Temporal Patterns in LGLN Opendata

Abstract

Orthoimagery and airborne laser scanning data in Lower Saxony is acquired on a regular basis by LGLN, for many applications the acquisition date is important. We use R to query and plot this information for orthophotos, image-based height models and ALS-based height models.

About LGLN opendata

The State Office for Geoinformation and Land Surveying Lower Saxony (LGLN) provides a lot of its geospatial datasets under open licences. These include orthoimages (DOP), image-based surface height models (bDOM), as well as terrain (DGM) and surface (DOM) models derived from airborne laser scanning (ALS). In recent years the data has been made available as Cloud Optimized GeoTIFFs (COG) and referenced through SpatioTemporal Asset Catalogs (STAC).

We use R to access and explore the acquisition dates of these datasets. The index of available data - stored as a GeoJSON file in a cloud bucket - contains numerous polygons representing image tiles (1 km² or 2 km²) covering Lower Saxony. Each tile contains some basic metadata and links to further metadata as well as links to the image data itself. We fetch the index files for orthoimages (DOP), image-based height models (bDOM) and ALS-based height models (DGM/DOM) to examine acquisition timing.

More information about the data is available at opengeodata.lgln.niedersachsen.de.

Data Acquisition Timing

Due to phenological effects there are different demands to the data in vegetated areas. For orthophotos, data collection during the leaf-on season is preferred to get information from tree canopy such as vitality and to get better results during image-matching for height model processing. In contrast, terrain models derived from ALS data benefit from leaf-off conditions, when more laser pulses can penetrate the canopy and reach the ground surface.

DOP and bDOM originate from the same underlying imagery, but the available tiles and acquisition dates may differ because bDOM is generated only for a subset of images. Similarly, DGM and DOM are both derived from ALS data, but since they share the same acquisition basis, we only plot information for DGM, which can be considered representative of DOM as well.

Note

The data is also indexed in as STAC which allows for advanced searching, filtering and processing. It can be e.g. explored via Radiant Earth STAC-Browser.

load libraries

library(dplyr)
library(sf)
library(ggplot2)

get data

# read geojson from cloud bucket
get_data <- function(geojson){
  dat <- st_read(geojson) |> 
  # modify date
  mutate(
    date = lubridate::ymd(Aktualitaet),
    year = lubridate::year(date),
    month = lubridate::month(date, label = TRUE),
    doy = lubridate::yday(date)
  )
  return(dat)
}

dop20<- get_data("https://arcgis-geojson.s3.eu-de.cloud-object-storage.appdomain.cloud/dop20/lgln-opengeodata-dop20.geojson")

Reading layer `lgln-opengeodata-dop20-rgbi' from data source 
  `https://arcgis-geojson.s3.eu-de.cloud-object-storage.appdomain.cloud/dop20/lgln-opengeodata-dop20.geojson' 
  using driver `GeoJSON'
Simple feature collection with 65997 features and 6 fields
Geometry type: POLYGON
Dimension:     XY
Bounding box:  xmin: 340000 ymin: 5682000 xmax: 676000 ymax: 5980000
Projected CRS: ETRS89 / UTM zone 32N

get data

bdom20<- get_data("https://arcgis-geojson.s3.eu-de.cloud-object-storage.appdomain.cloud/bdom20/lgln-opengeodata-bdom20.geojson")

Reading layer `lgln-opengeodata-bdom20' from data source 
  `https://arcgis-geojson.s3.eu-de.cloud-object-storage.appdomain.cloud/bdom20/lgln-opengeodata-bdom20.geojson' 
  using driver `GeoJSON'
Simple feature collection with 101218 features and 4 fields
Geometry type: POLYGON
Dimension:     XY
Bounding box:  xmin: 340000 ymin: 5682000 xmax: 676000 ymax: 5980000
Projected CRS: ETRS89 / UTM zone 32N

get data

dgm1<- get_data("https://arcgis-geojson.s3.eu-de.cloud-object-storage.appdomain.cloud/dgm1/lgln-opengeodata-dgm1.geojson")

Reading layer `lgln-opengeodata-dgm1' from data source 
  `https://arcgis-geojson.s3.eu-de.cloud-object-storage.appdomain.cloud/dgm1/lgln-opengeodata-dgm1.geojson' 
  using driver `GeoJSON'
Simple feature collection with 49573 features and 4 fields
Geometry type: POLYGON
Dimension:     XY
Bounding box:  xmin: 341000 ymin: 5682000 xmax: 675000 ymax: 5972000
Projected CRS: ETRS89 / UTM zone 32N

Plot coverage by acquisitioin dates

The data catalog contains information about the acquisition date of images. We can plot this data to get more information about the available images. Acquisition dates for bDOM should be in principle the same as for DOP as the underlying image data is the same, however not all images were processed to bDOM as the necessary data quality (overlap) has not always been fulfilled.

