Olga Leshchenko; Yuriy Kravchenko; Oksana Herasymenko; Nataliia Dakhno; Оlexandr Makhovych; Serhii Stavytskyi; Denisa Macekova

A Three-Level Model for Detecting and Identifying Aviation Objects using Deep Learning

PDF (638KB), PP.277-289

Views: 0 Downloads: 0

Author(s)

Olga Leshchenko ^1,* Yuriy Kravchenko ¹ Oksana Herasymenko ¹ Nataliia Dakhno ¹ Оlexandr Makhovych ¹ Serhii Stavytskyi ¹ Denisa Macekova ²

1. Taras Shevchenko National University of Kyiv, 60, Volodymyrska Str., Kyiv, 01033, Ukraine

2. Department of Informatics, Faculty of Management Science and Informatics, University of Zilina (Slovakia)

* Corresponding author.

DOI: https://doi.org/10.5815/ijwmt.2026.03.18

Received: 2 Mar. 2026 / Revised: 1 Apr. 2026 / Accepted: 23 Apr. 2026 / Published: 8 Jun. 2026

Index Terms

Object Detection, Artificial Intelligence, Deep Learning, Image Recognition, Three-Level Model, Computer Vision, Neural Networks, Yolov8n, SSD, Faster R-CNN, OCR

Abstract

This paper presents an intelligent software solution for object identification in images using deep learning models, designed for automated interpretation of monitoring results of aviation objects and infrastructure. The proposed approach addresses the growing demand for enhanced flight safety and improved efficiency of aviation operations. To meet this demand, a three-level model is proposed: Level 1 performs object detection, Level 2 provides optical character recognition (OCR) and text normalization, and Level 3 implements fuzzy matching with an object database. Based on comparative testing of detection models, YOLOv8n was selected as the core of the three-level architecture, providing an optimal balance between real-time processing speed and detection accuracy. A detailed analysis of model architectures revealed specific advantages and limitations in identifying monitoring results from image data. Training on a specialized dataset and subsequent testing confirmed the high efficiency of the proposed solution and its ability to reliably localize objects even under challenging visual conditions such as shadows, glare, and partial occlusion. The

obtained results demonstrate the significant potential of the proposed intelligent solution for extending computer vision capabilities in the monitoring of aviation objects and infrastructure. The experimental results also confirm the effectiveness of the OCR and fuzzy matching modules in improving object identification accuracy under real-world conditions.

Cite This Paper

Olga Leshchenko, Yuriy Kravchenko, Oksana Herasymenko, Nataliia Dakhno, Оlexandr Makhovych, Serhii Stavytskyi, Denisa Macekova, "A Three-Level Model for Detecting and Identifying Aviation Objects using Deep Learning", International Journal of Wireless and Microwave Technologies(IJWMT), Vol.16, No.3, pp. 277-289, 2026. DOI:10.5815/ijwmt.2026.03.18

