Energy Data Analytics Lab's tutorials on geo-spatial operations

Table of contents

• Geospatial set operations
• Geospatial plotting
• Geo-coding
• Rasterization and vectorization

Geospatial set operations

Find the intersection/union/difference between two geospatial dataset and output a geospatial file

Useful package:

• geopandas: Set-Operations with Overlay

Example:

• Find out how many building there are in each image tile

Geospatial plotting

Plot points/polygons/any other shapes on a map

Useful package:

• geopandas
• cartopy
• matplotlib's Basemap: not recommended, has been depreciated and replaced by cartopy

Example:

• Plot sampling locations of Connecticut annotated tiles on Connecticut town-level map
• Plot density map of solar PV panel area at any resolution

Geo-coding

Use case: Convert pixel coordinates of polygons/points/lines to geospatial coordinates

Useful package:

• geopandas

Example:

• Convert polygon's pixel coordinates to geospatial coordinates
• register_to_tile() in generatePolygons.py

Rasterization and vectorization

Convert between vector shapes and raster images

Useful package:

• geopandas
• rasterio
• opencv
• skimage

Example:

Rasterization:

• pyimannotate's binarymask.py: converts manual polygon annotations to binary raster masks
• Script for rasterizing GeoTIFF images

vectorization:

• make_polygons() in posrProcUtils.py
• opencv's cv2.findContours
• Usually coupled with a polygon approximation process based on the Douglas-Peucker algorithm to reduce the number of vertices. E.g. skimage.measure.approximate_polygon().

Managing projections

Geospatial datasets sometimes come with different coordinates reference systems (CRS). Convert them to the same CRS before making any overlay analysis or plotting.

Useful package

• geopandas

Example:

• Managing Projections with geopandas