When plotting and RGB image using rasterio, I have seen that it is necessary to normalize the values first (0-1) and the use show. I am using the following code for that:

def norm(band):
    band_min, band_max = band.min(), band.max()
    return ((band - band_min)/(band_max - band_min))

b2 = norm(S_images[0].astype(numpy.float))
b3 = norm(S_images[1].astype(numpy.float))
b4 = norm(S_images[2].astype(numpy.float))

# Create RGB
rgb = numpy.dstack((b4,b3,b2))

# Visualize RGB

When reading the rasterio.plot.show documentation, there is a paramter called `adjust'

adjust ('linear' | None) – If the plotted data is an RGB image, adjust the values of each band so that they fall between 0 and 1 before plotting. If ‘linear’, values will be adjusted by the min / max of each band. If None, no adjustment will be applied.

Is this parameter doing the same as my norm function? Does it work only if I load a multi-chanel RGB image or can I use it when providing a np.stack?

1 Answer 1


Yes you can pass an array. The documentation specifies:

rasterio.plot.show(source, etc...)


  • source (array or dataset object opened in 'r' mode or Band or tuple(dataset, bidx))

Yes, it's the same. Demo using rasterio.plot.adjust_band that show uses to do the adjustment:

import rasterio.plot as rp
import numpy as np

def norm(band):
    band_min, band_max = band.min(), band.max()
    return ((band - band_min)/(band_max - band_min))

arr = np.arange(255, dtype=np.float32)
print(np.all(rp.adjust_band(arr) == norm(arr)))

  • 1
    Great, I can avoid then the my norm function. I was trying to use adjust has a parameter when calling rasterio.plot.show and that is why I was getting the error. Using first rasterio.plot.adjust_band(arr) to normalize the band and then using that output for the plot works fine. Thanks
    – GCGM
    Oct 9, 2019 at 7:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.