[gdal-dev] gdal_calc.py, calculating the mean value of masked arrays

[christian yrrman]

Dear All,


I'm trying to use gdal_calc.py to solve the following problem:

I have a time series of NDVI data as 8bit tiff files with different
BadData pixels (>50 and <250; this is the same for all layers) from
which I would like to calculate the mean.
(i.e. the same pixel might not be valid in all members of the time
series, and accordingly: Only sum up valid values and divide by the
number of valid layers, not all).

There exists a numpy array function that does exactly this:
[0] ma.masked_outside(d, 0.1, 0.9).mean()  # calculating the mean only
within the interval [0.1 - 0.9]

As far as I understand it, gdal_calc.py can also use numpy array
funtions for calculations, but I'm not sure how this should look like;
something like this:

Pseudo-code:
gdal_calc.py -A layer1.tif  -B layer2.tif -C layer3.tif
--outfile=test.tif --calc="masked_outside((A,B,C), 50,250).mean" ?


Do you think this is possible? Thanks in advance, and help highly appreciated!
:-)


Greetings, Chris



[0] http://docs.scipy.org/doc/numpy/reference/maskedarray.generic.html#data-with-a-given-value-representing-missing-data