Chinese Optics Letters, Volume. 20, Issue 7, 071101(2022)

In-situ laser-induced surface damage inspection based on image super-resolution and adaptive segmentation method

Fengdong Chen1、*, Jingyang Sun1, Qian Wang1, Hongbo Zhu1, Fa Zeng2、**, Yueyue Han1, Cheng Lu1, and Guodong Liu1、***
Author Affiliations
  • 1Instrument Science and Technology, Harbin Institute of Technology, Harbin 150001, China
  • 2Research Center of Laser Fusion, China Academy of Engineering Physics, Mianyang 621900, China
  • show less
    Figures & Tables(14)
    (a) Sketch-map of the FODI. (b) An optic in IOM and an example of LID inside (196 µm). (c) The FODI image of the optics and the LID corresponding to (b).
    Sketch-map of the online optics damage inspection by using TIR illumination and remote imaging method. Edge illumination lights are the only optic to be imaged.
    Semiconductor laser light source (wavelength is 808 nm).
    Edge illumination is uneven. An initiated LID is difficult to detect or accurately segment to judge whether they have grown up. (a) Online image. (b) Uneven lighting simulation.
    Sketch-map of the SR and adaptive segmentation method.
    Examples showing only backlit illumination of the 12 classes in the “Damage Morphology” dataset[15].
    Cross section of the isolated peak and cross section of the peak super-imposed on a background feature, both shown with cross hatching to indicate the total area included when a fixed k is used to define the extent of an LID. For a peak that is not over a background feature, the area of the LID increases smoothly as the k is decreased. However, for a peak that overlays a background feature, the area will sharply increase when the k drops below the level of the background feature.
    Diagram of the FODI online experiment device.
    Optical system of the IOM and the optical imaging lens.
    SR results. (a) Online original FODI image of wedge lens. (b) An example window of original resolution. (c) The 2× resolution of the HAN_SR (SSIM = 0.92). (d) The 4× resolution of the HAN_SR (SSIM = 0.91). (e) The 2× resolution of the CDC_SR (SSIM = 0.92). (f) The 4× resolution of the CDC_SR (SSIM = 0.91). (g) The 2× resolution of the integrated result (SSIM = 0.95). (h) The 4× resolution of the integrated result (SSIM = 0.94).
    Example result of SR image of one LID. (a) Online original FODI image of wedge lens. (b) An example of original resolution of an LID inside (94.0 µm). (c) The 2× integrated resolution of the LID. (d) The 4× integrated resolution of the LID.
    Adaptive LID segmentation results. (a) Online original image of wedge lens segmentation result. (b) The example window of the original resolution segmentation result (46 LID sites found). (c) The 2× integrated resolution image segmentation result (75 LID sites found). (d) The 4× integrated resolution image segmentation result (90 LID sites found).
    Quartile statistical chart of detection rate of LIDs (>50) in the 12 images of the original, 2×, and 4× integrated resolution.
    • Table 1. Optical Resolution and Pixel Resolution of the Optical Lens at Different Locations

      View table
      View in Article

      Table 1. Optical Resolution and Pixel Resolution of the Optical Lens at Different Locations

      Distance (m)Optical ElementOptical Resolution (μm)Pixel Resolution (μm/pixel)
      5.1Shielding sheet136.17137.04
      3.7Vacuum window124.63125.42
    Tools

    Get Citation

    Copy Citation Text

    Fengdong Chen, Jingyang Sun, Qian Wang, Hongbo Zhu, Fa Zeng, Yueyue Han, Cheng Lu, Guodong Liu. In-situ laser-induced surface damage inspection based on image super-resolution and adaptive segmentation method[J]. Chinese Optics Letters, 2022, 20(7): 071101

    Download Citation

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

    Category: Imaging Systems and Image Processing

    Received: Dec. 31, 2021

    Accepted: Apr. 8, 2022

    Published Online: Apr. 29, 2022

    The Author Email: Fengdong Chen (chenfd@hit.edu.cn), Fa Zeng (cengfa@tsinghua.org.cn), Guodong Liu (lgd@hit.edu.cn)

    DOI:10.3788/COL202220.071101

    Topics