de dahil olmak üzere çeşitli alanlardaki nesne algılama görevlerinde yaygın olarak uygulanmaktadır. Savaş uçaklarının doğru ve verimli bir şekilde tespit edilmesi, hava savunma sistemlerinin güçlendirilmesinde ve etkili stratejik karar alma süreçlerinin desteklenmesinde kritik bir rol oynamaktadır. Bu çalışmada, 43 uçak modelini kapsayan 19.514 görüntüden oluşan bir veri kümesi kullanılarak YOLOv7, YOLOv8 ve RT-DETR modellerinin savaş uçaklarını tespit etme performansı değerlendirilmektedir. Veri kümesi, çeşitli açılardan ve kentsel, kırsal ve kıyı alanları gibi farklı arka planlardan çekilen görüntüleri içermekte ve gerçekçi test koşulları sağlamaktadır. Bununla birlikte, F14 ve F16 gibi belirli uçak modellerinin diğerlerine göre daha fazla temsil edildiği ve model genellemesini etkileyebilecek sınıf dengesizliği gözlemlenmiştir. Bu zorlukların üstesinden gelmek için hiperparametreler optimize edilmiş ve ortalama Ortalama Hassasiyet mAP ve geri çağırma dahil olmak üzere performans ölçütleri analiz edilmiştir. Deneysel sonuçlar, YOLOv8in 94 mAP ve 88,1 geri çağırma, YOLOv7nin 90,2 mAP ve 82,7 geri çağırma değerlerine ulaştığını, RT-DETRnin ise 92,7 mAP ve 90,4 geri çağırma ile tutarlı bir performans sergilediğini göstermektedir. Bu bulgular, değerlendirilen modellerin güçlü yönlerini ve kısıtlamalarını vurgulamakta ve savunma uygulamalarında tespit sistemlerinin iyileştirilmesi için çıkarımlar sağlamaktadır.

Article Details

Detection of Military Aircraft Using YOLO and Transformer-Based Object Detection Models in Complex Environments

Authors:

Fatih Şengül | Kemal Adem

Publisher:

Gazi University

Main Topics:

Derin Öğrenme | Yapay Zeka (Diğer)

Journal:

Bilişim Teknolojileri Dergisi Vol.18 Issue 1, pp.85-97

Document Type:

Article

Publication Date:

1/2025

Document Language:

English

Document ID (Org No.):

2001339773

pISSN:

1307-9697

military aircraft detection

YOLOv7

YOLOv8

RT-DETRsavaş uçağı tespiti

YOLOv7

YOLOv8

RT-DETR

Issues

Article Details

Authors: Fatih Şengül | Kemal Adem

Publisher: Gazi University

Main Topics: Derin Öğrenme | Yapay Zeka (Diğer)

Journal: Bilişim Teknolojileri Dergisi Vol.18 Issue 1, pp.85-97

Document Type: Article

Publication Date: 1/2025

Document Language: English

Document ID (Org No.): 2001339773

pISSN: 1307-9697

Issues

Authors:

Fatih Şengül | Kemal Adem

Publisher:

Gazi University

Main Topics:

Derin Öğrenme | Yapay Zeka (Diğer)

Journal:

Bilişim Teknolojileri Dergisi Vol.18 Issue 1, pp.85-97

Document Type:

Article

Publication Date:

1/2025

Document Language:

English

Document ID (Org No.):

2001339773

pISSN:

1307-9697