vapour_read_raster: Raster IO (read)

In vapour: Access to the 'Geospatial Data Abstraction Library' ('GDAL')

Raster IO (read)

Description

Read a window of data from a GDAL raster source. The first argument is the source name and the second is a 6-element window of offset, source dimension, and output dimension.

Usage

vapour_read_raster(
  x,
  band = 1,
  window,
  resample = "nearestneighbour",
  ...,
  sds = NULL,
  native = FALSE,
  set_na = TRUE,
  band_output_type = "",
  unscale = TRUE,
  nara = FALSE
)

Arguments

x	data source
band	index of which band to read (1-based)
window	src_offset, src_dim, out_dim
resample	resampling method used (see details)
...	reserved
sds	index of subdataset to read (usually 1)
native	apply the full native window for read, FALSE by default
set_na	specify whether NA values should be set for the NODATA
band_output_type	numeric type of band to apply (else the native type if ”), is mapped to one of 'Byte', 'Int32', or 'Float64'
unscale	default is TRUE so native values will be converted by offset and scale to floating point
nara	logical whether to return a (scaled) nativeRaster

Details

The value of window may be input as only 4 elements, in which case the source dimension Will be used as the output dimension.

This is analogous to the rgdal function readGDAL with its arguments offset, region.dim and output.dim. There's no semantic wrapper for this in vapour, but see ⁠https://github.com/hypertidy/lazyraster⁠ for one approach.

Resampling options will depend on GDAL version, but currently 'NearestNeighbour' (default), 'Average', 'Bilinear', 'Cubic', 'CubicSpline', 'Gauss', 'Lanczos', 'Mode' are potentially available. These are compared internally by converting to lower-case. Detailed use of this is barely tried or tested with vapour, but is a standard facility used in GDAL. Easiest way to compare results is with gdal_translate.

There is no write support in vapour.

Currently the window argument is required. If this argument unspecified and native = TRUE then the default window specification will be used, the entire extent at native resolution. If 'window' is specified and native = TRUE then the window is used as-is, with a warning (native is ignored).

'band_output_type' can be 'raw', 'integer', 'double', or case-insensitive versions of the GDAL types 'Byte', 'UInt16', 'Int16', 'UInt32', 'Int32', 'Float32', or 'Float64'. These are mapped to one of the supported types 'Byte' ('== raw'), 'Int32' ('== integer'), or 'Float64' ('== double').

Value

list of numeric vectors (only one for 'band')

Examples

f <- system.file("extdata", "sst.tif", package = "vapour")
## a 5*5 window from a 10*10 region
vapour_read_raster(f, window = c(0, 0, 10, 10, 5, 5))
vapour_read_raster(f, window = c(0, 0, 10, 10, 5, 5), resample = "Lanczos")
## find the information first
ri <- vapour_raster_info(f)
str(matrix(vapour_read_raster(f, window = c(0, 0, ri$dimXY, ri$dimXY)), ri$dimXY[1]))
## the method can be used to up-sample as well
str(matrix(vapour_read_raster(f, window = c(0, 0, 10, 10, 15, 25)), 15))

vapour documentation built on June 22, 2024, 10:01 a.m.