| 融合空间和时序遥感信息的深度学习水稻提取 |
| |
| 引用本文： | 周佳玮,涂理林,陈洪建,江挺,林佳佳.融合空间和时序遥感信息的深度学习水稻提取[J].地理空间信息,2022,20(2):39-44. |
| |
| 作者姓名： | 周佳玮 涂理林 陈洪建 江挺 林佳佳 |
| |
| 作者单位： | 宁波市鄞州区测绘院,浙江 宁波 315192,武汉大学 遥感信息工程学院,湖北 武汉 430079 |
| |
| 基金项目： | 湖北省自然科学基金创新群体项目 |
| |
| 摘 要： | 为了得到更加精细的水稻提取结果,提出一种结合高分辨率和多时序遥感影像的深度学习水稻提取方法.构建全卷积网络(FCN)对BJ-2高分辨率遥感影像进行分类,并利用长短期记忆网络(LSTM)和随机森林(RF)分类器对Sentinel-2多时序遥感影像进行分类,再通过面向对象的分割和投票对3种方法的分类结果进行融合,得到最终提...
|
| 关 键 词： | 水稻提取 高分辨率遥感影像 多时序遥感影像 全卷积网络(FCN) 长短期记忆网络(LSTM) 面向对象分割 |
Deep Learning-based Rice Paddy Extraction by Fusing Spatial and Temporal Remote Sensing Information |
| |
| ZHOU Jiawei,TU Lilin,CHEN Hongjian,JIANG Ting,LIN Jiajia.Deep Learning-based Rice Paddy Extraction by Fusing Spatial and Temporal Remote Sensing Information[J].Geospatial Information,2022,20(2):39-44. |
| |
| Authors: | ZHOU Jiawei TU Lilin CHEN Hongjian JIANG Ting LIN Jiajia |
| |
| Institution: | (Yinzhou District Institute of Surveying and Mapping of Ningbo City,Ningbo 315192,China;School of Remote Sensing and Information Engineering,Wuhan University,Wuhan 430079,China) |
| |
| Abstract: | Aiming at deriving finer rice paddy maps,we proposed a deep learning rice paddy extraction method combining high-resolution and multi-temporal remote sensing images.Firstly,we built FullyConvolutional Network(FCN)to classify high-resolution BJ-2 images.Then,we used Long Short Term Memory(LSTM)network and Random Forest(RF)classifier to classify multi-temporal Sentinel-2 images.Finally,we used object-oriented segmentation and voting to fuse three classificationresults and obtained final rice extraction results.The experiment results in Yinzhou District,Ningbo City demonstrate that the proposed method can achieve high accuracy and spatially detailed rice paddy map. |
| |
| Keywords: | rice paddy extraction high-resolution remote sensing image multi-temporal remote sensing image FCN LSTM object-oriented segmentation |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
