I'm working on MODIS NDVI composite data (single band). I converted the image to a numpy array and performed filtering operations. When i'm saving the new numpy array as an image I get a 3 band image instead of single band image. How can I save it as a single band image?

#read an image
img = cv2.imread("D:/trial/MOD09_NDVI.A2017001.tif")

cv2.namedWindow('Unfiltered Image', cv2.WINDOW_NORMAL)

cv2.imshow('Unfiltered Image',img)


#apply filter
myIMG = cv2.blur(img, (5,5))

cv2.namedWindow('Blur Image', cv2.WINDOW_NORMAL)

cv2.imshow('Blur Image', myIMG)


scipy.misc.imsave('D:/trial/blur_55.tif', myIMG)

scipy.misc.imsave('D:/trial/blur_55.tif', myIMG)

saves a 3 band image instead of original single band image


2 Answers 2


Why are you using for saving Scipy? I think you can use function imwrite from CV2.

For example:

cv2.imwrite('D:/trial/blur_55.tif', myIMG)

or you can use also another libraries. For exammple PIL or Matplotlib.

Example for Matplotlib:

from matplotlib import pyplot as plt
  • I was using imwrite as well but I needed to assign spatial references as well. I ended up using GDAL functions for it. Thank you for the reply Feb 16, 2018 at 9:54

You can use the gdal driver.Create(filename, xsize, ysize, [bands], [data_type], [options]) function.

For example:

from osgeo import gdal

driver = gdal.GetDriverByName('GTiff')
out_ds = driver.Create('raster.tif', array.shape[0], array.shape[1], 1)
out_band = out_ds.GetRasterBand(1)


del out_ds, out_band
  • I ended up using something similar. Thanks for the reply Feb 16, 2018 at 9:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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