- 'gdalUtilities::gdalinfo' causes an error with sf 1.0-19. Use 'sf::gdal_utils' instead in function get_gdal_info
- systematically use seq_len for loops and sequences of indices
- update fixed problem occurring when running map_alpha_div with default value nbCPU = 1
- update fixed problem occurring when performing PCA without mask
- fixed problem occurring when performing PCA without mask
- fixed issue related to plan(multisession), suggested by @rnedelec on github, following documented issue in ropensci/drake#255 (comment)
- option added to display progress bar. Set to FALSE as default
- added input variables Kmeans_info_path and Kmeans_info in map_spectral_species to allow production from previously computed clusters
- allow use of any vector format (including gpkg, shp... ) for validation plots provided in diversity_from_plots
- allow different crs between raster and vector layers
- update example_script and vignettes to fulfill removal to rgdal dependency in the package
- issue #18: remove the dependency to rgdal and rgeos using terra and gdalUtilities instead
- remove extract_pixels_coordinates.From.OGR, a unique extract_pixels_coordinates now works for both character or rast/vect inputs
- add get_gdal_info to get raster gdal info as a nested list.
- possibility to enable / disable progressbar in compute_spectral_species_FieldPlots and init_kmeans
- fixed computation of FD metrics
- removed import progress and emstreeR
- removed reference to multiprocess, multisession used exclusively from future
- fixed bug when producing simpson index from map_alpha_div function
- added importFrom corresponding to progressr in continum removal
- removed unused functions (VectorInRasterFootprint, get_BB, get_BB_from_fullImagem get_BB_from_Vector, get_polygonCoord_from_Shp, read_ListRasters, VectorInRasterFootprint, gdal_polygonizeR) in order to reduce dependency to rgeos
- these functions can be found in R package preprocS2
- fixed bug when computing functional diversity metrics from 1 component only
- function map_spectral_species checks if format and dimensions of Input_Mask_File are as expected
- function map_spectral_species now supports Input_Mask_File = FALSE. It writes a blank mask to ensure next processes
- fixed bug when defining SelectedPCs as a vector in map_spectral_species()
- Updated vignettes
- Updated alpha and beta diversity mapping when starting from a classification map
- now uses red band instead of red edge band for the computation of NDVI --> changes thresholding
- simplified inputs for functions such as map_spectral_species(), init_kmeans(), map_beta_div()
- updated vignettes
- optimized codes for the computation of spectral species, alpha and beta diversity
- addition of progress bars during the different steps of the computation of spectral species, alpha and beta diversity maps
- addition of functions Compute_ALPHA_SSD_per_window_list, Compute_ALPHA_SSD_per_window, Compute_ALPHA_per_window, prepare_HDR_SSD, prepare_HDR_Sunlit, RW_bytes_all, RW_bytes
- use multisession instead of multiprocess where it was forgotten
- use of list for computation of beta diversity
- addition of functions getBCdiss and Normalize_SSD
- added Hellinger distance as beta diversity metric
- removed NMDS as possible ordination method when computing beta diversity
- progress bar instead of messages
- future: multisession instead of multiprocess
- added tutorials for the PROGYSAT workshop
- install package dissUtils directly from github ('cran/dissUtils') as it was removed from official CRAN repo
- discard marginal spectral species based on number of sunlit pixels, instead of total number of pixels considered (window size or plot used for validation)
- harmonize default value for pcelim
- re-integrate package emstreeR (previously removed from CRAN) used to compute functional divergence
- uses future_lapply only if more than one CPU requested, otherwise use standard lapply
- fixed data type of spectral species derived from supervised classification and output directories for alpha and beta diversity
- fixed error occuring when using custom classification map instead of spectral species map
- Added dimMDS defining number of dimensions to run PCoA for beta diversity
- exported functions from beta library
- updated documentation
- corrected weighted distance from nearest neighbors: assign exact coordinates of a sample when dissimilarity = 0
- temporary: discarded evenness from functional diversity as it uses emstreeR and requires binding to mlpack
- Added functionality to allow for computation of spectral diversity maps based on spectral index stack or any raster stack
- added function to compute interquartile range (IQR) and identify outliers
- applied IQR instead of center reduction
- exported most of the functions in the Lib_ImageProcess
- Image sample was removed and placed in external repository
- Major update of vignettes for tutorial
- used file.path instead of paste
- capacity to produce diversity maps based on classification raster
- warning eliminated when testing Selected_Features in Lib_MapFunctionalDiversity.R
- Corrected data extraction function using nbands in extract.big_raster and get_random_subset_from_image
- updated extract.big_raster and get_random_subset_from_image in order to account for 2D rasters which cannot be read with brick
- added driver definition when using read_stars (extract.big_raster)
- corrected vignettes (TypePCA used before defined in previous version)
- added importFrom raster brick in function extract.big_raster
- Fixed bug when calling file.edit from linux terminal. initial parameter editor='internal' is not cross platforms. removed it.
- Added option to directly compute alpha and beta diversity maps from classification maps, even if it does not correspond to SpectralSpecies file produced from biodivMapR
- Fixed bug occuring when calling function diversity_from_plots if no functional duversity map was produced before
- Fixed bug occuring when writing image if initial raster is not a multiple of the window size. no bug occur but the raster files have wrong information
- fixed problem when performing estimation of biodiversity for plots outside of the raster: now the value is NA
- Added functional diversity metrics
- updated example script
- updated vignette
- implemented MNF
- discarded HDR as input variable from get_random_subset_from_image
- updated example file
- added contribution of F de Boissieu
- changed email address to teledetection.fr
- updated diversity_from_plots
- updated example script
- fixed identification and elimination of pixel samples with NA
- fixed bug occuring when input raster is BSQ interleave, and BSQ reported in PC file and following files
- finalized preparation for MNF
- improved information in the header files
- integrated stars package in order to read any file format, including TIFF format
- developed a generic function to write rasters
- prepared for MNF
- changed default red band for the computation of NDVI: closest band to 690 nm is now selected instead of closest band to 700 nm
- added option to convert the mask file into proper format with raster2BIL
- fixed bug raised when processing data over large number of pixels (image products >2^31 bits)
- fixed bug by adding elimination of bands disturbed by water vapor even when ContinuumRemoval set to FALSE
- Changed name of default directory when saving image after calling raster2BIL
- added documentation for raster data conversion using raster2BIL
- fixed bug in Lib_MapSpectralSpecies: remove constant bands from Subset$DataSubset if needed
- fix bug when continuum removal applied on pixels with constant values which were not filtered in previous stages
- Updated continuum removal
- Named Continuum_Removal instead of CR
- corrected documetation for perform_radiometric_filtering
- Added an error report when spectral information is not conform with expectations (due to too high noise level for example, usually leading to NaN or Inf values after PCA)
- Changed name for some variables in internal functions for consistency
- will add documentation on how to analyze error reports
- moved examples to repository root: example files are not installed with package anymore.
- removed Plots reprojection from
diversity_from_plots: now, Plots and Raster must be in the same projection. Changed example Plots projection accordingly. - removed
get_projection(useless)
- Added NEWS.md
- Updated README.md: transfered from gitlab.irstea to github
- changed return() into return(invisible())
- updated vignettes & tutorial with latest outputs and figures from code
First release in GitHub Submission accepted to Methods in Ecology and Evolution