Hi FME'ers,
Is there any way during using RasterMosaicker to use the maximum or minimum value of the entering ratser for the overlapping area instead of the value of the last raster which enters on top
Many Thanks
Abiy
GIS-Analyst
Page 1 / 1
I don't think so, I believe you'll need to sort the rasters before they enter the mosaicker so they are in the correct order before being conflated.
I am with Dave on this one, when mosaiking with FME the order of the features is key.
This is a shortcoming in FME and we plan to add additional modes for merging the values when there is an overlap while mosaiking. The issue is that the pure order of the rasters is not sufficient because you may want to make the decision pixel-by-pixel.
If I needed to do this within the current functionality, I would determine the area of overlap, clip the rasters to that area, then use a RasterExpressionEvaluator with a conditional statement to determine the max/min value and send the result of that to the RasterMosaiker on top of the original images.
This is a shortcoming in FME and we plan to add additional modes for merging the values when there is an overlap while mosaiking. The issue is that the pure order of the rasters is not sufficient because you may want to make the decision pixel-by-pixel.
It would be great if we could control the resolution too (max of inputs, min of inputs) rather than 1st feature in sets it.
We are tracking this internally via PR 54236
This will be in FME 2017.0 -- it is in the internal betas now. The commit comments say:
Added a new Overlapping Values option, which specifies how output cell values will be calculated when multiple rasters overlap. The choices are:
Last
Minimum
Maximum
Average
Sum
Composite Using Alpha Band