Create map tiles

Create geographic and non-geographic map tiles from a file.

Usage

tile(
  file,
  tiles,
  zoom,
  crs = NULL,
  resume = FALSE,
  viewer = TRUE,
  georef = TRUE,
  ...
)

Arguments

file

character, input file.

tiles

character, output directory for generated tiles.

zoom

character, zoom levels. Example format: "3-7". See details.

crs

character, Proj4 string. Use this to force set the CRS of a loaded raster object from file in cases where the CRS is missing but known, to avoid defaulting to non-geographic tiling.

resume

logical, only generate missing tiles.

viewer

logical, also create preview.html adjacent to tiles directory for previewing tiles in the browser using Leaflet.

georef

logical, for non-geographic tiles only. If viewer = TRUE, then the Leaflet widget in preview.html will add map markers with coordinate labels on mouse click to assist with georeferencing of non-geographic tiles.

...

additional arguments for projected maps: reprojection method or any arguments to raster::RGB(), e.g. col and colNA. See details. Other additional arguments lng and lat can also be passed to the tile previewer. See tile_viewer() for details.

Value

nothing is returned but tiles are written to disk.

Details

This function supports both geographic and non-geographic tile generation. When file is a simple image file such as png, tile() generates non-geographic, simple CRS tiles. Files that can be loaded by the raster package yield geographic tiles as long as file has projection information. If the raster object's Proj4 string is NA, it falls back on non-geographic tile generation and a warning is thrown.

Choice of appropriate zoom levels for non-geographic image files may depend on the size of the image. A zoom value may be partially ignored for image files under certain conditions. For instance using the example map.png below, when passing strictly zoom = n where n is less than 3, this still generates tiles for zoom n up through 3.

Supported file types

Supported simple CRS/non-geographic image file types include png, jpg and bmp. For projected map data, supported file types include three types readable by the raster package: grd, tif, and nc (requires ncdf4). Other currently unsupported file types passed to file throw an error.

Raster file inputs

If a map file loadable by raster is a single-layer raster object, tile coloring is applied. To override default coloring of data and noData pixels, pass the additional arguments col and colNA to .... Multi-layer raster objects are rejected with an error message. The only exception is a three- or four-band raster, which is assumed to represent red, green, blue and alpha channels, respectively. In this case, processing will continue but coloring arguments are ignored as unnecessary.

Prior to tiling, a geographically-projected raster layer is reprojected to EPSG:4326 only if it has some other projection. The only reprojection argument available through ... is method, which can be "bilinear" (default) or"ngb". If complete control over reprojection is required, this should be done prior to passing the rasterized file to the tile function. Then no reprojection is performed by tile(). When file consists of RGB or RGBA bands, method is ignored if provided and reprojection uses nearest neighbor.

It is recommended to avoid using a projected 4-band RGBA raster file. However, the alpha channel appears to be ignored anyway. gdal2tiles gives an internal warning. Instead, create your RGBA raster file in unprojected form and it should pass through to gdal2tiles without any issues. Three-band RGB raster files are unaffected by reprojection.

Tiles and Leaflet

gdal2tiles generates TMS tiles. If expecting XYZ, for example when using with Leaflet, you can change the end of the URL to your hosted tiles from {z}/{x}/{y}.png to {z}/{x}/{-y}.png.

This function is supported by two different versions of gdal2tiles. There is the standard version, which generates geospatial tiles in TMS format. The other version of gdal2tiles} handles basic image files like a matrix of rows and columns, using a simple Cartesian coordinate system based on pixel dimensions of the image file. See the Leaflet JS library and leaflet` package documentation for working with custom tiles in Leaflet.

See also

view_tiles(), tile_viewer()

Examples

# non-geographic/simple CRS
x <- system.file("maps/map.png", package = "tiler")
tiles <- file.path(tempdir(), "tiles")
tile(x, tiles, "2-3")
#> Creating tiles. Please wait...
#> Creating tile viewer...
#> Complete.

if (FALSE) { # \dontrun{
# unprojected map
x <- system.file("maps/map_wgs84.tif", package = "tiler")
tile(x, tiles, 0)

# projected map
x <- system.file("maps/map_albers.tif", package = "tiler")
tile(x, tiles, 0)
} # }