1 / 25

Destriping VIIRS brightness temperatures for SST

Destriping VIIRS brightness temperatures for SST. Karlis Mikelsons Marouan Bouali Alex Ignatov Yury Kihai. STAR JPSS Annual Meeting College Park, MD May 14, 2014. Striping in VIIRS data. Low amplitude Unidirectional artifact Strongly affects gradients. Destriping Method.

galya
Download Presentation

Destriping VIIRS brightness temperatures for SST

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Destriping VIIRS brightnesstemperatures for SST KarlisMikelsons MarouanBouali Alex Ignatov YuryKihai STAR JPSS Annual Meeting College Park, MD May 14, 2014

  2. Striping in VIIRS data • Low amplitude • Unidirectional artifact • Strongly affects gradients Destriping of brightness temperatures...

  3. Destriping Method • Start with original “stripy” image • Calculate gradients • Discard “y” gradients in striped, but otherwise smooth regions • Poisson reconstruction (with DCT using FFT) yields approximate destriped image • Split the original image into destripedand striped components original image destriped component residual striped component Algorithm: M. Bouali, A. Ignatov, J. Atmos. Oceanic Technol., 31, 150-163 (2014). Destriping of brightness temperatures...

  4. Destriping method: iterative refinement • At each iteration, contribution to destriped image is extracted from residual striped component • Repeat until destriped component contains (nearly) all useful information and residual is (nearly) reduced to stripes residual striped component destriped component Algorithm: M. Bouali, A. Ignatov, J. Atmos. Oceanic Technol., 31, 150-163 (2014). Destriping of brightness temperatures...

  5. Destriping method: nonlinear filter residual (striped) filtered residual filtering domain nonlinear filter last iteration destriped component final destriped image Destriping of brightness temperatures...

  6. Results – VIIRS band M12 (3.7mm) Destriping of brightness temperatures...

  7. Results – VIIRS band M12 (3.7mm) Destriping of brightness temperatures...

  8. Results – VIIRS band M15 (10.75mm) Destriping of brightness temperatures...

  9. Results – VIIRS band M15 (10.75mm) Destriping of brightness temperatures...

  10. Results – VIIRS band M16 (12.01mm) Destriping of brightness temperatures...

  11. Results – VIIRS band M16 (12.01mm) Destriping of brightness temperatures...

  12. Results – VIIRS band M12 (3.7mm) – day (glint) Striping in glint region primarily due to different viewing angle for detectors Study: Q. Liu, C. Cao, F. Weng, J. Atmos. Oceanic Technol., 30, 2478-2487 (2013). Destriping of brightness temperatures...

  13. Results – VIIRS band M12 (3.7mm) – day (glint) • Areas outside the glint region and onset of glint region are destriped • High amplitude striping in the center of glint region is not removed Destriping of brightness temperatures...

  14. Results – VIIRS band M15 (10.76mm) – day longer wavelength bands unaffected by glint Destriping of brightness temperatures...

  15. Results – VIIRS band M15 (10.76mm) – day longer wavelength bands unaffected by glint Destriping of brightness temperatures...

  16. Results – VIIRS band M12 (3.7mm) – effect on cloud mask Cloud mask identification affected by striping Destriping of brightness temperatures...

  17. Results – VIIRS band M12 (3.7mm) – effect on cloud mask • Destriping removes artifacts in cloud mask • Also helps to identify larger areas as cloud free Destriping of brightness temperatures...

  18. Results – MODIS Aqua band 20 (3.75mm) Destriping of brightness temperatures...

  19. Results – MODIS Aqua band 20 (3.75mm) Destriping of brightness temperatures...

  20. Results – MODIS Aqua band 31 (11.0mm) Destriping of brightness temperatures...

  21. Results – MODIS Aqua band 31 (11.0mm) Destriping of brightness temperatures...

  22. Results – MODIS Aqua band 32 (12.0mm) Destriping of brightness temperatures...

  23. Results – MODIS Aqua band 32 (12.0mm) Destriping of brightness temperatures...

  24. Performance – IDL vs C • C code is about 6 times faster • I/O is a significant factor for C version: ≈25% time (VIIRS) and ≈40% time (MODIS) Destriping of brightness temperatures...

  25. Summary • Fast, operational production ready destriping code developed at NOAA • Capable of working with S-NPP VIIRS and Terra/Aqua MODIS • Initially prototyped in GPU-IDL (VIIRS: ×0.25; 2.5min/10min granule) • Now rewritten into C – 6 times faster than GPU-IDL (×0.06, 35sec/10min granule) • Implemented at STAR in experimental mode with S-NPP VIIRS and Terra/Aqua MODIS • Brightness temperature & SST imagery, ACSPO cloud mask, and SST gradients significantly improved • Next Steps • Immediate • Incorporate destriping code as a preprocessor for ACSPO VIIRS in NDE operations • Destripe “optional” IR bands (VIIRS: M13, M14; MODIS: B22, B23, B29) • Near term • Destripe solar bands for cloud cover (VIIRS: M6, M7; MODIS: B6, B7) • Address saw-like modulations in glint areas (short wavelength bands) • Optimize codes for reprocessing of historical VIIRS and MODIS data Destriping of brightness temperatures...

More Related