郝才斐,桑会勇,翟亮,朱熀

• 1:山东科技大学测绘与空间信息学院
• 2:中国测绘科学研究院
• 3:北京虞禾源科技有限公司

摘要(Abstract)：

针对现有研究中少有针对深层分割卷积神经网络构型进行深入改进的问题，该文提出了一种半密集连接策略，以湛江市沿海5 km范围的红树林为研究区，以Sentinel-2中分辨率多光谱遥感影像为原始数据，并基于U型卷积神经网络分割架构构建了一种红树林提取方法。结果表明：该文所提方法比其他方法效果更好。网络在测试集中获得90.96%的红树林提取精度，高于基于面向对象的支持向量机和面向对象的随机森林方法。使用该文方法提取了2020年湛江市红树林分布，获得89.65%的精度。

关键词(KeyWords)： 红树林;遥感影像分类;卷积神经网络;多尺度学习;半密集连接

Abstract:

Keywords:

基金项目(Foundation): 中国测绘科学研究院基本科研业务费项目(AR2017,AR2001);; 2019年度自然资源部高层次人才培养工程杰出人才资助项目(12110600000018003908)

作者(Author): 郝才斐,桑会勇,翟亮,朱熀

DOI: 10.16251/j.cnki.1009-2307.2022.04.019

参考文献(References)：

• [1] WAN Luoma,LIN Yinyi,ZHANG Hongsheng,et al.GF-5 hyperspectral data for species mapping of mangrove in Maipo,Hong Kong[J].Remote Sensing,2020,12(4):656.
• [2] 马云梅，吴培强，任广波，等.基于国产高分遥感的人工种植红树林种间分类方法研究：以广西茅尾海红树林为例[J].海洋技术学报，2019,38(4):1-8.(MA Yunmei,WU Peiqiang,REN Guangbo,et al.Study on interspecific classification method of mangrove planted artificially based on high-resolution remote sensing in China:a case study of mangrove in Maoweihai,Guangxi[J].Journal of Ocean Technology,2019,38(4):1-8.)
• [3] CAO Jingjing,LENG Wanchun,LIU Kai,et al.Object-based mangrove species classification using unmanned aerial vehicle hyperspectral images and digital surface models[J].Remote Sensing,2018,10(2):89.
• [4] KUENZER C,BLUEMEL A,GEBHARDT S,et al.Remote sensing of mangrove ecosystems:a review[J].Remote Sensing,2011,3(5):878-928.
• [5] 郑艺，林懿琼，周建，等.基于资源三号的雷州半岛红树林种间分类研究[J].国土资源遥感，2019,31(3):201-208.(ZHENG Yi,LIN Yiqiong,ZHOU Jian,et al.Mangrove inter-species classification based on ZY-3 image in Leizhou Peninsula,Guangdong province[J].Remote Sensing for Land and Resources,2019,31(3):201-208.)
• [6] 李想，刘凯，朱远辉，等.基于资源三号影像的红树林物种分类研究[J].遥感技术与应用，2018(2):360-369.(LI Xiang,LIU Kai,ZHU Yuanhui,et al.Study on mangrove species classification based on ZY-3 image[J].Remote Sensing Technology and Application,2018(2):360-369.)
• [7] CARRANZA-GARCíA M,GARCíA-GUTIéRREZ J,RIQUELME J.A framework for evaluating land use and land cover classification using convolutional neural networks[J].Remote Sensing,2019,11(3):274.
• [8] FAZA S,NABABAN E B,EFENDI S,et al.An initial study of deep learning for mangrove classification[J].IOP Conference Series:Materials Science and Engineering,2018,420:012093.
• [9] 黄亦其，刘琪，赵建晔，等.基于深度卷积神经网络的红树林物种无人机监测研究[J].中国农机化学报，2020,41(2):141-146,189.(HUANG Yiqi,LIU Qi,ZHAO Jianye,et al.Research on mangrove drone monitoring based on convolutional neural network [J].China Agricultural Machinery and Chemistry News,2020,41(2):141-146,189.)
• [10] 高弋斌.基于深度学习与无人机监测的龙海九龙江口红树林群落识别研究[D].福州：福建农林大学，2020.(GAO Yibin.Study on Recognition of mangrove communities based on deep learning and UAV monitoring in Jiulongjiang Estuary,Longhai[D].Fuzhou:Fujian Agriculture and Forestry University,2020.)
• [11] WAN Luoma,ZHANG Hongsheng,LIU Mingfeng,et al.Early monitoring of exotic mangrove sonneratia in Hong Kong using deep convolutional network at half-meter resolution[J].IEEE Geoscience and Remote Sensing Letters,2021,18(2):203-207.
• [12] RONNEBERGER O,FISCHER P,BROX T.U-net:Convolutional networks for biomedical image segmentation[EB/OL].[2021-07-20].https://ieeexplore.ieee.org/document/9330594.
• [13] MISRA D.Mish:a self regularized non-monotonic neural activation function[EB/OL].[2021-07-20].https://arxiv.org/abs/1908.08681.
• [14] ZHOU Z W,SIDDIQUEE M M R,TAJBAKHSH N,et al.UNet++:redesigning skip connections to exploit multiscale features in image segmentation[J].IEEE Transactions on Medical Imaging,2020,39(6):1856-1867.
• [15] LOSHCHILOV I,HUTTER F.SGDR:stochastic gradient descent with warm restarts[EB/OL].[2021-07-20].https://arxiv.org/abs/1608.03983.