深度学习的SAR图像海洋涡旋自动检测及其特征提取  

Research on automatic detection of oceanic eddies in SAR image and extraction of its characteristic parameters based on depth learning

在线阅读下载全文

作  者:吴进群[1,2] 陈戈 马纯永[1] 郑益勤[2] WU Jin-qun;CHEN Ge;MA Chun-yong;ZHENG Yi-qin(Ocean University of China,Qingdao 266100,China;PIESAT International Information Technology Limited,Beijing 100097,China)

机构地区:[1]中国海洋大学,山东青岛266100 [2]航天宏图信息技术股份有限公司,北京100097

出  处:《舰船科学技术》2023年第6期170-173,共4页Ship Science and Technology

基  金:国家自然科学基金资助项目(42276179)。

摘  要:本文针对SAR卫星遥感图像海洋涡旋自动检测问题,提出一种基于YOLOX高性能目标检测的海洋涡旋检测模型YOLOX-EDDY,该模型能够精准检测到亚中尺度海洋涡旋。根据SAR图像海洋涡旋的螺旋线形态,本文提出了涡旋中心位置、涡旋尺度和涡旋边缘位置等特征参数自动提取方法,实现了从海洋涡旋自动检测到涡旋特征信息提取的自动化处理。利用国家卫星海洋应用中心和中国资源卫星应用中心提供的高分三号卫星海洋涡旋SAR图像,构建SAR图像涡旋样本库,进行涡旋自动检测模型训练与验证、涡旋特征参数提取实验。实验表明,本文提出方法具有较强的泛化能力,能够实现海洋涡旋精准检测和涡旋特征参数提取。Aiming at the problem of automatic detection of oceanic eddies in SAR satellite remote sensing images,this paper proposes an oceanic eddies detection model YOLOX-EDDY based on YOLOX high-performance target detection,which can accurately detect the sub-mesoscale eddies.According to the spiral shape of the oceanic eddies in the SAR image,this paper proposes the automatic extraction method of the feature parameters such as the eddy center position,the eddy scale and the eddy edge position,which realizes the automatic processing from the detection of the oceanic eddies to the extraction of the eddy feature information.Using the oceanic eddies SAR images of GF3 satellite provided by National Satellite Ocean Application Service and China Centre for Resources Satellite Data and Application,the eddy sample library of SAR images is constructed,and the eddy automatic detection model training and validation,and the eddy feature parameter extraction experiments are carried out.The experiment shows that the method proposed in this paper has strong generalization ability,and can achieve accurate detection of oceanic eddies and extraction of oceanic eddies characteristic parameters.

关 键 词:SAR图像 YOLOX-EDDY 深度学习 海洋涡旋检测 特征参数提取 

分 类 号:TP5[自动化与计算机技术]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象