numpy array to raster

MatthewBrown1

Hi,

I am iterating through a series of rasters and would like to mosaic the results, but don't have access to any tools that I would normally use because this is a Runtime SDK for WPF project using local server only. (Available tools: http://resources.arcgis.com/en/help/runtime-wpf/concepts/index.html#//01700000006r000000)

One option I have tried is numpy.concatenate but this is only partially successful:

[ATTACH=CONFIG]22625[/ATTACH]

r2 is in the correct location, but r1 has shifted. No idea where the backwards 'F' came from.

Here's the code I'm using:

import arcpy
import numpy
from arcpy.sa import *

r1 = Raster("ras1")
r2 = Raster("ras2")

arcpy.env.outputCoordinateSystem = r2.spatialReference

x = r2.extent.XMin
y = r2.extent.YMin

point = arcpy.Point(x,y)

a1 = arcpy.RasterToNumPyArray(r1)
a2 = arcpy.RasterToNumPyArray(r2)

b = numpy.concatenate((a1,a2))

c = arcpy.NumPyArrayToRaster(b,point,15)

Any ideas what I am missing?

Stock Movement.gif

MatthewBrown1

Hi Melanie,

Yes, both rasters are in the same projection/cell size as they have been extracted from the same raster. I had a go at changing the axis, and even thought axis=0 is correct, changing it to 1 made me realise that the raster extents are the problem. For some reason, the extract by mask tool did not update the raster extents and the original extents are still being used.

While I can fix this, I'm now wondering if the NoData cells are going to cause a problem with the alignment of the extracted rasters.

Update: seems like I'm also blocked by the differences in array dimension for each raster.