rasterTools

Obtain and Process Earth Observation Data

The rasterTools package provides a toolkit to obtain and process spatial (earth observation) data in a transparent and reproducible manner.


Getting Started

  1. Install the development version from github via:

    devtools::install_github("EhrmannS/rasterTools")
  2. Read a brief introduction to understand the philosophy of rasterTools:

    ?`rasterTools-package`
  3. The vignettes given an in detail Introduction and explain what the logic behind landscape metrics is.


Example workflow

  1. (Down)load a range of gridded datasets:

    myDatasets <- list(list(operator = "oGFC", years = 2006)),
                       list(operator = "oMODIS", product = "mod17a3",
                            period = 2006, layer = 2),
                       ...)
    myData <- obtain(data = myDatasets, mask = aMask)
  2. Determine forest patches in a raster with continuous integer values of tree-cover:

    get_patches <- list(list(operator = "rBinarise", thresh = 30),
                        list(operator = "rPatches"))
    myInput <- rtData$continuous
    myPatches <- modify(input = myInput, by = get_patches, sequential = TRUE)
    visualise(raster::stack(myInput, myPatches))

  1. Compute the Class proportional area in a raster with categorial values:

    myInput <- rtData$categorial
    myMetric <- list(a_l = list(operator = "mArea", scale = "landscape"),
                     a_c = list(operator = "mArea", scale = "class"),
                     mCPA = "a_c / a_l * 100")
    measure(input = myInput, with = myMetric)
  2. Share your algorithms (or a gist thereof) on twitter with #rtAlgos.


Planned for future versions

  • Support of the Sentinel, Landsat and Lidar datasets.

  • Support of various “national forest inventory” datasets (Germany, France, Italy, Spain, yours?)

  • rTilify() to segregate a gidded dataset into another tiling, for instance to align datasets to each other or produce a hexagonal tiling of a rectangluarly tiled dataset.

  • You are encouraged to participate by writing for instance an obtain operator for your favourite dataset.


Acknowledgements

I am grateful for financial support from the PROFOUND Cost-action, which gave me the opportunity to work in a concentrated effort a large part of the functionality. This package has been developed in support of the FunBo Project, which was made possible by the Grünewald-Zuberbier Scholarship handed out by the University of Freiburg.

Thanks are due to Prof. Arne Pommerening who was a great source of inspiration for what rasterTools is now.