Using GDAL to merge 3 bands to form a multispectral image?

Question

I have 3 separate *.tif image files as grayscale. They are representing data acquired with a scanner. Each of them is a separate band of the scanner; red, green, and blue. I would like to reassemble them in a single RGB *.tif image file.

Using gdal_merge.py -o RGB.tif -co PHOTOMETRIC=RGB red.tif green.tif blue.tif

doesn't work; I end up with a single grayscale image which size is the sum of all individual images. How to achieve this simple operation?

(They are not geotiff, just simple tif files; there is no tfw)

Use -separate: Place each input file into a separate band. http://www.gdal.org/gdal_merge.html — user30184, Nov 30 '16 at 22:27

score 14 · Accepted Answer · edited Feb 15 '24 at 17:24

14

Per user30184's comment, you can use gdal_merge like so:

gdal_merge.py -separate  -o new_rgb.tif -co PHOTOMETRIC=RGB C:\input_r.tif C:\input_g.tif C:\input_b.tif

edited Feb 15 '24 at 17:24

André Caceres

103
4

answered Jan 14 '18 at 00:12

cm1

2,280
14
26

score 7 · Answer 2 · answered Mar 09 '19 at 22:04

7

I had MemoryError with gdal_merge.py -separate ... when used on big input. Different answer helped my issue:

gdalbuildvrt -separate RGB.vrt red.tif green.tif blue.tif
gdal_translate RGB.vrt RGB.tif

answered Mar 09 '19 at 22:04

Martin Ždila

572
3
11

+1. You should always use VRT's when possible. They're much less of a memory-hog. – Kyle Barron Dec 06 '19 at 22:43
This was not the case for OP, but when merging GeoTiff files you may want to specify the CRS with e.g. -a_srs epsg:1234 to avoid further issues. – Skippy le Grand Gourou Oct 08 '20 at 13:03

Using GDAL to merge 3 bands to form a multispectral image?

2 Answers2