I have been struggling with this problem for some time - I am rather new to ArcGIS Pro and apologize in advance if this is trivial, but I really need to find out what I am doing wrong...
I have a set of Sentinel 2 images with just one band each (NDVI or NDRE) and a set of drone photos that are multispectral.
I used Extract Band to isolate NDVI respectively NDRE in the drone photos as well and now I want to combine drone and sattelite derived rasters so I can use Chart to see how well they fit together (they were taken at almost the same time over the same fields and I need to make an estimate of the correlation between NDVI/NDRE in the two...).
But when I try to do this I end up with a new raster that is identical to the original one-band drone raster..?
Let me understand this – before I suggest you a solution:
1) You have several Sentinel 2 images and from those you DERIVE NDVI/NDRE … so the result is one-band float data?
Q1: Do you have those in ONE MosaicDataset and derive the result with a raster function from it?
2) You have several drone images, 4 band at least with NIR and DERIVE NDVI/NDRE from those too – result being one-band float again?
Q2: Same as Q1
The easiest now would be to substract those from each other – by creating either (Option1 ) a referenced Mosaic on one of the two above and there use the
Arithmetic function to substract the other, or (Option 2) do the same integrated into one of the primary Mosaics. – my opinion is the first one is clearer for future use and maintenance and the step with the charts to follow
You can then use the chart function on the Mosaic to get you “a is higher – both are equal – b is higher” … and could add another raster function to that Referenced Mosaic (again Option 1) that classifies the result to your liking (REMAP and then add a ATTRIBUTE TABLE).
Does that help?
Thanks for your input so far!
To answer your questions first:
1) Yes, I have Sentinel 2 data as one-band float data rasters, one for NDVI and one for NDRE in each case (chosen from dates with minimal cloud coverage around the times where drone data where collected...).
Ad Q1: I do not have a mosaic dataset but several individual rasters, cut according to field polygons, as I only have a few crop fields til cover... (I might be mistaken here)
2) The original drone orthomosaics that I received are 8 band rasters (NDVI/NDRE/Blue/Green/Red/RedEdge/NIR/Temperature C) so, yes, they contain derived data as the first two bands...
Those do not cover the fields completely and are higher resolution than the satellite data (5x5 meters vs 10x10 meters) but they are georeferenced in the same coordinate system and also cut with the field polygons.
Ad Q2: From these I have extracted Band 1 and Band 2 respectively to make new single band rasters with each index.
I then tried to combine (with Band Composite) two single band rasters in each case - a drone and a satellite raster, both containing either NDVI or NDRE for the same area (although not fully same extent) in order to get one raster from which I could generate a Chart comparing Band 1 and Band 2 (which I expected to be two instances of NDVI respectively NDRE registered in the same field in each case...). But as I tried to explain they only contain the same (Band 1) data anyway...
I am not sure that it helps me yet - but at least it gets me thinking further about what is happening 🙂