EAlidaR

THIS PACKAGE IS NOW NOT WORKING DUE TO THE DEFRA PORTAL MOVING FROM THE ESRI PORTAL TO THE NEW OGC VERSION. DEVELOPMENT WILL CONTINUE AT: https://github.com/h-a-graham/gblidar WHERE WE HOPE TO BUILD ON THE OGC API

An R package to download EA LiDAR 'National LiDAR programme' (NLP) and composite data for England. So you’re aware, Data downloaded with this package is licensed under the Open Government License v 3.0. This package is in a somewhat developmental state so if things aren't working quite right, please start up a discussion or submit an issue

Background:

The Environment Agency (EA) provide wonderful high resolution, open source elevation datasets for much of England. At present, the best way to download this data is via the ESRI-based web map portal. This has some drawbacks - there are limits to the number of files that can be extracted at any given time, the spatial join between the requested area and available tiles is very slow and the data is provided in zipped files of varying raster formats (mainly ASCII and GeoTiff).

This package aims to provide a clean and easy way to download and interact directly with these excellent data in R. The challenge here is that the download URLs for LiDAR tiles are not static. This package uses {RSelenium} to automate the uploading of Ordnance Survey tiles that intersect a requested area.

Right now, this package only supports the most recent LiDAR data available. This includes the National LiDAR Programme (NLP) and Composite data. If 1m resolution is chosen, The 2020 Composite will be requested for DTM data; if the composite isn't available, NLP data will be requested. The DSM composite 2020/2019 has not yet been released. Therefore NLP, data is first requested before searching the 2017 composite. If 0.25, 0.5 or 2m resolution is chosen, Composite data will be requested.

Dependencies

At present, {EAlidar} only supports the chrome driver; you will therefore need to install Google Chrome to run this package.

Selenium requires java, therefore make sure to install Java before installing {EAlidaR}.

Installation

devtools::install_github('h-a-graham/EAlidaR')

Checking for available data

You can check the availability of data for your region by using check_coverage. To see national scale coverage use national_covaerage. At present, these functions only display the <2019 Composite data, and don't include NLP extents. For more information on data coverage see this web portal. At present composite data coverage for 1 m and 2 m resolutions is 87% and 81% respectively. With the inclusion of NLP data this is even higher! More info here

A note on Selenium, JAVA and the Chrome driver...

So, this package uses the RSelenium::rsDriver() function to open and manage the chrome driver. If you are getting errors relating to {Rselenium} then you may have to update/reinstall Java.

In windows, remove the folder 'C:/ProgramData/Oracle', remove references to javapath in your Path Environment and then reinstall java.

For Debian/Ubuntu the standard Java install steps should be sufficient.

The function find_chrome_v() can be used to retrieve your system's version of Chrome;This function is the default value of the 'chrome_version' argument in all the download functions. If this function does not work on your OS (only tested on Windows and Ubuntu) or returns an incorrect value, you can manually specify the version using the 'chrome_version' argument. To do this simply open chrome > menu button > help > About Google Chrome. Then inspect the version number and then run binman::list_versions('chromedriver'). Select the driver version which most closely matches your machine's Google Chrome version and include it as a character string in the 'chrome_version' argument. Some issues still remain for OSX and MacOS...

Examples:

Here is a simple use case where we download the available 2m DTM data for one of the example regions provided with the package Ashop_sf. Using the get_area() function, we retrieve a single raster as ‘merge_tiles’ is TRUE. We can save this data in a desired location with ‘dest_folder’, ‘out_name’ and ‘ras_format’ arguments but, in this case, rasters are stored in the tempfile() location and will be available only during the active R session (unless subsequently saved with raster::writeRaster).

library(EAlidaR)

# national_coverage(model_type = 'DSM', resolution = 2) # quite slow by the way...
# check_coverage(poly_area = Ashop_sf, model_type = 'DTM', resolution = 2) 

Ashop_Ras <- get_area(poly_area = Ashop_sf, resolution = 2, model_type = 'DTM', 
                      merge_tiles=TRUE, crop=TRUE)

terra::plot(Ashop_Ras, col=sun_rise())
plot(Ashop_sf, add = TRUE)

Alternatively, the functions get_OS_tile_5km() and get_OS_tile_10km() allow the users to specify 5 or 10m Ordnance Survey (OS) tile name(s) as a vector:

NY20nw <- get_OS_tile_5km(OS_5km_tile = c('NY20nw','NY10ne)', resolution = 1, model_type = 'DTM')

NY20 <- get_OS_tile_10km(OS_10km_tile = 'NY20', resolution = 1, model_type = 'DTM')

To download data around a specific location use get_from_xy(). The XY coordinates must be provided in OSGB/British National Grid (Lat, Long) format:

Scafell_Peak <- get_from_xy(xy=c(321555, 507208), radius = 500, resolution = 1, model_type = 'DSM')

Some Extras...

And just to really show off how great this data is, here are some 3D examples with the brilliant {rayshader} package.

First let’s try out the Ashop Valley data we downloaded previously. Note that multicore is set to TRUE, in these examples, as they are quite large rasters - set to FALSE if you don’t want to use multiprocessing.

library(rayshader)

AshopMat = raster_to_matrix(raster::raster(Ashop_Ras)) 

AshopMat %>%
  sphere_shade(texture = "imhof1") %>%
  add_shadow(ray_shade(AshopMat, zscale = 1, multicore =TRUE), 0.3) %>%
  add_shadow(ambient_shade(AshopMat, multicore=TRUE), 0) %>%
  plot_3d(AshopMat, zscale = 1.5, fov = 60, theta = 45, phi = 20, windowsize = c(1000, 800), zoom = 0.2,
          solid = FALSE)

Sys.sleep(0.2)
render_depth(focus = 0.7, focallength = 70, clear = TRUE)

In some parts of England you can download <1m resolution data - here is an example using the for the City of London using the get_from_xy() function

CoL <- get_from_xy(xy=c(532489 , 181358), radius = 500, resolution=0.5, model_type = 'DSM')

CoL_Mat = raster_to_matrix(raster::raster(CoL))

CoL_Mat %>%
  sphere_shade(texture = "bw") %>%
  add_shadow(ray_shade(CoL_Mat, zscale = 1, multicore =TRUE), 0.3) %>%
  add_shadow(ambient_shade(CoL_Mat, multicore=TRUE), 0.1) %>%
  plot_3d(CoL_Mat, zscale = 1, fov = 60, theta = 20, phi = 30, windowsize = c(1000, 800), zoom = 0.3,
          solid = FALSE)

Sys.sleep(0.2)
render_depth(focus = 0.7, focallength = 70, clear = TRUE)