Remote Sensing Technology and Application, Volume. 40, Issue 3, 545(2025)

Semantic Segmentation Model-based Mangrove Identification Method and Time-Series Variation Analysis in Wenzhou City

Yun WANG1, Mengguang LIAO1、*, Nan CHU2,3, Xing CHEN1, Shaoning LI3, and Junjie ZHOU1
Author Affiliations
  • 1College of Geoscience and Spatial Information Engineering, Hunan University of Science and Technology, Xiangtan411201,China
  • 2Institute for Local Sustainable Development Goals, Hunan University of Science and Technology, Xiangtan411201,China
  • 3National-Local Joint Engineering Laboratory of Geo-Spatial Information Technology, Hunan University of Science and Technology, Xiangtan411201,China
  • show less
    Figures & Tables(11)
    Map of mangrove distribution and growth in the study area
    Flowchart of mangrove extraction and analysis based on Sentinel-2 remote sensing image
    Dataset Production Process
    DeepLabV3+ network architecture
    Model training situation
    Extraction results of different semantic segmentation models
    Mangrove extraction results in the study area from 2019 to 2023
    Wenzhou mangrove forest field research situation
    • Table 1. Sentinel-2 remote sensing image selection

      View table
      View in Article

      Table 1. Sentinel-2 remote sensing image selection

      模型训练阶段
      编号成像日期潮高/cm编号成像日期潮高/cm
      120220103557.5720220729516
      220220311195820220821141.5
      320220403484.5920230125294
      420220410201.51020230128131.5
      520220629486.51120230306528.5
      6202207092431220230309516
      影像分析阶段
      120181217242420211114197
      220200131215520230108254
      320210108140
    • Table 2. Testing accuracy of different methods

      View table
      View in Article

      Table 2. Testing accuracy of different methods

      提取方法精确率/%召回率/%IoU/%Kappa系数训练时长
      语义分割方法U-Net++模型84.184.9973.230.812 h 32 min 59 s
      SegNet模型69.6284.0269.620.782 h 13 min 41 s
      DeepLabV3+模型84.8986.9175.260.822 h 24 min 56 s
      传统机器学习方法最大似然法61.6491.89\0.72\
      支持向量机61.4479.51\0.67\
      随机森林68.3685.28\0.74\
    • Table 3. Identified area and planting of mangroves in the study area

      View table
      View in Article

      Table 3. Identified area and planting of mangroves in the study area

      树排沙岛

      /hm2

      鳌江沿岸/hm2

      霓屿岛

      /hm2

      沿浦湾

      /hm2

      潮高

      /cm

      鳌江沿岸其中鳌江口

      种植

      面积

      2014年

      4 hm2

      2015年

      70.27 hm2

      2002年 10.07 hm2

      2018年 8.36 hm2

      2019年 10.62 hm2

      2020年 3.25 hm2

      2018年 13.71 hm2

      2019年 13.60 hm2

      2020年

      4.49 hm2

      2017年 22.65 hm2

      2018年 12.47 hm2

      2019年

      2.92 hm2

      2019年29.0028.4912.345.9714.22242
      2020年29.2729.8712.836.3117.41215
      2021年36.5828.7810.398.1816.60140
      2022年39.8624.597.938.5418.40197
      2023年39.6528.037.839.3517.63254
    Tools

    Get Citation

    Copy Citation Text

    Yun WANG, Mengguang LIAO, Nan CHU, Xing CHEN, Shaoning LI, Junjie ZHOU. Semantic Segmentation Model-based Mangrove Identification Method and Time-Series Variation Analysis in Wenzhou City[J]. Remote Sensing Technology and Application, 2025, 40(3): 545

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: Oct. 24, 2023

    Accepted: --

    Published Online: Sep. 28, 2025

    The Author Email: Mengguang LIAO (liaomengguang@163.com)

    DOI:10.11873/j.issn.1004-0323.2025.3.0545

    Topics