개선된 YOLO11 접근을 통한 지상무기체계의 객체 탐지와 예측: PLZ-05 자주포의 작전운용 환경과 무인항공기 촬영의 가정을 초점으로
Object prediction and detection of ground-based weapon with an improved YOLO11 approach: Focusing on assumptions underlying operational environments and UAV-captured features related to PLZ-05 Self-Propelled Howitzer
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
지상 무기체계의 탐지와 분석에서 UAV 활용한 탐지 기술은 실시간으로 전술적 의사결정 지원에서 중요하다. 기존 연구는 UAV 기반 군사 객체 탐지와 분류 성능 향상에 중점을 두었으나, 본 연구는 UAV 운용 환경을 가정하여 군사 객체를 탐지하고, 탐지된 포신의 고각과 방위각을 분석하는 새로운 체계를 제안한다. 분석방법으로 객체 탐지는 YOLO11 모델과 BCEF 손실 함수를 적용하여 탐지 성능을 강화하였으며, 복잡한 전장 조건을 모사한 생성된 합성 이미지를 활용하여 다양한 환경에서 시스템을 검증하였다. 강우, 복잡한 지형, 저조도 환경과 같은 도전적인 조건에서도 제안된 모델은 높은 탐지 정확도와 신뢰성이 유지되었다. 본 연구는 UAV 기반 객체 탐지 기술의 활용 가능성을 제시하여 정찰⋅식별 기능에서 전술적 의사결정 지원 도구로 확장 가능성을 제기한다. 향후 연구에서는 실제 UAV 운용 환경 조건을 구현하는 연구설계를 통해 실시간 데이터 수집과 분석 체계를 가정한 제안 모델의 성능을 검증하는 추가 실험연구가 필요하다.
Abstract
The utilization of UAV-based detection technologies in ground weapon system analysis plays a crucial role in supporting real-time tactical decision-making. While previous studies have primarily focused on improving the detection and classification performance of military objects using UAVs, the current study proposes a novel system that not only detects military objects in simulated UAV operational environments but also analyzes the elevation and azimuth angles of detected gun barrels. For object detection, the YOLO11 model was employed in conjunction with the BCEF loss function to enhance detection performance. The proposed system was validated across various environments using synthetically generated images simulating complex battlefield conditions, including rain, challenging terrain, and low-light environments. Even under these adverse conditions, the model demonstrated high detection accuracy and reliability. This study highlights the potential of UAV-based object detection technology as a tactical decision-making support tool, extending its utility from reconnaissance and identification to broader operational roles. Future research need to further evaluate the performance of the proposed model with experimental validation in real-world UAV operational conditions, emphasizing real-time data collection and analysis frameworks.
참고문헌 (22)
- Barros, J. P., "Key features and applications of military drones : a case study from the Portuguese military ground forces", Journal of Defense Analytics and Logistics, 8(2), pp.179~201 (2024)[DOI]
- Chang, J. O., "A Study on the Prediction of Future Warfare Changes and the Development Direction of the R. O. K. Army : Focusing on the relationship between war components and forms of war", Korean Journal of Military Art and Science, 78(3), pp.1~28 (2022)[DOI]
- "On denoising autoencoders trained to minimise binary cross-entropy" (2017)[URL]
- Gupta, C., "A novel finetuned YOLOv6 transfer learning model for real-time object detection", Journal of Real-Time Image Processing, 20, pp.1~19 (2023)[DOI]
- Hwang, J. Y., "A Study on Transfer Learning-based Ensemble Model for Classification of Maneuver Weapon System", Journal of Korean Institute of Information Technology, 19(12), pp.1~10 (2021)[DOI]
- Kim, D. B., "Development and Technological Strategy of Intelligent Ground Manned-Unmanned Composite Combat Systems: Focused on Future Ground Combat Systems", Journal of the Korea Association of Defense Industry Studies, Journal of Korean Institute of Information Technology, 30(3), pp.59~67 (2023)[DOI]
- Kim, M. G., "A Study on the Devel opment Direction of Next Generation Drone Through Analysis of the Role of Drones Used in Warfare : From Gulf War to Ukraine War", Journal of the Korea Academia-Industrial cooperation Society, 24(10), pp.656~664 (2023)[DOI]
- Kim, M. W., "Determining the Scal e of Mil itary Echel on Using Deep LearningBased Object Detection and Tracking Techniques", Journal of Korea Academia-Industrial cooperation Society, 24(5), pp.441~449 (2023)[DOI]
- Lee, S. H., "Effects of the Technological and Individual Characteristics of Army Tactical Command Information System on Situation Awareness and Decision Making", Journal of The Korean Operations Research and Management Science Society, 45(2), pp.25~42 (2020)[DOI]
- Lee, W. H., "Enabling reliable UAV control by utilizing multiple protocols and paths for transmitting duplicated control packets", Sensors, 21(9), pp.3295~ (2021)[DOI]
- Lim, S. G, "Identifications and Evaluation of Tank Nationality using YOLO Algorithm", KIISE Transactions on Computing Practices, 27(12), pp.555~562 (2021)[DOI]
- "Focal loss for dense object detection" (2017)[URL]
- Pan, S. J., "A survey on transfer learning", IEEE Transactions on Knowledge and Data Engineering, 22(10), pp.1345~1359 (2010)[DOI]
- Park, S. J., "Classification Type of Weapon Using Artificial Intelligence for Counter-battery RadarPaper Title", Journal of IKEEE, 24(4), pp.921~930 (2020)[DOI]
- "YOLOv3: An Incremental Improvement" (2018)[URL]
- Redmon, J., "You only look once : Unified, real-time object detection", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.779~788 (2016)
- Shorten, C., "A survey on image data augmentation for deep learning", Journal of Big Data, 6(60) (2019)[DOI]
- Stocker, C., "Review of the current state of UAV regulations", Remote Sensing, 9(5), pp.459~ (2017)[DOI]
- Tan, M., "EfficientNet: Rethinking model scaling for convolutional neural networks", Proceedings of the 36th International Conference on Machine Learning, pp.6105~6114 (2019)
- Vasavi, S., "Onboard processing of drone imagery for military vehicles classification using enhanced YOLOv5", Journal of Advances in Information Technology, 14(6), pp.1221~1229 (2023)
- "What is YOLOv8: An in-depth exploration of the internal features of the next-generation object detector" (2024)[URL]
- Yuk, S. M., "Research Trends on Deep Learning-based Image Detection of Weapon Systems in National Defense", Journal of Korea Academia-Industrial cooperation Society Spring conference, pp.1~3 (2023)