Compute Histogram about raster-functions

[jdrisdelle]

Hi @csap2109. What version of our software are you using? I just tried it on 10.4.1 and it is applying the colormap properly. @akferoz

[Johannes-R-Schmid]

I got the newest arcgis version. If 10.4.1 is actually the newest, then I have it

Did you subtract ndvi-raster from each other? If yes, does your ndvi have scientific values (-1 - 1) ?

@jdrisdelle

[jdrisdelle]

@csap2109

In this case the problem is likely with scaling and bit depth. Colormaps are only supported with 8 bit rasters. Colormaps in the Python raster functions are setup using 4 arrays.

[pixel value][Red][Green][Blue]

In the colormap you are using we are using a range of 0-255 in the pixel value array. In order for this to work on your output pixel block you will have to scale the values. You can to this by adding one then multiplying by 127.5.

You can also do scaling after applying the Python raster function in your function chain.

Add the stretch function on top of your Python raster function. When you do this go to the general tab in the function and set the output pixel type to 8 bit unsigned. Add the colormap function on top of the stretch function and you should be good.

Hope this helps

[Johannes-R-Schmid]

@jdrisdelle
Ich habe ersteres Versucht, da der Stretch im Nachhinein nicht mein Ziel war. Ich benötige ein einziges Script, damit ich dieses im Nachhinein online als Widget importieren kann. Deshalb kann ich auch keine bestehenden Funktionen verwenden.

Bei Ersterem funktioniert es leider nicht, aber vllt habe ich es auch falsch in das Script integriert.
Könntest du mir den 2-Zeiler Code zeigen, den du meintest und vllt wo du es in meinem Script einbaue würdest. Wie gesagt, ich habe es verschiedenst ausprobiert und das Ergebnis war nicht wie erwünscht.

[jdrisdelle]

@csap2109

You will be able to access both out of the box functions and Python raster functions using widgets in your web application so you can still use the workflow that i sent in my previous message. If you would still like to do everything in your Python raster function you can scale you can do the following in your updatePixels() method.

r1 = np.array(pixelBlocks['r1_pixels'], dtype='f4', copy=False)
r2 = np.array(pixelBlocks['r2_pixels'], dtype='f4', copy=False)

np.seterr(divide='ignore')

outblock = self.op(r1, r2).astype(props['pixelType'], copy=False)

outBlock = (outBlock + 1) * 127.5

pixelBlocks['output_pixels'] = outBlock.astype(props['pixelType'])
return pixelBlocks