Laser & Optoelectronics Progress, Volume. 59, Issue 24, 2428001(2022)

Multiscale and Adaptive Morphology for Remote Sensing Image Segmentation of Vegetation Areas

Xinna Li, Xiaopeng Wang*, and Tongyi Wei
Author Affiliations
  • School of Electronic and Information Engineering, Lanzhou Jiaotong University, Lanzhou 730070, Gansu, China
  • show less
    Figures & Tables(7)
    Basic process of the proposed method
    GAN morphological filtering. (a) Original image; (b) GAN morphology dilate-open filtering; (c) GAN morphology erode-close filtering; (d) GAN morphology compound filtering
    Results of segmentation at different scales. (a) Original image; (b) segmentation scale is 60; (c) segmentation scale is 150
    Segmentation results of different methods. (a) Original image; (b) method 1; (c) method 2; (d) proposed method; (e) manual drawing
    Segmentation results of different methods for images with different complexity. (a) Original image; (b) method 1; (c) method 2; (d) proposed method
    • Table 1. Segmentation error and index comparison of different methods

      View table

      Table 1. Segmentation error and index comparison of different methods

      Image size/pixelMethodRMSEPSNR /dBSegmentation error /%Operation time/ s
      355×374Method 156.56030.11929.8751.797
      Method 244.40334.95918.3863.088
      Proposed method38.44137.8433.1893.836
    • Table 2. Segmentation error and index comparison of different methods for images with different complexity

      View table

      Table 2. Segmentation error and index comparison of different methods for images with different complexity

      Imege No.Image size /pixelMethodRMSEPSNR /dBSegmentation error /%Operation time /s
      Image 1299×277Method 189.47019.61527.2861.986
      Method 287.06621.49228.3863.158
      Proposed method86.53421.49222.1633.936
      Image 2298×277Method 152.38331.65318.2851.628
      Method 246.14734.95919.9422.935
      Proposed method43.42436.6923.8223.683
      Image 3163×152Method 167.12819.93325.3281.986
      Method 254.96522.38222.8753.158
      Proposed method50.70326.6936.6323.936
      Image 4350×300Method 170.98720.80925.8261.882
      Method 268.84425.57518.7853.328
      Proposed method63.44632.6634.2374.026
    Tools

    Get Citation

    Copy Citation Text

    Xinna Li, Xiaopeng Wang, Tongyi Wei. Multiscale and Adaptive Morphology for Remote Sensing Image Segmentation of Vegetation Areas[J]. Laser & Optoelectronics Progress, 2022, 59(24): 2428001

    Download Citation

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

    Category: Remote Sensing and Sensors

    Received: Sep. 8, 2021

    Accepted: Oct. 27, 2021

    Published Online: Nov. 28, 2022

    The Author Email: Xiaopeng Wang (wangxp1969@sina.com)

    DOI:10.3788/LOP202259.2428001

    Topics