🌐 Local Tile Server for Geospatial Rasters

Need to visualize a rather large raster (gigabytes) you have locally? This is for you.

This is a simple Flask application for serving tiles from local raster files in the Slippy Maps standard (e.g. /zoom/x/y.png)

🌟 Highlights

• Create a local tile server for large geospatial images
• Extract regions of interest (ROIs) interactively
• View local raster files with ipyleaflet

Under the hood, this uses large_image to launch a tile server in a background thread which will serve raster imagery to a tile viewer (see ipyleaflet examples below). This tile server can efficiently deliver varying levels of detail of your raster imagery to your viewer; it helps to have pre-tiled, Cloud Optimized GeoTIFFs (COG), but no wories if not as large_image will tile and cache for you when opening the raster.

There is an included, standalone web viewer leveraging GeoJS. You can use the web viewer to select and extract regions of interest from rasters.

⬇️ Installation

Install from PyPI: https://pypi.org/project/flask-tileserver/

pip install flask-tileserver

📝 A Brief Note on Installing GDAL

GDAL can be a pain in the 🍑 to install, and you may want to handle GDAL before installing flask-tileserver.

If on linux, I highly recommend using the large_image_wheels from Kitware.

pip install --find-links=https://girder.github.io/large_image_wheels --no-cache GDAL

Otherwise, I recommend using conda:

conda install -c conda-forge GDAL

💭 Feedback

Please share your thoughts and questions on the Discussions board. If you would like to report any bugs or make feature requests, please open an issue.

🚀 Usage

🖥️ Local Web Application

Launch the tileserver from the commandline to use the included web application where you can view the raster and extract regions of interest.

python -m tileserver path/to/raster.tif

🍃 ipyleaflet Tile Layers

There are utilities included here for launching a tile server as a background thread to serve image tiles from any raster file on your local file system. Further, I have included a utility for automatically launching a tile server and creating an ipyleaflet.TileLayer. Here is an example:

from tileserver import get_leaflet_tile_layer, TileServer
from ipyleaflet import Map, ScaleControl, FullScreenControl

# First, create a tile server from local raster file
tile_server = TileServer('~/Desktop/TC_NG_SFBay_US_Geo.tif')

# Create ipyleaflet tile layer from that server
t = get_leaflet_tile_layer(tile_server)

# Create ipyleaflet, add layers, add draw control, display
m = Map(center=tile_server.center())

m.add_layer(t)
m.add_control(ScaleControl(position='bottomleft'))
m.add_control(FullScreenControl())
m

Two Raster at Once

from tileserver import get_leaflet_tile_layer
from ipyleaflet import Map, ScaleControl, FullScreenControl, SplitMapControl

m = Map(
        center=(37.7249511580583, -122.27230466902257),
        zoom=9,
       )

# Create 2 tile layers from 2 separate raster files
l = get_leaflet_tile_layer('~/Desktop/TC_NG_SFBay_US_Geo.tif',
                           band=1, palette='matplotlib.Viridis_20', vmin=50, vmax=200)
r = get_leaflet_tile_layer('~/Desktop/small.tif',
                           band=2, palette='matplotlib.Plasma_6', vmin=0, vmax=150)

control = SplitMapControl(left_layer=l, right_layer=r)
m.add_control(control)

m.add_control(ScaleControl(position='bottomleft'))
m.add_control(FullScreenControl())
m

Note: the color palette choices come form palettable

📍 Using ipyleaflet for ROI Extraction

from tileserver import get_leaflet_tile_layer, TileServer
from ipyleaflet import Map, ScaleControl, FullScreenControl, DrawControl

# First, create a tile server from local raster file
tile_server = TileServer('~/Desktop/TC_NG_SFBay_US_Geo.tif')

# Create ipyleaflet tile layer from that server
t = get_leaflet_tile_layer(tile_server)

# Create ipyleaflet, add layers, add draw control, display
m = Map(
        center=(37.7249511580583, -122.27230466902257),
        zoom=9,
       )
m.add_layer(t)
m.add_control(ScaleControl(position='bottomleft'))
m.add_control(FullScreenControl())
draw_control = DrawControl()
m.add_control(draw_control)
m

from shapely.geometry import Polygon

# Inspect `draw_control.data` to get the ROI
bbox = draw_control.data[0]
p = Polygon([tuple(l) for l in bbox['geometry']['coordinates'][0]])
left, bottom, right, top = p.bounds

roi_path = tile_server.extract_roi(left, right, bottom, top)
roi_path

r = get_leaflet_tile_layer(roi_path)

m2 = Map(
        center=(37.7249511580583, -122.27230466902257),
        zoom=9,
       )
m2.add_layer(r)
m2.add_control(ScaleControl(position='bottomleft'))
m2.add_control(FullScreenControl())
m2

🗺️ Example Datasets

A few example datasets are included with tileserver. A particularly useful one has global elevation data which you can use to create high resolution Digital Elevation Models (DEMs) of a local region.

from tileserver import get_leaflet_tile_layer, examples
from ipyleaflet import Map, DrawControl

# Load example tile layer from publicly available DEM source
tile_server = examples.get_elevation()

# Create ipyleaflet tile layer from that server
t = get_leaflet_tile_layer(tile_server,
                           band=1, vmin=-500, vmax=5000,
                           palette='mycarta.Cube1_19',
                           opacity=0.75)

m = Map(zoom=2)
m.add_layer(t)
draw_control = DrawControl()
m.add_control(draw_control)
m

Then you can follow the same routine as described above to extract an ROI.

I zoomed in over Golden, Colorado and drew a polygon of the extent of the DEM I would like to create:

And perform the extraction:

from shapely.geometry import Polygon

# Inspect `draw_control.data` to get the ROI
bbox = draw_control.data[0]
p = Polygon([tuple(l) for l in bbox['geometry']['coordinates'][0]])
left, bottom, right, top = p.bounds

roi_path = tile_server.extract_roi(left, right, bottom, top)

r = get_leaflet_tile_layer(roi_path, band=1,
                           palette='mycarta.Cube1_19', opacity=0.75)

m2 = Map(
        center=(39.763427033262175, -105.20614908076823),
        zoom=12,
       )
m2.add_layer(r)
m2