Topography: How to get the x,y,z coordinates of highest or lowest point/pixel of a raster GIS?

Question

Given a .tiff topographic raster image with mountains and deep sea, I am looking for the x,y,z specifics of the highest or lowest point.

The x,y,z values being:

• altitude (z-coordinate)
• latitude (y-coordinate)
• longitude (x-coordinate)

While gdalinfo raster.tif would give a summary of the bounding box and the highest / lowest elevation values (see my previous question here, and note comments below), I would like to find the elevation, latitude, longitude of the highest and/or lowest pixels of my raster.

Would be better if these values are got as isolated output integer. I think raster processing iterating on all pixels of the raster and keeping values of the z-highest point may be a good way to go.

Answer 1

Is Python an option?

Use RasterIO (a Python GDAL/ numpy bridge) to load the raster to NumPy array, then use numpy.amax() to find the maximum value, followed by numpy.where() to find the row/column indices, then calculate the lat and lon from the raster extents.

Answer 2

I recommend using Python or R (or a GIS software), as @Marc Pfister has suggested.

However, you can do it with bash and gdal only, and heavy usage of grep.

First get the Min/Max values without coordinates:

Obtain the Min / Max values with gdalinfo or gdalinfo -mm like explained in your other question about Min/Max values. Use grep (and possibly some awk) to extract the values.

Get their corresponding coordinates:

Convert your tiff file to xyz using gdal_translate. (Beware, large file size ahead!)

gdal_translate -of XYZ input.tiff /tmp/output.xyz

Now grep for the Minimum or the Maximum value. For example, if your Maximum value is 2878, type

cat /tmp/output.xyz | grep -E ' 2878$'

to get the corresponding Lat / Lon / Elevation output:
36.2358333333333249 -0.123611111111111116 2878

If you are working with a float dataset, you should adapt the grep expression in order to find rounded Min/Max values.

That's it.

Alternative if you got the pixel row/column coordinates with other software:

If you don't like the xyz file method, you could also first get the pixel row/column coordinates of the Min/Max values and then convert them to the corresponding lat/lon pair with gdaltransform:

For example, if you found out that the maximum value is at pixel coordinates x=12 y=300:
echo 12 300 | gdaltransform image.tif will output the corresponding geo coordinate pair depending on the CRS of your source image:
36.2358333333333249 -0.123611111111111116 0

If you keep this final coordinate conversion as an extra step, you may be able to use some other software packages (without geospatial features) (e.g. imagemagick) or a small Python / R script to locate pixel coordinates of the the Min/Max values in a more efficient way.

Answer 3

Since neither gdal_translate -stats SRTM_NE_250m.tif , nor gdalinfo -stats SRTM_NE_250m.tif, nor gdalinfo -mm SRTM_NE_250m.tif can work I think I have to iterare through the bitmap via a FOR_loop upon all pixels. The following may help.

• Gdal: How iterate through all pixels of a raster image?

• Gdal: How to get a pixel's value from raster image?

  $ gdallocationinfo crop.tif 50 50 -valonly

  73