卷积神经网络在SAR遥感海冰分类中的应用可行性分析  

Feasibility analysis of convolutional neural networks in remote sensing sea ice classification

作　　者：崔艳荣 邹斌 韩震[1] 石立坚[2,3] 刘森[2] CUI Yan-rong;ZOU Bin;HAN Zhen;SHI Li-jian;LIU Sen(College of Marine Science,Shanghai Ocean University,Shanghai 201306,China;National Satellite Ocean Application Service,Beijing 10081 China;Key Laboratory of Space Ocean Remote Sensing and Application,State Oceanic Administration,Beijing 100081 China)

机构地区：[1]上海海洋大学海洋科学学院,上海201306 [2]国家卫星海洋应用中心,北京100081 [3]国家海洋局空间海洋遥感与应用研究重点实验室,北京100081

出　　处：《海洋预报》2019年第5期77-85,共9页Marine Forecasts

基　　金：国家重点研发计划（2018YFC1407200;2018YFC1407206）

摘　　要：分别介绍了卫星遥感海冰监测、分类的传统方法,以及卷积神经网络在遥感影像分类识别中的应用成果。尝试将在图像识别、语言检测等方面取得成功的卷积神经网络算法应用在海冰图像分类中,利用其能够应对非线性、网络结构简单、可并行运算等能力去解决海冰数据分类问题。This paper analyzes the traditional methods of sea ice monitoring and classification based on satellite remote sensing, and the application results of convolutional neural network in remote sensing image classification and recognition. Moreover, the convolutional neural networks algorithm, which has been successfully used in image recognition and language detection, is applied to the classification of sea ice images, and to solve the issues of sea ice data classification based on its simple network structure, capability in coping with nonlinearity and parallel computing.

关 键 词：卫星遥感 海冰分类 应用 卷积神经网络 深度学习 

分 类 号：P731.15[天文地球—海洋科学]