Python geospatial analysis and remote sensing toolkit

Pygaarst is a Python library to simplify the manipulation of remote sensing and other geospatial data. Features include data and metadata access for a number of frequently used freely available satellite remote sensing data such as Landsat TM/ETM+/OLI/TIRS, VIIRS, and MODIS; access and processing of generic raster formats such as GeoTIFF, HDF-EOS, HDF5 and netCDF and combination with vector formats such as ESRI Shapefile and GeoJSON; georeferencing and re-projecting; combining raster and vector datasets; mapping and plotting.

Pygaarst relies heavily on numpy for data structures and GDAL for data access and geospatial functionality. The main goal is to provide a friendly API that makes common tasks on widely used data formats easy to accomplish in few lines of code.

The project was initiated on 2013-09-18 and in an early stage of development.

The documentation for this project can be found [on ReadTheDocs] 2.

NOTE: Most of the dependencies require the isntallation of C (or even Fortran) libraries beforehand.

Pygaarst requires Python 2.7. Python 3(>=3.3) is planned.

For all raster handling:

• numpy
• GDAL for gdal, ogr and osr
• pyproj

For HDF5 rasters (such as VIIRS):

• h5py

For HDF4 rasters (such as MODIS, ASTER)

• python-hdf4

For vector overlays:

• Shapely
• Fiona

Future probably:

• netCDF4

For plotting and mapping:

• matplotlib and mpl_toolkits.basemap for pygaarst.basemaputils

Documentation is also still in the very early stages of development.

The following modules are available and fulfil useful functions at the current time:

This is a REST-full web service client that implements the NASA LAADSWEB data API (http://ladsweb.nascom.nasa.gov/data/api.html)

class ModapsClient(object)

Usage:

from pygaarst import modapsclient as m
a = m.ModapsClient()
retvar = a.[methodname](args)

Implements the methods from http://ladsweb.nascom.nasa.gov/data/api.html , except for (currently) those related to OpenSearch (which don't appear to be working reliably server-side) and ordering (TBD). Implemented methods are marked with an x.

• x getAllOrders
• x getBands
• x getBrowse
• x getCollections
• x getDataLayers
• x getDateCoverage
• x getFileOnlineStatuses
• x getFileProperties
• x getFileUrls
• x getMaxSearchResults
• getOpenSearch
• getOrderStatus
• getOrderUrl
• getOSDD
• x getPostProcessingTypes
• x listCollections (deprecated)
• x listMapProjections
• x listProductGroups
• x listProducts
• x listProductsByInstrument
• x listReprojectionParameters
• x listSatelliteInstruments
• orderFiles
• orderFilesProcessed (deprecated)
• releaseOrder
• searchDatasets
• x searchForFiles
• x searchForFilesByName

This module provides several classes to represent common raster dataset formats. Implemented currently:

class GeoTiff(object)

A generic representation of a GeoTIFF file.

Args: filepath - path to GeoTiff file

class HDF5(object)

A representation of a HDF5 data file. This is currently tailored to NPP/VIIRS SDR or EDR files.

Args: filepath - path to HDF5 file

class Landsatscene(object)

A representation of a Landsat 4, 5, 7 or 8 scene for TM, ETM+ or OLI/TIRS sensor data. Will parse the metadata file in both pre and post August 2012 formats.

Args: dirname - name of the directory to which the scene files are unzipped

Individual Landsat bands (for the above sensors and Landsat versions) can be accessed via:

class Landsatband(GeoTIFF)

A representation of a single-band dataset. Is aware of its metadata. Inherits from GeoTIFF.

Further examples are provided in a [tutorial walkthrough] 1 of data access and map plotting for GeoTIFF and HDF5/VIIRS.

TODO:

• move the VIIRS-specific functionality to a separate class
• GeoTIFF improvements: clean handling of multiband datasets
• lat/long and native coordinate arrays for both "middle of pixel" and "corner" specs
• HDF-EOS, HDF4, netCDF
• MODIS and ASTER swath and gridded data including XML metadata file parsing

pygaarst.mtlutils.parsemeta(metadatafilepath)

Parses a metadata file and returns a dictionary of metadata values in the same nested structure as the USGS metadata file decription language. Used for pygaarst.raster.Landsatscene, pygaarst.raster.ALIscene and pygaarst.raster.Hyperionscene.

Contains some Landsat-specific data and helper functions that facilitate the conversion of digintal numbers (DN) to radiance, reflectance, and brighness temperature data, the calculation of vegetation indices and cloud masking.

Provides functions used by the pygaarst.raster.Landsatscene and pygaarst.raster.Landsatband classes, including:

pygaarst.landsatutils.get_bands(spacecraftid)

Returns permissible band labels (TM, ETM+ or OLI/TIRS) for a given spacecraft (L4, L5, L7 or L8).

pygaarst.landsatutils.getKconstants(spacecraftid)

Functions to work with infrared raster data.

pygaarst.irutils.gainbias(lmax, lmin, qcalmax, qcalmin)

pygaarst.irutils.dn2rad(data, gain, bias)

pygaarst.irutils.rad2kelvin(data, k1, k2)

pygaarst.irutils.rad2celsius(data, k1, k2)

TODO:

• shapefile class
• geojson class

pygaarst.basemaputils.map_interiorAK() returns a Basmap object for a map of the bulk of the Alaska mainland for convenient plotting.

(TODO: Other quick maps, possibility to override more parameters.)