Edge-enhanced CNN Training for Fine-grained Ship Classification in Remote Sensing

Changju Bak^*

LT, ROK Navy/Instructor, Dept. of Navigation Operation, Republic of Korea Naval Academy

세분 선박 원격탐사 이미지 분류를 위한 엣지 강화 컨볼루션 신경망 학습

박창주^*

해군 대위/해군사관학교 항해운용학과 교관

Correspondence to: *Changju Bak Dept. of Navigation Operation, Republic of Korea Naval Academy 1 Jungwon-ro, Jinhae-gu, Changwon-si, Gyungsangnam-do, 51704, Republic of Korea Tel: +82-55-907-5083 E-mail: cjbak1225@navy.ac.kr

Abstract

Fine-grained ship remote sensing image classification is essential for maritime surveillance and defense. However, CNN-based fine-grained classification often struggle from subtle inter-class differences, low resolution, and limited data. This study proposes a framework that enhances CNN learning using edge-enhanced images generated through the holistically-nested edge detection (HED) algorithm. Applied to ResNet-50, DenseNet-121, and EfficientNet-B0, proposed framework improved classification accuracy over models trained on original images. The results demonstrate that edge-based learning effectively enhances ship classification accuracy with limited datasets.

초록

선박 원격 센싱 이미지 분류는 해양 감시와 국방 분야에서 중요한 역할을 한다. 그러나 CNN 기반 세분 이미지 분류는 클래스 간 미세한 차이, 낮은 해상도, 제한된 데이터로 인해 성능이 저하되는 문제가 있다. 본 연구는 HED 엣지 검출 알고리즘으로 추출한 엣지 정보를 강화한 이미지를 활용해 CNN의 학습 성능을 개선하는 프레임워크를 제안한다. 이를 ResNet-50, DenseNet-121, EfficientNet-B0 모델에 적용한 결과, 원본 이미지 학습 대비 분류 정확도가 향상되었다. 이는 엣지 기반 학습이 세분 선박 이미지 분류 성능 개선에 적용 가능함을 보여준다.

Keywords:

Fine-Grained Image Classification, Image Classification, Image Processing, Maritime Remote Sensing, Convolutional Neural Network, Edge Detection, Maritime Surveillance

키워드:

세분 이미지 분류, 이미지 분류, 이미지 처리, 해양 원격탐사, 컨볼루션 신경망, 에지 검출, 해양 감시

Acknowledgments

본 논문은 해군사관학교 해양연구소 학술연구과제 연구비의 지원으로 수행된 연구임.

References

Zhang, R., Ji, X., Loughney, S., Wang, J., & Yang, Z. (2025). Visual Perception for Long-distance and Small Target Detection in Autonomous Maritime Navigation. Ocean Engineering, 332, 121447.
Sun, S., Gu, Y., & Ren, M. (2022). Fine-grained Ship Recognition from the Horizontal View Based on Domain Adaptation. Sensors, 22(9), 3243.
Guo, M., Wu, M., Shen, Y., Li, H., & Tao, C. (2025). IFShip: Interpretable Fine-grained Ship Classification with Domain Knowledge-enhanced Vision-language Models. Pattern Recognition, 166, 111672.
Karus, H., Schwenker, F., Munz, M., & Teutsch, M. (2024). Towards Explainable Visual Vessel Recognition Using Fine-Grained Classification and Image Retrieval. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 82-92).
Si, G., Xiao, Y., Wei, B., Bullock, L. B., Wang, Y., & Wang, X. (2023). Token-Selective Vision Transformer for Fine-grained Image Recognition of Marine Organisms. Frontiers in Marine Science, 10, 1174347.
Chen, Y., Zheng, J., & Zhou, Z. (2020). Airbus Ship Detection-traditional vs Convolutional Neural Network Approach. CS229 Course Report, 637.
Zhou, G., Huang, L., & Sun, Q. (2023). Fine-grained Classification of Remote Sensing Ship Images Based on Improved VAN. Computers, Materials & Continua, 77(2).
Xie, S., & Tu, Z. (2015). Holistically-nested Edge Detection. In Proceedings of the IEEE international Conference on Computer Vision (pp. 1395-1403).
Di, Y., Jiang, Z., & Zhang, H. (2021). A Public Dataset for Fine-grained Ship Classification in Optical Remote Sensing Images. Remote Sensing, 13(4), 747.
Biederman, I., & Ju, G. (1988). Surface Versus Edge-based Determinants of Visual Recognition. Cognitive psychology, 20(1), 38-64.
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet Classification with Deep Convolutional Neural Networks. Advances in Neural Information Processing Systems, 25.
He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep Residual Learning for Image Recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 770-778).
Huang, G., Liu, Z., Van Der Maaten, L., & Weinberger, K. Q. (2017). Densely Connected Convolutional Networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 4700-4708).
Tan, M., & Le, Q. (2019, May). Efficientnet: Rethinking Model Scaling for Convolutional Neural Networks. In International Conference on Machine Learning (pp. 6105-6114). PMLR.
Xia, G. S., Bai, X., Ding, J., Zhu, Z., Belongie, S., Luo, J., ... & Zhang, L. (2018). DOTA: A Large-scale Dataset for Object Detection in Aerial Images. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 3974-3983).
Liu, Z., Wang, H., Weng, L., & Yang, Y. (2016). Ship Rotated Bounding Box Space for Ship Extraction from High-resolution Optical Satellite Images with Complex Backgrounds. IEEE Geoscience and Remote Sensing Letters, 13(8), 1074-1078.
Cheng, G., Zhou, P., & Han, J. (2016). Learning Rotation-invariant Convolutional Neural Networks for Object Detection in VHR Optical Remote Sensing Images. IEEE Transactions on Geoscience and Remote Sensing, 54(12), 7405-7415.
Canny, J. (2009). A Computational Approach to Edge Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 6, 679-698.
Sobel, I. (2014). History and Definition of the Sobel Operator. Retrieved from the World Wide Web, 1505.
Simonyan, K., & Zisserman, A. (2014). Very Deep Convolutional Networks for Large-scale Image Recognition. arXiv preprint arXiv:1409.1556, .
Arbelaez, P., Maire, M., Fowlkes, C., & Malik, J. (2010). Contour Detection and Hierarchical Image Segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 33(5), 898-916.
Bu, T., Lazarou, M., & Stathaki, T. (2024). Image Edge Enhancement for Effective Image Classification. arXiv preprint arXiv:2401.07028, .
Wang, B., Yu, X., Huang, C., Sheng, Q., Wang, Y., & Liu, W. (2022). DCNN-based Ship Classification Using Enhanced Edge Information and Inception Module. J. Imaging Sci. Technol, 30501, 1.