02-28-2017 04:06 AM
I am trying to creat a spatial model for drought monitoring using Land Surface Temperature(LST) model and NDVI. However I am unable to create a correlation between LST and NDVI using spatial modeler. Any idea on how to correlate the two output.
03-06-2017 11:35 AM
You could use the Correlation operator at the end of your model. This operator will return a correlation matrix of the layers in the input raster. You will need to stack the temperature raster and the NDVI raster together first as illustrated in the example below.
Hexagon Geospatial Support
03-08-2017 03:26 AM
I modified the model as proposed and run. The output was a an MTX file but what i had in mind as the output is a raster image showing areas having positive or negative correlation of the two variables. Find attached the modified model.
03-23-2017 08:06 AM
It may be possible to output a raster image showing areas of positive or negative correlation, but it won’t be easy. How would the different areas showing positive or negative correlation be defined? I suppose you could perform an unsupervised classification or image segmentation to define the regions. You would need to separate each class or segment into its own raster object to calculate the correlation between NDVI and temperature for that area. Then you would have to write the resulting correlation value to the pixels for each raster object that represents a class or segment and finally piece all of the individual raster objects back together again.