应用卷积神经网络VGG16的星载GNSS-R海冰检测
Sea Ice Detection Using Spaceborne GNSS-R Data by VGG16
胡媛 1华曦帆 1刘卫 2江志豪1
作者信息
- 1. 上海海洋大学工程学院,上海 201306
- 2. 上海海事大学商船学院,上海 201306
- 折叠
摘要
针对全球卫星导航系统反射计(global navigation satellite system-reflection,GNSS-R)海冰检测中延迟-多普勒图(delay-Doppler map,DDM)数据噪声大、消融期精度低等问题,提出将VGG16卷积神经网络模型应用于海冰检测.通过深层的网络结构提取DDM多层次特征进行海冰海水分类,以提高海冰检测的精度和稳定性.实验结果表明,与美国国家海洋和大气管理局地表类型数据对比,所提出的基于VGG16海冰检测方法检测准确率为98.02％,有效提升了海冰检测的准确率和稳定性.
Abstract
To address the problems of high noise and low accuracy in the melting period of delay-Doppler map(DDM)data in global navigation satellite system-reflection(GNSS-R)sea ice detection,the VGG16 convolutional neural network model is proposed to be applied to sea ice detection.The multi-level features of DDM are extracted by deep network structure to improve the accuracy and stability of sea ice detection.The experimental results show that the detection accuracy of the proposed VGG16-based sea ice detection method is 98.02％compared with the National Oceanic and Atmospheric Administration(NOAA)surface type data,which effectively improves the accuracy and stability of sea ice detection.
关键词
海冰遥感/海冰检测/星载GNSS-R/卷积神经网络/延迟-多普勒图/NOAAKey words
sea ice remote sensing/sea ice detection/spaceborne GNSS-R/convolutional neural network/delay-doppler map/NOAA引用本文复制引用
出版年
2024
NETL
NSTL
万方数据