留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

深度边缘光谱U-Net海水网箱养殖信息提取

柯丽娜 翟宇宁 范剑超

柯丽娜,翟宇宁,范剑超. 深度边缘光谱U-Net海水网箱养殖信息提取[J]. 海洋学报,2022,44(2):132–142 doi: 10.12284/hyxb2022026
引用本文: 柯丽娜,翟宇宁,范剑超. 深度边缘光谱U-Net海水网箱养殖信息提取[J]. 海洋学报,2022,44(2):132–142 doi: 10.12284/hyxb2022026
Ke Li’na,Zhai Yuning,Fan Jianchao. Marine cage aquaculture information extraction based on deep edge spectral U-Net[J]. Haiyang Xuebao,2022, 44(2):132–142 doi: 10.12284/hyxb2022026
Citation: Ke Li’na,Zhai Yuning,Fan Jianchao. Marine cage aquaculture information extraction based on deep edge spectral U-Net[J]. Haiyang Xuebao,2022, 44(2):132–142 doi: 10.12284/hyxb2022026

深度边缘光谱U-Net海水网箱养殖信息提取

doi: 10.12284/hyxb2022026
基金项目: 国家自然科学基金(42076184, 41876109, 41806207, 41706195);国家重点研发计划(2016YFC1401007, 2017YFC1404902);国家高分重大科研专项(41-Y30F07-9001-20/22)。
详细信息
    作者简介:

    柯丽娜(1978-),女,辽宁省庄河市人,教授,博士生导师,研究方向为海岸带遥感技术应用。E-mail: kckcsunny@163.com

    通讯作者:

    范剑超,博士,研究员,研究方向为海洋遥感影像人工智能分析。E-mail:jcfan@nmemc.org.cn

  • 中图分类号: TP75;S967.3

Marine cage aquaculture information extraction based on deep edge spectral U-Net

  • 摘要: 网箱养殖是海水养殖的重要类型之一,传统网箱养殖目标光谱特征受近岸植被、水体影响较大,易出现噪声问题。新型深海网箱养殖目标离岸较远,但养殖目标海面框体部分较小,与自然水体光谱相似性较高,难以实现有效提取。本文提出深度边缘光谱U-Net模型对两种海水网箱养殖类型进行养殖信息提取。该模型通过Canny算子双边滤波算法去除波段运算后冗余光谱信息,提取边缘光谱特征并利用U-Net跳跃连接结构将其与深度卷积网络特征相融合,经softmax分类器逐像素分类实现网箱养殖信息提取。以海南近岸网箱养殖与深海网箱养殖为研究对象进行养殖信息提取,经实验对比所提方法在传统近岸网箱目标上精确度达到97.35%,新型深海网箱目标上提取精度达98.99%,其提取结果明显优于传统无监督算法和典型深度学习网络模型。
  • 图  1  研究区及网箱养殖目标示意图

    Fig.  1  Study area and cage culture images

    图  2  传统近岸网箱(a)与新型深海网箱养殖(b)

    Fig.  2  Traditional offshore cages (a) and new deep-sea cages culture (b)

    图  3  整体流程图

    Fig.  3  The overall flow chart

    图  4  不同波段计算效果

    Fig.  4  Calculation effect of different bands

    图  5  DES-Unet结构图

    Fig.  5  Schematic diagram of DES-Unet structure

    图  6  镜像对称操作

    Fig.  6  Mirror symmetry operation

    图  7  深度特征提取效果

    Fig.  7  The effect of depth feature extraction

    图  8  养殖区域遥感影像(a)与真值(b)

    Fig.  8  Remote sensing images (a) and true value (b) of the breeding area

    图  9  近岸网箱提取效果

    Fig.  9  Extraction effect of offshore cages

    图  10  万宁市养殖提取结果

    Fig.  10  The results of aquaculture extraction in Wanning City

    图  11  陵水县养殖提取结果

    Fig.  11  The results of aquaculture extraction in Lingshui Country

    图  12  深海网箱提取效果

    Fig.  12  Extraction effect of deep-sea cages

    图  13  临高县养殖提取结果

    Fig.  13  Breeding extraction results in Lingao

    表  1  数据介绍

    Tab.  1  Data introduction

    卫星位置接收日期传感器云覆盖/%空间分辨率/m
    纬度经度
    GF-219.9°N109.5°E2020年6月12日PMS11
    GF-218.3°N111.3°E2020年5月13日PMS41
    GF-218.6°N111.3°E2020年5月13日PMS41
    下载: 导出CSV

    表  2  算法参数表

    Tab.  2  Parameters of algorithm

    提取算法参数设置
    Canny算子[11]近岸网箱:低阈值为0.08;高阈值为0.2
    深海网箱:低阈值为0.04;高阈值为0.1
    滤波器方差为2
    Otsu[12]阈值根据图像自适应计算
    PCA_Kmeans[13]近岸网箱:分块大小为3×3
    特征向量维度为3
    深海网箱:分块大小为3×3
    特征向量维度为1
    下载: 导出CSV