Overview about acquisition dates

get overview

# first and last available acquisition dates
get_timespan <- function(dat){
  min <- dat|> arrange(date) |>  slice(1) |> pull(date)
  max <- dat|> arrange(desc(date)) |>  slice(1) |> pull(date)
  
  cat("Images available from", format(min), "to", format(max), "\n")
}

plot_distribution <- function(dat){
  dat |> 
  ggplot(aes(x = doy)) +
  geom_bar() +
  scale_x_continuous(
    breaks = c(1, 32, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335),
    labels = month.abb,
    limits = c(1, 365)
  ) +
  labs(x = NULL, y = NULL) +
  theme_void() +
  theme(
    axis.text.x = element_text(),
    legend.position = "top"
  )
}

Images available from 2012-03-15 to 2025-04-29

Distribution of orthophoto acquisitions in Lower Saxony by Day of Year

Images available from 2021-03-29 to 2025-04-29

Distribution of orthophoto acquisitions for height models in Lower Saxony by Day of Year

Images available from 2010-05-14 to 2022-04-23

Distribution of DGM/DOM data in Lower Saxony by Day of Year

Most recent coverage

plot latest data

# latest acquisition date per tile
plot_latest <- function(dat){
  dat|> 
  group_by(tile_id) |> 
  slice_max(date, n = 1) |> 
  ggplot() + 
    geom_sf(aes(fill = as.factor(year)), color = NA) + 
    theme_void()  +
    scale_fill_viridis_d(option = "plasma", name = "Year")
}

Most recent available orthophoto coverage in Lower Saxony

Most recent available image based height model coverage in Lower Saxony

Most recent available DGM/DOM data coverage in Lower Saxony

First coverage

plot first data

# first acquisition date per tile
plot_first <- function(dat){
  dat|> 
  group_by(tile_id) |> 
  slice_min(date, n = 1) |> 
  ggplot() + 
    geom_sf(aes(fill = as.factor(year)), color = NA) + 
    theme_void()  +
    scale_fill_viridis_d(option = "plasma", name = "Year")
}

First available orthophoto coverage in Lower Saxony

First available coverage of image based height model in Lower Saxony

First available DGM/DOM data in Lower Saxony

Leaf-off coverage

plot leaf-off data

# latest leaf-off data per tile
plot_leafoff <- function(dat){
  dat|>  
  mutate(Year = if_else(!between(doy, 115, 315), 
                        as.character(year), 
                        NA_character_)) |> 
  arrange(!is.na(Year), date) |> 
  ggplot() + 
    geom_sf(aes(fill = Year), color = NA) + 
    theme_void() +
    scale_fill_viridis_d(option = "plasma", na.value = "#f2f2f2", name = "Year")
}

Available orthophoto coverage in Lower Saxony during leaf-off season (mid November - mid April)

Available coverage of image based height model in Lower Saxony during leaf-off season (mid November - mid April)

Available coverage of DGM/DOM data in Lower Saxony during leaf-off season (mid November - mid April)

Leaf-on coverage

plot leaf-on data

# latest leaf-on data per tile
plot_leafon <- function(dat){
  dat|>
  mutate(Year = if_else(between(doy, 115, 315), 
                        as.character(year), 
                        NA_character_)) |> 
  arrange(!is.na(Year), date) |> 
  ggplot() + 
    geom_sf(aes(fill = Year), color = NA) + 
    theme_void() +
    scale_fill_viridis_d(option = "plasma", na.value = "#f2f2f2", name = "Year")
}

Available orthophoto coverage in Lower Saxony during leaf-on season (mid April - mid November)

Available coverage of image based height model in Lower Saxony during leaf-on season (mid April - mid November)

Available coverage of DGM/DOM data in Lower Saxony during leaf-on season (mid April - mid November)

Acquisition month

plot month

# month of latest data per tile
plot_month <- function(dat){
  dat|>  
  group_by(tile_id) |> 
  slice_max(date, n = 1) |> 
  ggplot() + 
    geom_sf( aes(fill = month), color = NA) + 
    theme_void()  +
    scale_fill_viridis_d(option = "plasma", name = "Month")
}

Latest orthophoto coverage in Lower Saxony

  Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Sep   Oct   Nov   Dec 
  328   164 14599 15078 16146  6989  7336  2834  2271     0   107   145

Latest coverage of image based height model in Lower Saxony

  Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Sep   Oct   Nov   Dec 
    0   656 32316 20160 24408  9960  4568  1932  6210     0   428   580

Latest coverage DGM/DOM data in Lower Saxony

  Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Sep   Oct   Nov   Dec 
 1167  6712 17020 22853  1656     0     0     0     0     0     0   165

Number of Images

plot number of images

# month of latest data per tile
plot_number <- function(dat){
  dat|>  
  st_drop_geometry() |>  # Work with data first
  group_by(tile_id) |> 
  summarise(n = n()) |> 
  left_join(dat|> select(tile_id) |> distinct(), by = "tile_id") |>
  st_as_sf() |>
  ggplot() + 
    geom_sf( aes(fill = as.factor(n)), color = NA) + 
    theme_void()  +
    scale_fill_viridis_d(option = "plasma", name = "# Campaigns")
}

Number of image based height models per tile

Citation

For attribution, please cite this work as:

Wiesehahn, Jens. 2023. “Temporal Patterns in LGLN Opendata.” March 1, 2023. https://wiesehahn.github.io/posts/lgln_opendata/.