Küçük Boyutlu İHA Tespiti için Ankraj Tabanlı YOLOv11 ve Ankrajsız FCOS’un Karşılaştırmalı Değerlendirmesi

Yıl 2025, Cilt: 7 Sayı: 2, 214 - 221

Hilal İkra Yücel , İlyas Özer , Adem Dalcalı

https://doi.org/10.46387/bjesr.1739026

Öz

Bu çalışmada, kritik altyapı ve kentsel alanlarda insansız hava araçlarının (İHA) düşük maliyetli görüntü tabanlı tespiti incelenmiştir. Histogram of Oriented Gradients (HOG), Scale-Invariant Feature Transform (SIFT) ve Local Binary Patterns (LBP) gibi geleneksel el işi özellik yöntemleri ile Support Vector Machines (SVM) ve Adaptive Boosting (AdaBoost) gibi tek aşamalı sınıflandırıcıların dinamik koşullarda küçük hedefleri tanımada yetersiz kaldığı gösterilmiştir. Bu sorunu aşmak amacıyla, “UAV Drone” Kaggle veri seti üzerinde çapa tabanlı YOLOv11 ve ankrajsız FCOS mimarileri karşılaştırılmıştır. Her iki model de 640×640 boyutlandırma, normalizasyon ve veri artırma içeren birleşik bir ön işleme hattı ve AdamW optimizatörü ile OneCycleLR öğrenme planı kullanılarak üç katlı çapraz doğrulama ile 50 epoch boyunca eğitilmiştir. Sonuçlar, YOLOv11’in yaklaşık 10 FPS hızda %66,6 mAP@[0.5–0.95], FCOS’un ise yaklaşık 20 FPS hızda %64,1 mAP ve daha düşük bellek kullanımı sağladığını ortaya koymuştur. Bu nedenle, yüksek doğruluk gerektiren araştırmalar için YOLOv11, gerçek zamanlı ve kaynak kısıtlı uygulamalar için ise FCOS türevleri önerilmektedir.

Anahtar Kelimeler

İHA Tespiti , Ankraj Tabanlı Nesne Tespiti , Tam Konvolüsyonel Tek Aşamalı

Proje Numarası

BAP-23-1004-008

Comparative Evaluation of Anchor-Based YOLOv11 and Anchor-Free FCOS for Small-Object UAV Detection

Öz

In this study, low-cost image-based detection of Unmanned Aerial Vehicle (UAVs) over critical infrastructure and urban areas is investigated. Traditional hand-crafted feature methods (Histogram of Oriented Gradients (HOG), Scale-Invariant Feature Transform (SIFT), Local Binary Patterns (LBP)) and single-stage classifiers (Support Vector Machines (SVM), Adaptive Boosting (AdaBoost)) are shown to struggle with small targets under dynamic conditions. To address this, anchor-based YOLOv11 and anchor-free Fully Convolutional One-Stage (FCOS) architectures are compared on the “UAV Drone” Kaggle dataset. Both models use a unified preprocessing pipeline (resize to 640×640, normalization, data augmentation) and three-fold cross-validation for 50 epochs with the AdamW optimizer and a OneCycleLR schedule. Results reveal that YOLOv11 achieves 66.6 % mAP@[0.5–0.95] at ~10 FPS, while FCOS attains 64.1 % mAP at ~20 FPS with lower memory use. Thus, YOLOv11 is recommended for high-accuracy research, and FCOS variants for real-time, resource-constrained applications.

Anahtar Kelimeler

UAV Detection , Anchor-Based Object Detection , Fully Convolutional One-Stage.

Etik Beyan

The authors of the articles declare that they have no conflict of interest

Destekleyen Kurum

This work was supported by the Bandırma Onyedi Eylül University Scientific Research Projects Coordination Unitunder Project BAP-23-1004-008.

