The Journal of the Convergence on Culture Technology (문화기술의 융합)

The International Promotion Agency of Culture Technology (국제문화기술진흥원)
권호 표지
DOI QR코드

DOI QR Code

Performance Comparison of CNN-Based Image Classification Models for Drone Identification System

드론 식별 시스템을 위한 합성곱 신경망 기반 이미지 분류 모델 성능 비교

  • 김영완 (해군 제2함대사령부) ;
  • 조대균 (국군방첩사령부 국방보안연구소) ;
  • 박건우 (대전대학교 컴퓨터공학과)
  • Received : 2024.04.22
  • Accepted : 2024.06.15
  • Published : 2024.07.31
Download PDF
⟨ Previous Next ⟩

Abstract

Recent developments in the use of drones on battlefields, extending beyond reconnaissance to firepower support, have greatly increased the importance of technologies for early automatic drone identification. In this study, to identify an effective image classification model that can distinguish drones from other aerial targets of similar size and appearance, such as birds and balloons, we utilized a dataset of 3,600 images collected from the internet. We adopted a transfer learning approach that combines the feature extraction capabilities of three pre-trained convolutional neural network models (VGG16, ResNet50, InceptionV3) with an additional classifier. Specifically, we conducted a comparative analysis of the performance of these three pre-trained models to determine the most effective one. The results showed that the InceptionV3 model achieved the highest accuracy at 99.66%. This research represents a new endeavor in utilizing existing convolutional neural network models and transfer learning for drone identification, which is expected to make a significant contribution to the advancement of drone identification technologies.

최근 전장에서의 드론 활용이 정찰뿐만 아니라 화력 지원까지 확장됨에 따라, 드론을 조기에 자동으로 식별하는 기술의 중요성이 더욱 증가하고 있다. 본 연구에서는 드론과 크기 및 외형이 유사한 다른 공중 표적들인 새와 풍선을 구분할 수 있는 효과적인 이미지 분류 모델을 확인하기 위해, 인터넷에서 수집한 3,600장의 이미지 데이터셋을 사용하고, 세 가지 사전 학습된 합성곱 신경망 모델(VGG16, ResNet50, InceptionV3)의 특징 추출기능과 추가 분류기를 결합한 전이 학습 접근 방식을 채택하였다. 즉, 가장 우수한 모델을 확인하기 위해 세 가지 사전 학습된 모델(VGG16, ResNet50, InceptionV3)의 성능을 비교 분석하였으며, 실험 결과 InceptionV3 모델이 99.66%의 최고 정확도를 나타냄을 확인하였다. 본 연구는 기존의 합성곱 신경망 모델과 전이 학습을 활용하여 드론을 식별하는 새로운 시도로써, 드론 식별 기술의 발전에 크게 기여 할 것으로 기대된다.

Keywords

References

  1. K.J. Lee, "A Study on the Development Plan and Analysis of Military Drone Standardization Promotion," Journal of Review of Korean Military Studies, Vol. 12, No. 1, pp.49-66, March 2023. DOI : 10.34166/rokms.2023.12.1.49. 
  2. B.Y. Um, "Study On Military Utility and Developmental Strategy for the Drones," Journal of Military Research and Development, Vol. 12, No. 2, pp.115-145, December 2018. 
  3. S.H. Park, S.K. Kim and S.P. Namgung, "A Study on Overseas battle Cases Using Combat Drone", The Journal of the Convergence on Culture Technology(JCCT), Vol. 9, No. 5, pp.403-407, 2023. DOI : 10.17703/JCCT.2023.9.5.403 
  4. J.Y. Hwang, B.A. Choi, H.J. Lee and J.H. Bae, "A Study on Transfer Learning based Ensemble Model for Classification of Maneuver Weapon System," Journal of Korean Institute Of Information Technology, Vol. 19, No. 12, pp.1-10, 2021. DOI : 10.14801/jkiit.2021.19.12.1 
  5. J.Y. Park, J.H. Kim, H.S. Moon, "A Study on Expert System for Nationality Classification of Coast Guard Ship using Convolution Neural Network," Journal of Industrial Studies(J.I.S), Vol. 44, No. 1, pp.45-60, May 2020. 
  6. M.A. Jeong, J.M. Ma, "A Study on the improvement of submarine detection based on mast images using an ensemble model of convolutional neural networks," Journal of the Korea Institute of Military Science and Technology, Vol. 23, No. 2, pp.115-124, 2020. DOI : 10.9766/KIMST.2020.23.2.115 
  7. Y.K. Lee, J.M. Ma, "A Comparative Study on the Performance of CNN Models for Classification of Artillery Weapon Systems," Journal of The Korea Academia-Industrial cooperation Society, Vol. 24, No. 1, pp.344-350, 2023. DOI : 10.5762/KAIS.2023.24.1.344 
  8. Z. Li, W. Yang, S. Peng, F. Liu, "A Survey of Convolutional Neural Networks: Analysis, Applications, and Prospects," IEEE transactions on neural networks and learning systems, 2021. DOI : 10.1109/TNNLS.2021.3084827 
  9. K. Simonyan and A. Zisserman, "Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1904.1556. 2014. 
  10. K. He, X. Zhang, S. Ren, J. Sun, "Deep residual learning for image recognition," In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.770-778, 2016. 
  11. C. Szegedy, V. Vanhoucke, S. loffe, J. Shlens, Z. Wojna, "Rethinking the inception architecture for computer vision," In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.2818-2826, 2016. 
  12. https://www.kaggle.com/datasets/.[accessed: January. 8. 2024