Given a .tiff topographic raster image.

How to get the altitudes of the highest and lowest point/pixel ?


5 Answers 5


With Python, you can access raster statistics using the Python GDAL/OGR API.

from osgeo import gdal

# open raster and choose band to find min, max
raster = r'C:\path\to\your\geotiff.tif'
gtif = gdal.Open(raster)
srcband = gtif.GetRasterBand(1)

# Get raster statistics
stats = srcband.GetStatistics(True, True)

# Print the min, max, mean, stdev based on stats index
print "[ STATS ] =  Minimum=%.3f, Maximum=%.3f, Mean=%.3f, StdDev=%.3f" % (
    stats[0], stats[1], stats[2], stats[3])

>>> print "[ STATS ] =  Minimum=%.3f, Maximum=%.3f, Mean=%.3f, StdDev=%.3f" % (
...     stats[0], stats[1], stats[2], stats[3])
[ STATS ] =  Minimum=1.000, Maximum=4.000, Mean=1.886, StdDev=0.797
  • 1
    I used to sneer when reading answers using the Python GDAL/OGR API, thinking "Why would anyone suggest that when there's a perfectly good command-line way to do the same?" Now, that I have to programmatically do nearly everything, I'm super-grateful to answers like yours. Thanks! Apr 27, 2016 at 20:23

With bash alone, you can use :

gdalinfo -mm input.tif

It returns a range of infos within which is the string Computed Min/Max=-425.000,8771.000, for my Eurasian raster.

Some cleanup and you get your vertical min/max variables:

$zMin=`gdalinfo -mm ./input.tif | sed -ne 's/.*Computed Min\/Max=//p'| tr -d ' ' | cut -d "," -f 1 | cut -d . -f 1`
$zMax=`gdalinfo -mm ./input.tif | sed -ne 's/.*Computed Min\/Max=//p'| tr -d ' ' | cut -d "," -f 2 | cut -d . -f 1`
$echo $zMin $zMax
>-425 8771

I trimed both the digits after decimal point and the spaces in case via cut -d <separator> -f <selected_field> and tr -d ' ', respectively. Adapt as needed.

  • 3
    Strangely, this does not have to be the same as the metadata. For an original SRTM hgt file: Min=-32.000 Max=202.000 Computed Min/Max=-58.000,205.000. The computed values reflect the real content of the cells.
    – AndreJ
    Mar 27, 2014 at 8:43

If the stats have already been calculated:

gdalinfo input.tif

If they haven't been calculated yet, do:

gdal_translate -stats input.tif output.tif
gdalinfo output.tif
  • The altitude info is not there. Did you test it successfully ?
    – Hugolpz
    Mar 24, 2014 at 20:04
  • That should be the MINIMUM and MAXIMUM values returned by gdalinfo. You may have a problem if the NODATA value hasn't been set, in which case you may have a very low MINIMUM value (rarely a very high MAXIMUM instead). Mar 24, 2014 at 20:11
  • gdal_translate -stats SRTM_NE_250m.tif & gdalinfo -stats SRTM_NE_250m.tif fails here. No min/max. Found with gdalinfo -mm SRTM_NE_250m.tif and appears as Computed Min/Max=-425.000,8771.000 for my data.
    – Hugolpz
    Mar 24, 2014 at 20:16
  • 1
    Well, I don't have room to show all the details but using a SPOT4 scene: gdalinfo SPOT4_HRVIR1_042-209_0_080831_111638.tif gave metadata but no min-max. the same command with -mm produced the min-max. gdal_translate -stats SPOT4_HRVIR1_042-209_0_080831_111638.tif output.tif produces a new file and gdalinfo output.tif on that gives metadata with min-max values plus more Mar 26, 2014 at 15:56
  • 1
    I've posted a more complete example on my blog (mercergeoinfo.blogspot.com/2014/03/gdalinfo-and-statistics.html). If that's against the rules here then I apologise now. Feel free to remove this if it is the case. Mar 26, 2014 at 20:19

Using gdalinfo and jq you can get what you need in a succinct fashion. This assumes that band that has your min/max is the first one (change the index if not).

gdalinfo -json -mm  input.tif | jq .bands[0].computedMin

gdalinfo -json -mm  input.tif | jq .bands[0].computedMax
  • I needed to quote the jq selector gdalinfo -json -mm input.tif | jq '.bands[0].computedMax' Apr 6, 2023 at 12:18

For Maximum and Minimum Height I will Prefer this

from osgeo import gdal

raster = r'Path of your image'
gtif = gdal.Open(raster)
srcband = gtif.GetRasterBand(1)

stats = srcband.GetStatistics(True, True)

print ('[ STATS ] =  Minimum=%.3f, Maximum=%.3f, Mean=%.3f' % (
stats[0], stats[1], stats[2]))

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.