Acta Optica Sinica, Volume. 37, Issue 9, 0910002(2017)

A Color Underwater Image Clearness Algorithm Based on Tetrolet Transform

Yu Shen, Jianwu Dang*, Yangping Wang, and Bowei Wang
Author Affiliations
  • School of Electronic and Information Engineering, Lanzhou Jiaotong University, Lanzhou, Gansu 730070, China
  • show less
    Figures & Tables(16)
    Flow chart of the proposed algorithm
    Maping underwater color image to lαβ three channels. (a) Original image; (b) l channel; (c) α channel; (d) β channel
    Five basic tetrominoes
    Structure diagram of Tetrolet transformation
    Comparison of high-frequency details. (a) Tetrolet transform; (b) improved Tetrolet transform
    Effects on image edge of bilateral filtering. (a) Image region with noises; (b) value domain filtering; (c) bilateral filtering; (d) spatial domain filtering; e) filtering combination of spatial domain and value domain
    Images before and after Tetrolet processing. (a) Original images; (b) after Tetrolet processing
    Comparison of experiment results. (a) Original images; (b) Retinex method; (c) three-channel method; (d) proposed method
    • Table 1. Quantitative results comparison of three methods in Image1

      View table

      Table 1. Quantitative results comparison of three methods in Image1

      Image1SDAvGKPSNR
      Original image21.94783.244611.4797
      Retinex method59.30208.787811.724518.7463
      Three-channel method64.989210.536711.010022.0152
      Proposed method69.126014.052310.748940.9846
    • Table 2. Quantitative results comparison of three methods in Image2

      View table

      Table 2. Quantitative results comparison of three methods in Image2

      Image2SDAvGKPSNR
      Original image37.43333.078922.7547
      Retinex method67.73735.671523.009124.4258
      Three-channel method70.84376.383320.788931.5185
      Proposed method75.879616.164217.599843.8117
    • Table 3. Quantitative results comparison of three methods in Image3

      View table

      Table 3. Quantitative results comparison of three methods in Image3

      Image3SDAvGKPSNR
      Original image22.18214.523133.9549
      Retinex method60.028412.188233.159813.8773
      Three-channel method64.787915.730832.091227.8534
      Proposed method68.663015.929126.913539.2829
    • Table 4. Quantitative results comparison of three methods in Image4

      View table

      Table 4. Quantitative results comparison of three methods in Image4

      Image4SDAvGKPSNR
      Original image30.31213.555412.9045
      Retinex method58.03906.808113.166319.6965
      Three-channel method70.86719.274413.092139.1319
      Proposed method73.371216.348413.654841.8504
    • Table 5. Quantitative results comparison of three methods in Image5

      View table

      Table 5. Quantitative results comparison of three methods in Image5

      Image5SDAvGKPSNR
      Original image35.27301.821813.2963
      Retinex method63.56053.416513.540620.0278
      Three-channel method74.85184.420112.194832.1510
      Proposed method55.704115.960913.299041.4206
    • Table 6. Quantitative results comparison of three methods in Image6

      View table

      Table 6. Quantitative results comparison of three methods in Image6

      Image6SDAvGKPSNR
      Original image41.37505.025715.0149
      Retinex method74.26539.096714.645916.4409
      Three-channel method74.87399.573913.568331.4975
      Proposed method66.964920.897211.586842.6501
    • Table 7. Quantitative results comparison of three methods in Image7

      View table

      Table 7. Quantitative results comparison of three methods in Image7

      Image7SDAvGKPSNR
      Original image34.27294.912210.0089
      Retinex method63.39749.287010.049311.7240
      Three-channel method74.853411.18969.406236.0456
      Proposed method72.899418.30519.347438.6173
    • Table 8. Quantitative results comparison of three methods in Image8

      View table

      Table 8. Quantitative results comparison of three methods in Image8

      Image8SDAvGKPSNR
      Original image23.21334.912210.0089
      Retinex method55.706413.29759.193418.8354
      Three-channel method74.751717.69596.726140.1223
      Proposed method59.886515.76956.575144.0395
    Tools

    Get Citation

    Copy Citation Text

    Yu Shen, Jianwu Dang, Yangping Wang, Bowei Wang. A Color Underwater Image Clearness Algorithm Based on Tetrolet Transform[J]. Acta Optica Sinica, 2017, 37(9): 0910002

    Download Citation

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

    Category: Image Processing

    Received: Apr. 5, 2017

    Accepted: --

    Published Online: Sep. 7, 2018

    The Author Email: Dang Jianwu (18609311366@163.com)

    DOI:10.3788/AOS201737.0910002

    Topics