skip to navigation
skip to content

rasterio 0.8

Fast and direct raster I/O for Python programmers who use Numpy

Latest Version: 1.0a5

Rasterio employs GDAL under the hood for file I/O and raster formatting. Its functions typically accept and return Numpy ndarrays. Rasterio is designed to make working with geospatial raster data more productive and more fun.


Here’s an example of the basic features rasterio provides. Three bands are read from an image and summed to produce something like a panchromatic band. This new band is then written to a new single band TIFF.

import numpy
import rasterio
import subprocess

# Register format drivers with a context manager

with rasterio.drivers():

    # Read raster bands directly to Numpy arrays.
    with'rasterio/tests/data/RGB.byte.tif') as src:
        b, g, r = map(src.read_band, (1, 2, 3))

    # Combine arrays using the 'iadd' ufunc. Expecting that the sum
    # will exceed the 8-bit integer range, initialize it as 16-bit.
    # Adding other arrays to it in-place converts those arrays up
    # and preserves the type of the total array.

    total = numpy.zeros(r.shape, dtype=rasterio.uint16)
    for band in (r, g, b):
        total += band
    total /= 3
    assert total.dtype == rasterio.uint16

    # Write the product as a raster band to a new 8-bit file. For
    # keyword arguments, we start with the meta attributes of the
    # source file, but then change the band count to 1, set the
    # dtype to uint8, and specify LZW compression.

    kwargs = src.meta

    with'example-total.tif', 'w', **kwargs) as dst:
        dst.write_band(1, total.astype(rasterio.uint8))

# At the end of the ``with rasterio.drivers()`` block, context
# manager exits and all drivers are de-registered.

# Dump out gdalinfo's report card and open the image.

info = subprocess.check_output(
    ['gdalinfo', '-stats', 'example-total.tif'])
print(info)['open', 'example-total.tif'])

The rasterio.drivers() function and context manager are new in 0.5. The example above shows the way to use it to register and de-register drivers in a deterministic and efficient way. Code written for rasterio 0.4 will continue to work: opened raster datasets may manage the global driver registry if no other manager is present.

Simple access is provided to properties of a geospatial raster file.

with rasterio.drivers():

    with'rasterio/tests/data/RGB.byte.tif') as src:
        print(src.width, src.height)

# Output:
# (791, 718)
# {u'units': u'm', u'no_defs': True, u'ellps': u'WGS84', u'proj': u'utm', u'zone': 18}
# [101985.0, 300.0379266750948, 0.0, 2826915.0, 0.0, -300.041782729805]
# 3
# [1, 2, 3]

Rasterio also affords conversion of GeoTIFFs, on copy, to other formats.

with rasterio.drivers():

        driver='JPEG')['open', 'example-total.jpg'])


The rio_insp program opens the hood of any raster dataset so you can poke around using Python.

$ rio_insp rasterio/tests/data/RGB.byte.tif
Rasterio 0.8 Interactive Inspector (Python 3.3.5)
Type "src.meta", "src.read_band(1)", or "help(src)" for more information.
>>> src.shape
(1024, 1024)
>>> import pprint
>>> pprint.pprint(
{u'a': 6378137,
 u'b': 6378137,
 u'k': 1,
 u'lat_ts': 0,
 u'lon_0': 0,
 u'nadgrids': u'@null',
 u'no_defs': True,
 u'proj': u'merc',
 u'units': u'm',
 u'wktext': True,
 u'x_0': 0,
 u'y_0': 0}
>>> b = src.read_band(1)
>>> b
array([[255, 255, 255, ...,   0,   0,   0],
       [255, 255, 255, ...,   0,   0,   0],
       [255, 255, 255, ...,   0,   0,   0],
       [255, 255, 255, ..., 255, 255, 255],
       [255, 255, 255, ..., 255, 255, 255],
       [255, 255, 255, ..., 255, 255, 255]], dtype=uint8)
>>> b.min(), b.max(), b.mean()
(0, 255, 224.75362300872803)


C library dependecies:

  • GDAL

Python package dependencies (see also requirements.txt):

  • Numpy
  • setuptools

Development also requires (see requirements-dev.txt)

  • Cython
  • pytest


Rasterio is a C extension and to install on Linux or OS X you’ll need a working compiler (XCode on OS X etc). Unofficial Windows binary packages created by Christoph Gohlke are available here.

To install from the source distribution on PyPI, do the following:

$ pip install -r
$ pip install rasterio>=0.8

To install from a forked repo, do this (in a virtualenv, preferably):

$ pip install -r requirements-dev.txt
$ python install

The Numpy headers are required to run the rasterio setup script. Numpy has to be installed (via the indicated requirements file) before rasterio can be installed. See rasterio’s Travis configuration for more guidance.


From the repo directory, run py.test

$ py.test







File Type Py Version Uploaded on Size
rasterio-0.8.tar.gz (md5) Source 2014-03-31 352KB