Whitecap detection analysis using instance segmentation models
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
In the ocean, whitecaps resulting from various factors such as wind and ships induce considerable disturbances in imagebased object detection and recognition models. These disturbances limit the accuracy of these models. Previous studies have employed object detection models to detect and remove irregular phenomena of nonuniform sizes and shapes. Because these models are trained on stereotyped rectangular objects, expressing the exact shape of an actual object remains challenging. Therefore, to ensure the accuracy of real-time optimal whitecap detection in a marine environment, we applied and analyzed core instance segmentation models based on real images captured by a drone. The instance segmentation models used in the experiment were selected as two models, namely anchor- and anchor-free models, and four models were analyzed to ensure real-time accuracy and processing speed.
Abstract
In the ocean, whitecaps resulting from various factors such as wind and ships induce considerable disturbances in imagebased object detection and recognition models. These disturbances limit the accuracy of these models. Previous studies have employed object detection models to detect and remove irregular phenomena of nonuniform sizes and shapes. Because these models are trained on stereotyped rectangular objects, expressing the exact shape of an actual object remains challenging. Therefore, to ensure the accuracy of real-time optimal whitecap detection in a marine environment, we applied and analyzed core instance segmentation models based on real images captured by a drone. The instance segmentation models used in the experiment were selected as two models, namely anchor- and anchor-free models, and four models were analyzed to ensure real-time accuracy and processing speed.
참고문헌 (19)
- C. E. Stringari, "Deep neural networks for active wave breaking classification", Scientific Reports, 11(1), pp.3604~ (2021)
- Y. Sugihara, "Variation of whitecap coverage with wave-field conditions", Journal of Marine Systems, 66(1-4), pp.47~60 (2007)
- B. Scanlon, "Oceanic wave breaking coverage separation techniques for active and maturing whitecaps", Methods in Oceanography, 8, pp.1~12 (2013)
- X. Hu, "Using optical flow trajectories to detect whitecaps in light-polluted videos", Remote Sensing, 14(22), pp.5691~ (2022)
- E. A. Atkin, "Machine-learned peel angles for surfing wave quality monitoring", Australasian Coasts & Ports 2021: Te Oranga Takutai, Adapt and Thrive: Te Oranga Takutai, Adapt and Thrive, pp.91~97 (2022), New Zealand Coastal Society
- T. Vrecica, "Deep learning applied to sea surface semantic segmentation: Filtering sunglint from aerial imagery", Proceedings of the ICML 2021 Workshop on Tackling Climate Change with Machine Learning, 23 (2020)
- J. Long, "Fully convolutional networks for semantic segmentation", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.3431~3440 (2015)
- H. Zhao, "Pyramid scene parsing network", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.6230~6239 (2017)
- K. He, "Mask rcnn", Proceedings of the IEEE International Conference on Computer Vision, pp.2980~2988 (2017)
- X. Chen, "Tensormask:A foundation for dense object segmentation", Proceedings of the IEEE, 2019, 2061~2069
- Y. Lee, "Centermask: Real-time anchor-free instance segmentation", Proceedings of the IEEE, 2020, 13906~13915
- Z. Tian, "Conditional convolutions for instance segmentation", Computer Vision–ECCV 2020:16th European Conference, Proceedings, Part I 16, 8, pp.282~298 (2020), Springer International Publishing
- Z. Tian, "FCOS: A simple and strong anchor-free object detector", IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(4), pp.1922~1933 (2022)
- T. Lin, "Feature pyramid networks for object detection", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.936~944 (2017)
- K. Khan, "DBSCAN: Past, present and future", The Fifth International Conference on the Applications of Digital Information and Web Technologies (ICADIWT 2014), pp.232~238 (2014), IEEE
- M. Shafiq, "Deep residual learning for image recognition: A survey", Applied Sciences, 12(18), pp.8972~ (2022)
- Z. Zou, "Object detection in 20 years: A survey", Proceedings of the IEEE, 111(3), pp.257~276 (2023)
- W. Gu, "A review on 2D instance segmentation based on deep neural networks", Image and Vision Computing, 120, pp.104401~ (2022)
- T. Lin, "Microsoft coco: Common objects in context", Computer Vision–ECCV 2014: 13th European Conference, Proceedings, Part V 13 (2014), Springer International Publishing