    表  3  近岸网箱精度验证结果

    Tab.  3  Accuracy verification results of offsore cages

    提取算法准确率Acc召回率R精确率pKappa系数
    Canny算子[9]86.28%86.05%54.14%0.77
    Otsu[10]92.86%93.33%57.32%0.70
    PCA_Kmeans[11]92.85%93.22%66.48%0.67
    Segnet[20]91.59%88.45%85.10%0.69
    U-Net[21]95.31%92.53%72.97%0.69
    DES-Unet97.35%97.38%89.37%0.88
    下载: 导出CSV

    表  4  深海网箱提取精度验证

    Tab.  4  Verification of the extraction accuracy of deep-sea cages

    提取算法准确率/%召回率/%精确率/%Kappa系数
    Canny算子[9]93.0185.1075.690.79
    Otsu[10]88.1082.6158.510.45
    PCA_Kmeans[11]88.1689.1264.580.49
    Segnet[20]92.1586.0988.600.65
    U-Net[21]96.3297.2589.350.81
    DES-Unet98.9998.8195.140.91
    下载: 导出CSV
  • [1] Yucel-Gier G, Eronat C, Sayin E. The impact of marine aquaculture on the environment; the importance of site selection and carrying capacity[J]. Agricultural Sciences, 2019, 10(3): 259−266. doi: 10.4236/as.2019.103022
    [2] 王学哲, 闫吉顺, 王鹏, 等. 辽宁省围海养殖的用海效益和退出机制[J]. 海洋开发与管理, 2019, 36(4): 17−19. doi: 10.3969/j.issn.1005-9857.2019.04.004

    Wang Xuezhe, Yan Jishun, Wang Peng, et al. The benefit and withdrawal mechanism for sea reclamation aquaculture in Liaoning Province[J]. Ocean Development and Management, 2019, 36(4): 17−19. doi: 10.3969/j.issn.1005-9857.2019.04.004
    [3] 包樱钰, 李菲菲, 温东辉. 我国海水养殖业的抗生素污染现状[J]. 海洋环境科学, 2021, 40(2): 294−302. doi: 10.12111/j.mes.20190312

    Bao Yingyu, Li Feifei, Wen Donghui. Antibiotic contamination in mariculture in China[J]. Marine Environmental Science, 2021, 40(2): 294−302. doi: 10.12111/j.mes.20190312
    [4] 彭广海, 付婧, 马增岭, 等. 基于3种生物指数的三沙湾养殖活动底栖环境效应研究[J]. 海洋学报, 2018, 40(4): 106−118.

    Peng Guanghai, Fu Jing, Ma Zengling, et al. Effects of mariculture activities on benthic environment based on analysis of three biotic indices in Sansha Bay[J]. Haiyang Xuebao, 2018, 40(4): 106−118.
    [5] 王宗灵, 傅明珠, 周健, 等. 黄海浒苔绿潮防灾减灾现状与早期防控展望[J]. 海洋学报, 2020, 42(8): 1−11.

    Wang Zongling, Fu Mingzhu, Zhou Jian, et al. Current situation of prevention and mitigation of the Yellow Sea green tide and proposing control measurements in the early stage[J]. Haiyang Xuebao, 2020, 42(8): 1−11.
    [6] 王志华, 杨晓梅, 苏奋振, 等. 我国海岸带海岛礁遥感研究进展及建议[J]. 中国工程科学, 2019, 21(6): 59−63.

    Wang Zhihua, Yang Xiaomei, Su Fenzhen, et al. Remote sensing application in China’s coastal zones and islands: recent progress and some suggestions[J]. Strategic Study of CAE, 2019, 21(6): 59−63.
    [7] 徐京萍, 赵建华, 张丰收, 等. 面向对象的池塘养殖用海信息提取[J]. 国土资源遥感, 2013, 25(1): 82−85.

    Xu Jingping, Zhao Jianhua, Zhang Fengshou, et al. Object-oriented information extraction of pond aquaculture[J]. Remote Sensing for Land and Resources, 2013, 25(1): 82−85.
    [8] 李海涛, 王腾, 王印庚. 基于VGG-16卷积神经网络的海水养殖病害诊断[J]. 计算机系统应用, 2020, 29(7): 222−227.

    Li Haitao, Wang Teng, Wang Yingeng. Diagnosis of marine aquaculture diseases based on VGG-16 convolutional neural network[J]. Computer Systems & Applications, 2020, 29(7): 222−227.
    [9] 杨英宝, 江南, 殷立琼, 等. 东太湖湖泊面积及网围养殖动态变化的遥感监测[J]. 湖泊科学, 2005, 17(2): 133−138. doi: 10.3321/j.issn:1003-5427.2005.02.007