Reference

[1]M. Farhadmanesh, A. Rashidi, and N. Marković, "General aviation aircraft identification at non-towered airports using a two-step computer vision-based approach," IEEE Access, vol. 10, pp. 48778–48791, 2022, doi: 10.1109/ACCESS.2022.3172963.
[2]V. Tetarwal, M. Kaur, and S. Kumar, "A comprehensive review on computer vision analysis of aerial data," Eng. Appl. Artif. Intell., vol. 156, p. 111206, 2025.
[3]P. Li and H. Li, "Research on FOD detection for airport runway based on YOLOv3," in Proc. 39th Chinese Control Conf. (CCC), Shenyang, China, 2020, pp. 7096–7099, doi: 10.23919/CCC50068.2020.9188724.
[4]S. Zhai, D. Shang, S. Wang, and S. Dong, "DF-SSD: An improved SSD object detection algorithm based on DenseNet and feature fusion," IEEE Access, vol. 8, pp. 24344–24357, 2020.
[5]P. Bharati and A. Pramanik, "Deep learning techniques—R-CNN to mask R-CNN: a survey," in Proc. Int. Conf. Comput. Intell. Pattern Recognit. (CIPR), 2019, pp. 657–668.
[6]Y. Shao, Z. Sun, A. Tan, and T. Yan, "Efficient three-dimensional point cloud object detection based on improved Complex-YOLO," Front. Neurorobot., vol. 17, p. 1092564, 2023.
[7]F. Yu, "YOLO, Faster R-CNN and SSD for cloud detection," Appl. Comput. Eng., vol. 37, pp. 239–247, 2024.
[8]M. Bobko and A. Sivokhin, "Method for identifying combat vehicles based on YOLO," Sci. Tech. Collection "Information Protection," no. 4, pp. 57–63, 2021.
[9]V. Petrivskyi et al., "Development of a modification of the method for constructing energy-efficient sensor networks using static and dynamic sensors," Eastern-Eur. J. Enterprise Technol., vol. 1, no. 9 (115), pp. 15–23, 2022, doi: 10.15587/1729-4061.2022.252988.
[10]T. Tran, M. D. Smith, and P. Nguyen, "Application of YOLO in automated retail shelf monitoring," ACM Trans. Intell. Syst., vol. 3, no. 1, pp. 1–18, 2023, doi: 10.1145/3581957.
[11]S. Bhaggiaraj, M. Priyadharsini, K. Karuppasamy, and R. Snegha, "Deep-learning-based self-driving cars using computer vision," in Proc. Int. Conf. Networking and Communications (ICNWC), Chennai, India, 2023, pp. 1–9, doi: 10.1109/ICNWC57852.2023.10127448.
[12]K. O. Stanley, "Efficient evolution of neural networks through complexification," 2004. [Online]. Available: https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=e158baaec08a54313c74ea0e1af1b72a6863406c
[13]X. Yao, "Evolving artificial neural networks," Proc. IEEE, vol. 87, no. 9, pp. 1423–1447, 1999.
[14]Y. Kravchenko, H. Dakhno, O. Leshchenko, A. Dudnik, and A. Miroshnyk, "Development of the intelligent control system of an unmanned car," CEUR Workshop Proc., vol. 3806, pp. 276–287, 2024. [Online]. Available: https://ceur-ws.org/Vol-3806/S_37_Dakhno.pdf
[15]Y. Kravchenko, O. Leshchenko, A. Dudnik, N. Dakhno, and H. Dakhno, "Machine learning systems for analyzing and predicting user behavior," in Proc. XI Int. Sci. Conf. Information Technology and Implementation (IT&I), CEUR Workshop Proc., 2024, pp. 138–152. [Online]. Available: http://ceur-ws.org/Vol-3909/Paper_11.pdf
[16]O. Kurchenko, K. Dukhnovska, O. Kovtun, A. Nikolaienko, and I. Yurchuk, "The Method of Effective Numerical Solution of the System of Equations of Thermal Conductivity," in Proc. Int. Sci. Conf. Information Technology and Implementation (IT&I Workshops), 2023, pp. 206–214.
[17]P. Prystavka, K. Dukhnovska, O. Kovtun, O. Leshchenko, O. Cholyshkina, and A. Zhultynska, "Devising information technology for determining the redundant information content of a digital image," Eastern-Eur. J. Enterprise Technol., vol. 6, no. 2 (114), pp. 59–70, 2021, doi: 10.15587/1729-4061.2021.248698.
[18]O. Trush, A. Dudnik, M. Trush, O. Leshchenko, K. Shmat, and R. Mykolaichuk, "Mask mode monitoring systems using IT technologies," in Proc. IEEE 4th Int. Conf. Advanced Trends Inf. Theory (ATIT), Dec. 2022, pp. 219–224.
[19]Ultralytics YOLO. [Online]. Available: https://github.com/ultralytics/ultralytics (accessed Mar. 23, 2026).
[20]M. K. A. Maaroof and M. S. Bouhlel, "Real-time object detection using YOLO-8 model: A drone-based approach," J. Wirel. Mob. Netw. Ubiquitous Comput. Dependable Appl., vol. 16, no. 1, pp. 190–204, 2025.
[21]A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, "YOLOv4: Optimal Speed and Accuracy of Object Detection," arXiv:2004.10934, 2020.
[22]NVIDIA, "SSD model on PyTorch Hub." [Online]. Available: https://pytorch.org/hub/nvidia_deeplearningexamples_ssd/ (accessed Mar. 2026).
[23]PyTorch, "Faster R-CNN model." [Online]. Available: https://pytorch.org/vision/main/models/faster_rcnn.html (accessed Mar. 2026).
[24]Google, "Open Images Dataset." [Online]. Available: https://storage.googleapis.com/openimages/web/index.html (accessed Mar. 2026).
[25]J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, "You Only Look Once: Unified, Real-Time Object Detection," in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2016, pp. 779–788.
[26]Y. Wen, X. Gao, L. Luo, and J. Li, "Improved YOLOv8-based target-precision detection algorithm for train-wheel-tread defects," Sensors, vol. 24, no. 11, p. 3477, 2024.
[27]T. Wu and Y. Dong, "YOLO-SE: Improved YOLOv8 for remote-sensing object detection and recognition," Appl. Sci., vol. 13, no. 24, p. 12977, 2023.

International Journal of Wireless and Microwave Technologies (IJWMT)