Proje Numarası

BAP-23-1004-008

Kaynakça

• Allied Market Research. Drone Service Market Overview, 2030. Acsses: June 2025.
• H.B. Srivastava, V. Kumar, H. K. Verma, S.S. Sundaram, “Image Pre-processing Algorithms for Detection of Small/Point Airborne Targets,” Defence Science Journal, vol. 59, no. 2, pp. 166-174, 2009.
• S. Al-Emadi, A. Al-Ali, A. Mohammad, A. Al-Ali, “Audio Based Drone Detection and Identification using Deep Learning,” 15th International Wireless Communications & Mobile Computing Conference, Morocco, pp. 459-464, 2019.
• N. Jiang, K. Wang, X. Peng, X. Yu, Q. Wang, J. Xing, G. Li, G. Guo, Q. Ye, J. Jiao, J. Zhao, Z. Han, “Anti-UAV: A Large-Scale Benchmark for Vision-Based UAV Tracking,” IEEE Transactions on Multimedia, vol. 25, pp. 486- 500, 2023.
• R. Girshick, “Fast R-CNN.” IEEE International Conference on Computer Vision, Santiago, 2015.
• S. Ren, K. He, R. Girshick, J. Sun, “Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 39, no. 6, pp. 1137-1149, 1 June 2017.
• J. Redmon, S. Divvala, R. Girshick, A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 779-788, 2016.
• J. Redmon, A. Farhadi, “YOLOv3: An Incremental Improvement,” arXiv 1804.02767, 2018.
• A. Bochkovskiy, C.Y. Wang, H.Y.M. Liao, “YOLOv4: Optimal Speed and Accuracy of Object Detection,” arXiv 2004.10934, 2020.
• Y. Zhao, Z. Ju, T. Sun, F. Dong, J. Li, R. Yang, Q. Fu, C. Lian, P. Shan, “TGC-YOLOv5: An Enhanced YOLOv5 Drone Detection Model Based on Transformer, GAM & CA Attention Mechanism,” Drones, vol. 7, no. 7, 446, 2023.
• S. Singha, B. Aydin, “Automated Drone Detection Using YOLOv4,” Drones, vol. 5, no. 3, 95, 2021.
• H. Liang, J. Yang, M. Shao, “FE-RetinaNet: Small Target Detection with Parallel Multi-Scale Feature Enhancement,” Symmetry, vol. 13, no. 5, 950, 2021.
• A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, J. Uszkoreit, N. Houlsby, “An Image is Worth 16 × 16 Words: Transformers for image recognition at scale,” The Ninth International Conference on Learning Representations, 2021.
• Z. Liu, Y. Lin, Y. Cao, H. Hu, Y. Wei, Z. Zhang, S. Lin, B. Guo, “Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows,” Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 10012-10022, 2021.
• S. Jamil, M.S. Abbas, A.M. Roy, “Distinguishing Malicious Drones Using Vision Transformer,” AI, vol. 3, no. 2, pp. 260-273, 2022.
• H. Guo, X. Lin, S. Zhao, “YOLOMG: Vision-based Drone-to-Drone Detection with Appearance and Pixel-Level Motion Fusion,” arXiv 2503.07115, 2025.
• X. Luo, Y. Wu, F. Wang, “Target Detection Method of UAV Aerial Imagery Based on Improved YOLOv5,” Remote Sensing, vol. 14, no. 19, 5063, 2022.
• https://www.kaggle.com/datasets/dasmehdixtr/drone-dataset-uav
• Z. Tian, C. Shen, H. Chen, & T. He, “Fcos: Fully convolutional one-stage object detection,” IEEE/CVF international conference on computer vision, pp. 9627-9636, 2019.
• Z. Tian, C. Shen, H. Chen, & T. He, “FCOS: A simple and strong anchor-free object detector,” IEEE Transactions On Pattern Analysis and Machine Intelligence, vol. 44, no. 4, 1922-1933, 2020.
• Y. H. Lee, & H. J. Kim, “Comparative Analysis of YOLO Series (from V1 to V11) and Their Application in Computer Vision,” Journal of the Semiconductor & Display Technology, vol. 23, no. 4, pp. 190-198, 2024.
• M. Mao, & M. Hong, “YOLO object detection for real-time fabric defect inspection in the textile industry: A review of YOLOv1 to YOLOv11,” Sensors, vol. 25, no. 7, 2270, 2025.
• M. A. Bayram, İ. Özer, & F. Temurtaş, “Deep learning methods for autism spectrum disorder diagnosis based on fMRI images,” Sakarya University Journal of Computer and Information Sciences, vol. 4, no. 1, pp. 142-155, 2021.
• I. Ozer, C. K. Ozer, A. C. Karaca, K. Gorur, I. Kocak, & O. Cetin, “Species-level microfossil identification for globotruncana genus using hybrid deep learning algorithms from the scratch via a low-cost light microscope imaging,” Multimedia Tools and Applications, vol. 82, no. 9, pp. 13689-13718, 2023.