I have read many articles regarding this topic. But how can i dissagregate MODIS LST and Landsat LST in software environment. Please help me. Suggestions welcome for R, Matlab, ArcGIS, ERDAS Imagine and ENVI. Thank you.
Hi Susanta Mahato !
You can use machine learning techniques to do the downscaling, the methodology utilized in the article below will work good!
hope it helps!!
http://www.mdpi.com/2072-4292/8/3/215
Dear Susanta,
In R statistical software there are easy and difficult ways of downscaling/interpolation.
1) Easiest: use disaggregate(x, fact) function of package 'raster'. It can simply disaggregate the RasterLayer/RasterStack/RasterBrick, or you can set parameter 'method' to 'bilinear' for bilinear interpolation.
2) More difficult: you can use idw() function of package 'gstat' for IDW (Inverse Distance Weighted) and Tps() function of package 'fields' for TPS (Thin Plate Spline). Use raster::interpolate() after training the TPS model.
3) Difficult: Several type of kriging are available from package 'gstat'. Fitting a semivariogram model (fit.variogram()) will be needed before running function krige() or (functions gstat() and predict()).
If you have auxiliary variables, I recommend you to use regression kriging. This post may be of interest: https://www.researchgate.net/post/Can_I_downscale_climate_data_from_Worldclim_which_will_be_used_in_species_distribution_modelling#view=5908286b5b49523cc909005f
For more interpolation methods, please refer to this site:
http://rspatial.org/analysis/rst/4-interpolation.html
Here you can find a very good overview of the GIS packages of R:
https://cran.r-project.org/web/views/Spatial.html
HTH,
Ákos
Thank you sir... If I have any issue regarding the work then I will contact you...
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R