The goal of delorean is to make it easy to explore, download and
georeference french historical aerial photo from Remonter le temps -
IGN.
You can install the development version of delorean like so:
devtools::install_github("paul-carteron/delorean")library(delorean)
library(sf)
#> Linking to GEOS 3.13.1, GDAL 3.11.0, PROJ 9.6.0; sf_use_s2() is TRUE
library(terra)
#> terra 1.8.70
library(happign)
#> Please make sure you have an internet connection.
#> Use happign::get_last_news() to display latest geoservice news.delorean allows exploration of available aerial photos with
find_photos(). It intersects the bbox of x (area of interest) with
all photo footprints. Additional filters can be applied: year,
color, and oblique (see ?delorean::find_photos).
x <- get_apicarto_cadastre("29158")
all_photos <- find_photos(x)
photos_1997 <- find_photos(x, year = 1997)
plot(st_geometry(photos_1997), col = "grey90", border = "grey50", main = "1997 Photo footprints")
plot(st_geometry(x), col = "firebrick", add = TRUE)For downloading, use get_photos() with the source returned by
find_photos(). By default, the mode is set to raw, meaning the image
has no spatial reference.
# Using photo numero = 430 for the sake of example
photo_1997 <- photos_1997[photos_1997$numero == 430, ]
url <- photo_1997$url
filepath <- get_photos(url, outdir = tempdir())
#> Downloading [1/1] 0...10...20...30...40...50...60...70...80...90...100 - done.
photo <- rast(filepath)
#> Warning: [rast] unknown extent
plot(photo)Historical photos from IGN are provided with a centroid, resolution, and orientation. From these, a default georeferencing can be performed — but be aware it is significantly inaccurate ! (as shown below).
get_photos() offers two other modes: gcp and warp. The gcp mode
only attaches Ground Control Points (GCPs), so the photo is
georeferenced but not resampled onto a regular grid. The warp mode
takes longer but resamples the photo to a grid, ensuring compatibility
with most GIS software.
filepath <- get_photos(url, outdir = "C:\\Users\\PaulCarteron\\Desktop\\temp\\dolorean", mode = "warp")
#>
#>
#> Downloading [1/1] 0...10...20...30...40...50...60...70...80...90...100 - done.
#> Warping [1/1]
#> Warning in .warp(src_datasets, dst_filename, list(), t_srs, cl_arg, quiet): GDAL FAILURE 1: Deleting C:\Users\PaulCarteron\Desktop\temp\dolorean/IGNF_PVA_1-0__1997-04-10__CA97S00621_1997_F0517-0520_0430.tif failed:
#> Permission denied
#> 0...10...20...30...40...50...60...70...80...90...100 - done.
photo <- rast(filepath)
plot(photo)
plot(project(vect(x), crs(photo)), border = "red", lwd = 2, add = T)Because the default georeferencing remains significantly inaccurate,
delorean provides helpers to support manual georeferencing through the
function get_georef_helpers(). This wrapper around
happign::get_wfs() downloads prominent features that can assist in
georeferencing, such as hydrological elements, infrastructure, roads,
and buildings. All these layers are retrieved from IGN’s BD TOPO®
V3 dataset.
georef_helpers <- get_georef_helpers(x, "infra", "other")
#> Features downloaded : 81
#> Features downloaded : 32
#> Warning: No data found, NULL is returned.
#> Features downloaded : 19
plot(project(vect(x), crs(photo)), col = "grey90", border = "grey50", main = "Georef helpers on Penmarc'h")
plot(vect(georef_helpers$construction_lineaire), lwd = 2, col = "forestgreen", add = TRUE)
plot(vect(georef_helpers$construction_ponctuelle), col = "firebrick", add = TRUE)
plot(vect(georef_helpers$point_de_repere), col = "royalblue", add = TRUE)