BROWSE
Detail
Conference Paper
한국 해상 환경에 적합한 영상 기반 선박 인식 기술에 관한 연구
Cited - time in 
Share 
Share
- Issue Date
- 2021-07
- Citation
- 대한전자공학회 학술 대회 (하계) 2021, pp.1520-1524
- Publisher
- 대한전자공학회
- Language
- Korean
- Type
- Conference Paper
- Abstract
- Recently, the introduction of deep learning in various fields has achieved great achievements, but the development of image-based ship recognizers suitable for Korean maritime environment is still insufficient, and learning data has not been properly established. In this paper, in order to solve these problems, we first defined and built ship learning data suitable for the Korean maritime environment. In addition, a deep learning-based ship recognizer structure that can be used in actual applications was proposed. As a result, the ship detection performance of mAP 0.64 and the ship identification performance of 92.37% were confirmed. The proposed technology can be used for maritime traffic control and autonomous driving for ships, or it can be used for military purposes.
- KSP Keywords
- Identification performance, Image-based, Learning data, Learning-based, Maritime Traffic, Maritime environment, Traffic control, autonomous driving, deep learning(DL), detection performance, ship detection