The R-package scapesClassification is designed to translate users’ mental models of seascapes, landscapes and, more generally, of geo-spaces into computer representations (classifications). Spaces and geographic objects are classified with user-defined rules taking into account spatial data as well as the spatial relationships existing among different classes and objects.

Raster surfaces and objects

  • scapesClassifications are performed on raster surfaces. The suit of available functions can be used to segment the raster space and to identify and manipulate distinct raster objects (i.e., groups of raster cells identified by a unique ID).

Key ideas

  • Landscapes and seascapes tend to have prominent features easy to identify. These features can be considered as anchor locations, locations around which a classification process can start and evolve.

  • A classification process can take into account the spatial relationships that are expected to exist among different classes, i.e., where a segment of space or an object is expected to be in relation to other segments and objects.

  • Based on such relationships, it is possible to estimate where a certain class is expected to exist and to perform focal evaluations of classification rules: rules are only evaluated at suitable locations, thus, limiting possible misclassification cases.

  • A classification process is seen as multi-step: as new portions of a raster are classified they can be used to define new focal areas over which classification rules are evaluated.

Get started

If you are just getting started with scapesClassification and you would like to have a general overview of the package capabilities you can consult the github page and the working example articles. For a deeper understanding of how the package works you can consult the implementation articles and the examples throughout the package documentation.

Installation

You can install the released version of scapesClassification from CRAN with:

install.packages("scapesClassification", dependencies = TRUE)

And the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("ghTaranto/scapesClassification", dependencies = TRUE)