    Yang Yingbao, Jiang Nan, Yin Liqiong, et al. RS-based dynamic monitoring of lake area and enclosure culture in East Taihu Lake[J]. Journal of Lake Science, 2005, 17(2): 133−138. doi: 10.3321/j.issn:1003-5427.2005.02.007
    [10] Seto K C, Fragkias M. Mangrove conversion and aquaculture development in Vietnam: a remote sensing-based approach for evaluating the Ramsar convention on wetlands[J]. Global Environmental Change, 2007, 17(3/4): 486−500.
    [11] Canny J. A computational approach to edge detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1986, PAMI-8(6): 679−698. doi: 10.1109/TPAMI.1986.4767851
    [12] Ostu N. A threshold selection method from gray level histograms[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1): 62−66. doi: 10.1109/TSMC.1979.4310076
    [13] Celik T. Unsupervised change detection in satellite images using principal component analysis and k-means clustering[J]. IEEE Geoscience and Remote Sensing Letters, 2009, 6(4): 772−776. doi: 10.1109/LGRS.2009.2025059
    [14] 卢业伟, 李强子, 杜鑫, 等. 基于高分辨率影像的近海养殖区的一种自动提取方法[J]. 遥感技术与应用, 2015, 30(3): 486−494.

    Lu Yewei, Li Qiangzi, Du Xin, et al. A method of coastal aquaculture area automatic extraction with high spatial resolution images[J]. Remote Sensing Technology and Application, 2015, 30(3): 486−494.
    [15] Kang Junmei, Sui Lichun, Yang Xiaomei, et al. Sea surface-visible aquaculture spatial-temporal distribution remote sensing: a case study in Liaoning Province, China from 2000 to 2018[J]. Sustainability, 2019, 11(24): 7186. doi: 10.3390/su11247186
    [16] Wang Zhihua, Yang Xiaomei, Liu Yueming, et al. Extraction of coastal raft cultivation area with heterogeneous water background by thresholding object-based visually salient NDVI from high spatial resolution imagery[J]. Remote Sensing Letters, 2018, 9(9): 839−846. doi: 10.1080/2150704X.2018.1468103
    [17] 程博, 刘岳明, 刘旭楠, 等. 基于多源特征融合的高分辨率遥感图像近海养殖区提取方法研究[J]. 遥感技术与应用, 2018, 33(2): 296−304.

    Cheng Bo, Liu Yueming, Liu Xu’nan, et al. Research on extraction method of coastal aquaculture areas on high resolution remote sensing image based on multi-features fusion[J]. Remote Sensing Technology and Application, 2018, 33(2): 296−304.
    [18] Shen Yu, Zhu Sijie, Chen Chen, et al. Efficient deep learning of nonlocal features for hyperspectral image classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 59(7): 6029−6043. doi: 10.1109/TGRS.2020.3014286
    [19] Cheng Bo, Liang Chenbin, Liu Xunan, et al. Research on a novel extraction method using deep learning based on GF-2 images for aquaculture areas[J]. International Journal of Remote Sensing, 2020, 41(9): 3575−3591. doi: 10.1080/01431161.2019.1706009
    [20] Badrinarayanan V, Kendall A, Cipolla R. SegNet: a deep convolutional encoder-decoder architecture for image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(12): 2481−2495. doi: 10.1109/TPAMI.2016.2644615
    [21] Ronneberger O, Fischer P, Brox T. U-Net: Convolutional networks for biomedical image segmentation[C]//Medical Image Computing and Computer-Assisted Intervention (MICCAI). Munich: Springer International Publishing, 2015, 9351: 234−241.
    [22] 郑智腾, 范海生, 王洁, 等. 改进型双支网络模型的遥感海水网箱养殖区智能提取方法[J]. 国土资源遥感, 2020, 32(4): 120−129.

    Zheng Zhiteng, Fan Haisheng, Wang Jie, et al. An improved double-branch network method for intelligently extracting marine cage culture area[J]. Remote Sensing for Land and Resources, 2020, 32(4): 120−129.
    [23] 刘岳明, 杨晓梅, 王志华, 等. 基于深度学习RCF模型的三都澳筏式养殖区提取研究[J]. 海洋学报, 2019, 41(4): 119−130.

    Liu Yueming, Yang Xiaomei, Wang Zhihua, et al. Extracting raft aquaculture areas in Sanduao from high-resolution remote sensing images using RCF[J]. Haiyang Xuebao, 2019, 41(4): 119−130.
    [24] Cui Binge, Fei Dong, Shao Guanghui, et al. Extracting raft aquaculture areas from remote sensing images via an improved U-net with a PSE structure[J]. Remote Sensing, 2019, 11(17): 2053. doi: 10.3390/rs11172053
    [25] Liu Chenxi, Jiang Tao, Zhang Zhen, et al. Extraction method of offshore mariculture area under weak signal based on multisource feature fusion[J]. Journal of Marine Science and Engineering, 2020, 8(2): 99. doi: 10.3390/jmse8020099
    [26] Plaut D C, Nowlan S J, Hinton G E. Experiments on learning by back propagation[R]. Pittsburgh: Carnegie-Mellon University, 1986: 54.
  • 加载中
图(13) / 表(4)
计量
  • 文章访问数:  304
  • HTML全文浏览量:  87
  • PDF下载量:  61
  • 被引次数: 0
出版历程
  • 收稿日期:  2021-06-05
  • 修回日期:  2021-09-01
  • 网络出版日期:  2021-10-25
  • 刊出日期:  2022-02-01

目录

    /

    返回文章
    返回