Working with Rasters

Making Rasters from Geometries

Turn geometric shapes into geo-registered raster images.

from import ocean_geom
from odc.geo.xr import rasterize

xx = rasterize(ocean_geom(), 0.5)
_ = xx.plot.imshow(aspect=2, size=3)

Creating PNG Images

  • Use odc.geo.xr.colorize() to turn data into RGBA image, matplotlib colormaps are supported

  • Use odc.geo.compress() to generate PNG data

  • We then display it with IPython.display.Image, but one can save to a file or send to an HTTP client from an API.

from import country_geom
from odc.geo.xr import rasterize
from IPython.display import Image

xx = rasterize(country_geom("AUS", "epsg:3577"), 20_000)
_images/raster-tools_1_0.png _images/raster-tools_1_1.png

Plotting on a map

import folium
import xarray as xr
from numpy.random import uniform
from import country_geom
from odc.geo.xr import rasterize

# Make some sample images
def gen_sample(iso3, crs="epsg:3857", res=60_000, vmin=0, vmax=1000):
   xx = rasterize(country_geom(iso3, crs), res)
   return xr.where(xx, uniform(vmin, vmax, size=xx.shape), float("nan")).astype("float32")

aus, png, nzl = [gen_sample(iso3) for iso3 in ["AUS", "PNG", "NZL"]]

# Create folium Map (ipyleaflet is also supported)
m = folium.Map()

# Plot each sample image with different colormap
aus.odc.add_to(m, opacity=0.5)
png.odc.add_to(m, opacity=0.5, cmap="spring", robust=True)   # vmin=2%,vmax=98%
nzl.odc.add_to(m, opacity=0.5, cmap="jet", vmin=0, vmax=800) # force vmin/vmax

# Zoom map to Australia
Make this Notebook Trusted to load map: File -> Trust Notebook

Saving Data

Use odc.geo.xr.write_cog() to generate cloud optimized GeoTIFF images. There is also in-memory version odc.geo.xr.to_cog() that returns compressed image bytes, useful for uploading data to the cloud storage.