lcfMapping

Repository for the MSc Thesis on land cover fraction mapping

Be aware: still under development!

Data

4 reference datasets used:

• IIASA (2015)
• IIASA (2015-2018)
• WUR (2015)
• WUR (2015-2018)

(Current) Usage

Scripts should at this moment be run in the following order:

1. temporalFilter.R
2. calcVIs.R
3. GetHarmonics.R

The first three scripts process the IIASA (2015) Landsat-8 timeseries and creates features.

4. loadValiData.R  (processes the WUR (2015) Landsat-8 timeseries and creates features)

5. spliTimeSeries.R  (splits the WUR (2015) Landsat-8 timeseries into yearly datasets)

6. wurChange.R  (processes the WUR (2015-2018) Landsat-8 timeseries and creates yearly datasets)

7. iiasaChange.R  (processes the IIASA (2015-2018) Landsat-8 timeseries and creates yearly datasets)

8. newFeatures.R  (calculates new features for all 4 reference datasets)

9. runRF.R  (random forest regression, basic and recurrent, run after processing of the Landsat-8 timeseries)
10. markov/Markov-August.ipynb or markov/Markov-final.ipynb  (runs Markov chain postprocessing)