Skip to content

Latest commit

 

History

History

satellite_imagery

README.md

Satellite Imagery

This directory documents my calculation of the NDVI of Munich. Via USGS Earth Explorer I retrieve the proper satellite imagery form Landsat 8 – Wikipedia. The steps include:

  • Cropping Munich's geopolygon
  • Preprocessing data for higher quality results (f.i. not whole 16bit color range is exploited)
  • Calculating an index, f.e. NDVI – Wikipedia or NDWI – Wikipedia, evaluating vegetation or water respectively
  • Saving image/generating plot I use numpy, rasterio, geopandas, shapely and matplotlib to conduct all operations.

Roadmap

Exploring more indices (maybe related to water)
Evaluate quantitatively the NDVI Use a Geopolygon to filter the city of Munich and calculate it's NDVI.

Overview

  • adjust_values.py: Satellite data must be clipped to improve quality. Full explanation given in the file directly.
  • geojson2shapefile_downsampling.py: Converts a GeoJSON file of the city of Munich containing it's districts into a Shapefile resembling the border of Munich (without districts) and applies downsampling because the USGS Earth Explorer only permits <500 vertices.
  • isolate_shape.py: Corps a geometry saved as a shapefile from a GeoTIFF and also it's according mask as boolean array.
  • Index.py: Class resembling the logic needed to calculate and process an Index like the NDVI.
  • index_over_time.py Calculate the difference over time of consecutive indices (s. #### NDVI over time)
  • [WIP] make_rgb.py: Combines the red, green and blue bands to an RGB file.

Produced images

NDVI

The NDVI, data was collected on 2022-05-15, the greener, the more (healthy) vegetation: (matplotlib's scaling of the image is lower than the original file. I don't know why and how to zoom/scale it.)

NDVI over time

The image below shows the development of the NDVI over two periods in Munich and it's surroundings. Red indicates a decrease in the NDVI, green an increase, ie. reduced/improved vegetation health. (The NDVI is in [0, 1], hence it's difference in [-1, 1].)

NDWI

The Normalized difference water index – Wikipedia is used to monitor changes related to water content in water bodies, using green and NIR wavelengths. Data is from 2022-05-15. Clearly visible is the Isar, some lakes and in the north the olympic regatta area.
Disclaimer: The values were multiplied by 10 to increase visibility.