基于机器学习遥感算法的大型漂浮藻面积提取方法研究

李兆鑫<sup>1</sup>; 丘仲锋<sup>1*</sup>; 李旭文<sup>2</sup>; 姜晟<sup>2</sup>; 张悦<sup>2</sup>

引用本文:	LI Zhao-xin, QIU Zhong-feng, LI Xu-wen, JIANG Sheng, ZHANG Yue.Study on Area Extraction Method of Floating Macroalgae Based on Machine Learning Remote Sensing Algorithm[J].Environmental Monitoring and Forewarning,2019,11(5):46~51

【打印本页】【HTML】【下载PDF全文】【View/Add Comment】【EndNote】【RefMan】【BibTex】

←前一篇|后一篇→

过刊浏览高级检索

本文已被：浏览 6074次下载 1626次	码上扫一扫！
分享到：微信更多字体:加大+\|默认\|缩小-
基于机器学习遥感算法的大型漂浮藻面积提取方法研究
李兆鑫，丘仲锋，李旭文，姜晟，张悦^1,2
1.南京信息工程大学海洋科学学院，江苏南京 210044;2. 江苏省环境监测中心，江苏南京 210019

摘要:

采用基于机器学习的多层感知机算法，利用GOCI（Geostationary Ocean Color Imager）传感器获取的瑞利校正反射率数据，对东中国海大型漂浮藻进行遥感自动识别，采用线性混合像元分解来计算大型漂浮藻的覆盖面积，并利用膨胀和侵蚀法进行大型漂浮藻的分布面积计算。利用L8/OLI（Landsat 8/ Operational Land Imager）高空间分辨率资料进行验证，结果表明，基于机器学习遥感算法针对GOCI提取的大型漂浮藻覆盖面积，与L8/OLI结果十分接近，R²达到0.959，平均绝对误差和平均相对误差分别为39.32 km²和18.15%。

关键词: 大型漂浮藻机器学习地球静止海洋水色成像仪遥感

DOI：

分类号:X87

文献标识码:C

基金项目:国家自然科学基金资助项目（41506200,41576172）；国家水体污染控制与治理科技重大专项基金资助项目（2017ZX07302-003）；2018年太湖水污染治理省级专项基金资助项目（TH2018304）；江苏省环境监测科研基金资助项目（1312，1704）。

Study on Area Extraction Method of Floating Macroalgae Based on Machine Learning Remote Sensing Algorithm

LI Zhao-xin, QIU Zhong-feng, LI Xu-wen, JIANG Sheng, ZHANG Yue^1,2

1.School of Marine Sciences, Nanjing University of Information Science & Technology, Nanjing, Jiangsu 210044, China;2. Jiangsu Provincial Environmental Monitoring Center, Nanjing, Jiangsu 210019, China

Abstract:

This paper used the multi layer perceptron algorithm based on machine learning, aiming at the Rayleigh corrected reflectance data obtained by GOCI (Geostationary Ocean Color Imager), to achieve the remotely automatic identification of floating macroalgae in the eastern China sea, and used the linear mixed pixel decomposition to calculate the coverage area and also used the methods of dilation and erosion to derive the distribution area of floating macroalgae. High spatial resolution data of L8/OLI (Landsat 8/ Operational Land Imager) were used to verify the proposed algorithm. The results showed that the coverage area of floating macroalgae extracted using GOCI based on machine learning remote sensing algorithm in this paper had a good agreement with that of L8/OLI, with R² reaching 0.959, MAE and MRE approaching 39.32 km² and 18.15%, respectively.

Key words: Floating macroalgae Machine learning GOCI Remote sensing