Deep learning classification and object detection in helicopter images: Performance analysis of GoogleNet, AlexNet and YOLOv9c architectures

Year 2025, Volume: 14 Issue: 1, 1 - 1

İrem Hatice Doğan , Ozan Arslan , Ayşe Betül Tat , Burhan Şahin , Ege Erberk Uslu , İbrahim Yülüce , Orhan Dağdeviren

https://doi.org/10.28948/ngumuh.1556995

Abstract

Helicopter imaging classification and detection are crucial for autonomous navigation, military operations, search and rescue, and civil aviation management. This study utilized two helicopter image datasets, applying data augmentation techniques such as random resizing, cutting, horizontal rotation, rotation, and color adjustments, along with histogram equalization for contrast enhancement. Twenty-four helicopter classes were trained using GoogleNet and AlexNet architectures, while the YOLOv9c model was employed for object detection. The results revealed that the GoogleNet classification model achieved an 81% F1 score, and AlexNet reached 73%. In contrast, the YOLOv9c model demonstrated an average mean Average Precision (mAP) of 87%. These findings indicate that CNN architectures and YOLO are effective for helicopter image classification and detection, highlighting their potential applications in military, search and rescue, and civil aviation contexts.

Keywords

Helicopter Image Classification, Object Detection, Convolutional Neural Networks, Data Augmentation

Helikopter görüntülerinde derin öğrenme ile sınıflandırma ve nesne tespiti: GoogleNet, AlexNet ve YOLOv9c mimarilerinin performans analizi

Abstract

Helikopter görüntülerinin sınıflandırılması ve tespiti, otonom navigasyon sistemlerinin, askeri operasyonların, arama kurtarma görevlerinin ve sivil havacılık yönetiminin önemli bileşenlerindendir. Bu çalışmada helikopter görüntüleri için iki farklı veri seti kullanılmıştır. Sınıflandırma için rastgele yeniden boyutlandırma, kesme, yatay döndürme, döndürme ve renk değişiklikleri gibi veri artırma teknikleri uygulanmıştır. Görüntülerin kontrastı da histogram eşitleme yöntemi kullanılarak yeniden düzenlenmiştir. 24 sınıf helikopter veri seti GoogleNet ve AlexNet mimarileri kullanılarak eğitilmiştir. Nesne tespiti için YOLOv9c mimarisi kullanılmıştır. Deneysel sonuçlar, GoogleNet tabanlı sınıflandırma modelinin test setinde %81 F-1 skoru elde ettiğini ve AlexNet modelinde genel F1 skorunun %73 olduğunu göstermektedir. YOLOv9c modeli ise ortalama %87 mAP oranları elde etmiştir. Bu sonuçlar, bir tür derin öğrenme modeli olan CNN mimarilerinin ve YOLO nesne tespitinin helikopter görüntüsü sınıflandırma ve tespitinde iyi olduğunu gösteriyor. Çalışma, helikopter ve bileşenlerini tespit etme ve sınıflandırmada iyi performans gösteren modellerin askeri, arama kurtarma ve sivil havacılık dahil olmak üzere çeşitli alanlarda kullanılabileceğini göstermiştir.

Keywords

Helikopter Görüntü Sınıflandırma, Nesne Algılama, Evrişimli Sinir Ağları, Veri Arttırma

References

• Anagün, Y., Işik, Ş., & Çakir, F. H. (2023). Surface roughness classification of electro discharge machined surfaces with deep ensemble learning. Measurement, 215,112855. https://doi.org/10.1016/j.measurement.2023.112855
• Kurt, Z., Işık, Ş., Kaya, Z., Anagün, Y., Koca, N., & Çiçek, S. (2023). Evaluation of EfficientNet models for COVID-19 detection using lung parenchyma. Neural Computing and Applications, 35(16), 12121–12132. https://doi.org/10.1007/s00521-023-08344-z
• Can, Z., Isik, S., & Anagun, Y. (2024). CVApool: using null-space of CNN weights for the tooth disease classification. Neural Computing and Applications, 36, 16567–16579. https://doi.org/10.1007/s00521-024 09995-2
• Gozukara, G., Anagun, Y., Isik, S., Zhang, Y., & Hartemink, A. E. (2023). Predicting soil EC using spectroscopy and smartphone-based digital images. Catena, 231, 107319. https://doi.org/10.1016/j.catena.2023.107319
• Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You Only Look Once: Unified, Real-Time Object Detection. arXiv preprint arXiv:1506.02640. https://doi.org/10.48550/arXiv.1506.02640
• Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., ... & Rabinovich, A. (2015). Going deeper with convolutions. *Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)*, 1–9. https://doi.org/10.1109/CVPR.2015.7298594
• Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems, 25.
• Wang, C. Y., Yeh, I. H., & Liao, H. Y. M. (2024). Yolov9: Learning what you want to learn using programmable gradient information. arXiv preprint arXiv:2402.13616. https://doi.org/10.48550/arXiv.2402.13616
• Wu, Z. Z., Wan, S. H., Wang, X. F., Tan, M., Zou, L., Li, X. L., & Chen, Y. (2020). A benchmark data set for aircraft type recognition from remote sensing images. Applied Soft Computing, 89, 106132. https://doi.org/10.1016/j.asoc.2020.106132
• Rong, H. J., Jia, Y. X., & Zhao, G. S. (2014). Aircraft recognition using modular extreme learning machine. Neurocomputing, 128, 166-174. https://doi.org/10.1016/j.neucom.2012.12.064
• Zhao, C., Zhang, S., Luo, R., Feng, S., & Kuang, G. (2023). Scattering features spatial-structural association network for aircraft recognition in SAR images. IEEE Geoscience and Remote Sensing Letters, 20, 1-5. https://doi.org/10.1109/LGRS.2023.3280442
• Wang, Y., Chen, Y., & Liu, R. (2022). Aircraft image recognition network based on hybrid attention mechanism. Computational Intelligence and Neuroscience, 2022(1), 4189500. https://doi.org/10.1016/j.measurement.2023.113098
• Huang, X., Xu, K., Huang, C., Wang, C., & Qin, K. (2021). Multiple instance learning convolutional neural networks for fine-grained aircraft recognition. Remote Sensing, 13(24), 5132. https://doi.org/10.3390/rs13245132
• Zhao, Y., Zhao, L., Li, C., & Kuang, G. (2020). Pyramid attention dilated network for aircraft detection in SAR images. IEEE Geoscience and Remote Sensing Letters, 18(4), 662-666. https://doi.org/10.1109/LGRS.2020.2981255
• Wu, Q., Feng, D., Cao, C., Zeng, X., Feng, Z., Wu, J., & Huang, Z. (2021). Improved mask R-CNN for aircraft detection in remote sensing images. Sensors, 21(8), 2618. https://doi.org/ 10.3390/s21082618
• Zhang, F., Du, B., Zhang, L., & Xu, M. (2016). Weakly supervised learning based on coupled convolutional neural networks for aircraft detection. IEEE Transactions on Geoscience and Remote Sensing, 54(9), 5553-5563. https://doi.org/10.1109/TGRS.2016.2569141
• Alkharji, N., Almazrouei, H., Alzaabi, S., Nassif, A. B., Elsalhy, M., & Talib, M. A. (2024). Aircraft-type classification using deep learning algorithms. Proceedings of the 12th International Conference on Intelligent Systems (IS), Varna, Bulgaria, 1–6. https://doi.org/10.1109/IS61756.2024.10705261