Pyresample is a python package for resampling geospatial image data. It is the primary method for resampling in the Satpy library, but can also be used as a standalone library. Resampling or reprojection is the process of mapping input geolocated data points to a new target geographic projection and area.
Pyresample can operate on both fixed grids of data and geolocated swath data.
To describe these data Pyresample uses various "geometry" objects including
the AreaDefinition
and SwathDefinition
classes.
Pyresample offers multiple resampling algorithms including:
- Nearest Neighbor
- Elliptical Weighted Average (EWA)
- Bilinear
For nearest neighbor and bilinear interpolation pyresample uses a kd-tree approach by using the fast KDTree implementation provided by the pykdtree library. Pyresample works with numpy arrays and numpy masked arrays. Interfaces to XArray objects (including dask array support) are provided in separate Resampler class interfaces and are in active development. Utility functions are available to easily plot data using Cartopy.
Pyresample is tested with Python 2.7 and 3.6, but should additionally work on Python 3.4+. Pyresample will drop Python 2.7 at the end of 2019.
See pytroll.github.io for more information on the PyTroll group and related packages.
Hoese, D., Raspaud, M., Lahtinen, P., Roberts, W., Lavergne, et al. (2020). pytroll/pyresample: Version 1.16.0. Zenodo. https://doi.org/10.5281/zenodo.3372769