━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
                 DATA CLUSTERING FOR THE PRODUCTION OF
                           ECO-CLIMATIC MAPS
           ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━


Land cover map production over large areas (national or continental
scales) can be performed by automatic classification of high resolution
satellite image time series [1]. Machine learning algorithm performances
are better when an eco-climatic stratification is used: specialised
classifiers are trained for each eco-climatic area. The eco-climatic
partition comes usually from maps produced decades ago and therefore,
their pertinence can be argued.

The goal of the internship is to design a procedure for the generation
of eco-climatic map using data from climate, meteorological (Worldclim),
topographical (digital elevation models) and phenology (medium and
low-resolution vegetation indices) data bases. The method will use
unsupervised machine learning techniques (clustering). The maps produced
will be compared to existing ones and their usefulness for land cover
mapping will be evaluated using standard machine learning metrics.

• Skills: statistics, machine learning, image processing or remote
  sensing; Python programming
• Period: 5 months between February and September 2019
• Monthly compensation: 568.76 €
• Location: CESBIO, Toulouse.
• Applications: CV and motivation letter to jordi.inglada@cesbio.eu



Footnotes
─────────

[1] Inglada, J., Vincent, A., Arias, M., Tardy, B., Morin, D., &
Rodes, I., Operational high resolution land cover map production at
the country scale using satellite image time series, Remote Sensing,
9(1), 95 (2017). [http://dx.doi.org/10.3390/rs9010095]