I am trying to estimate land surface evapotranspiration using Landsat 8 imagery. In the equation, we need to use LST, NDVI and DEM data, and so on. However, I should use them with the same spatial resolution. I`ve already got the NDVI value in 15 m resolution after GSS processing using Pan data. Besides, TIR (100 m) and DEM (30 m) data do not meet the 15 m resolution. So, my question is: Can I also use GSS technique to get TIR or LST and DEM data with 15m resolution? And in my case, in terms of upscaling and downscaling, which one is proper?
In my opinion, you can't use this technique. The sharperning follows the principle that a spatially fine but spectrally broad band (pan) integrates more or less the same 'amount of radiation' as a set of spatially coarse but spectrally fine bands (vnir or all bands within the pan band range). This is then not valid for the TIR or the DEM. Even if you use the first band of a PCA the correlation with the DEM is too weak. Just utilize bicubic interpolation to reduce your problem.
Cheers
Chris
yes can do but i feel fused image will be too coarse resolution because the ratio of resolution between two image is high. and to met 15 resolution you have to downscale the 100m and 30m resolution. 30m resolution to 15m is factor of 2 and at the same time 100m to 15m resolution is factor of 6 .i feel u loss solid amount of information when u downsample 100m to 15m .
In my opinion, you can't use this technique. The sharperning follows the principle that a spatially fine but spectrally broad band (pan) integrates more or less the same 'amount of radiation' as a set of spatially coarse but spectrally fine bands (vnir or all bands within the pan band range). This is then not valid for the TIR or the DEM. Even if you use the first band of a PCA the correlation with the DEM is too weak. Just utilize bicubic interpolation to reduce your problem.
Cheers
Chris
I am definitely agree with previous opinions. For TIR the factor of resolution change is quite big. TIR besides differs much from other channels VIS and NIR. Pansharpening can lead here to very strange results. I do not see rational possibility of fusion DTM in raster format with spectral images. What would be a data with higher resolution in pansharpening - brightness? You should use interpolation as mentioned before or find elevation source with a better resolution i.e. e DTM from airborne platform (image matching, LIDAR, interferometry) or satellite (1" SRTM or TerrasarX_TandemX height data)
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing