Reading, modifying and writing a geotiff with GDAL in python

Question

I'm trying the learn the ropes of Remote Sensing image processing using Python GDAL bindings and numpy. As a first attempt, I'm reading a Landsat8 geotiff file, do a simple manipulation and write the result to a new file. The code below appears to work fine, except that the original raster is dumped in the output file, rather than the manipulated raster.

Any comments or suggestions are welcome, but particularly notes on why the manipulated raster does not show in the result.

import os
import gdal

gdal.AllRegister()

file = "c:\~\LC81980242015071LGN00.tiff"
(fileRoot, fileExt) = os.path.splitext(file)
outFileName = fileRoot + "_mod" + fileExt

ds = gdal.Open(file)
band = ds.GetRasterBand(1)
arr = band.ReadAsArray()

[cols, rows] = arr.shape
arr_min = arr.Min()
arr_max = arr.Max()
arr_mean = int(arr.mean())

arr_out = numpy.where((arr < arr_mean), 10000, arr)

driver = gdal.GetDriverByName("GTiff")
outdata = driver.Create(outFileName, rows, cols, 1, gdal.GDT_UInt16)
outband = outdata.GetRasterBand(1)
outband.WriteArray(arr_out)
outdata = None

print arr_min
> 0
print arr_max
> 65535
print arr_mean
> 4856

I use Python 2.7.1 on a Windows 7 32 bit machine.

Answer 1

Your script is missing the ds.FlushCache method, that saves to disk what you have in memory at the end of the modifications. See below a corrected version of your example. Notice that I also added two lines to set projection and geotransform as input

import os
import gdal

file = "path+filename"
ds = gdal.Open(file)
band = ds.GetRasterBand(1)
arr = band.ReadAsArray()
[rows, cols] = arr.shape
arr_min = arr.min()
arr_max = arr.max()
arr_mean = int(arr.mean())
arr_out = numpy.where((arr < arr_mean), 10000, arr)
driver = gdal.GetDriverByName("GTiff")
outdata = driver.Create(outFileName, cols, rows, 1, gdal.GDT_UInt16)
outdata.SetGeoTransform(ds.GetGeoTransform())##sets same geotransform as input
outdata.SetProjection(ds.GetProjection())##sets same projection as input
outdata.GetRasterBand(1).WriteArray(arr_out)
outdata.GetRasterBand(1).SetNoDataValue(10000)##if you want these values transparent
outdata.FlushCache() ##saves to disk!!
outdata = None
band=None
ds=None