GDAL - Geospatial Data Abstraction Library

• sudo apt install gdal
• export GDAL_NUM_THREADS=ALL_CPUS to parallelize

Vectors

• Prefer GPKG.
• Consider compressing GeoJSON files.

Reproject:

ogr2ogr -f GeoJSON out.json in.json -t_srs EPSG:4326
ogr2ogr -f GPKG out.gpkg in.gpkg -t_srs EPSG:4326

Note: This moves things by two meters for me, better use QGIS?

Rasters

• Use compression, DEFLATE for single-band (eg DSM), YCbCr JPEG for RGB (eg orthophotos).
• Use COG (Cloud Optimized GeoTIFF) for whatever needs to be opened in QGIS.

Inspect a GeoTIFF:

gdalinfo in.tif

See details of *.tif in the current dir:

for f in *.tif; do           
    printf "%s: %s %s with %s in EPSG:%s\n" "$f" \
        "$(gdalinfo -json "$f" | jq -r '"\(.size[0])x\(.size[1])"')" \
        "$(gdalinfo -json "$f" | jq -r '.metadata.IMAGE_STRUCTURE.LAYOUT')" \
        "$(gdalinfo -json "$f" | jq -r '.metadata.IMAGE_STRUCTURE.COMPRESSION')" \
        "$(gdalinfo -json "$f" | jq -r '.stac."proj:epsg"')"
done

Turn a single-band GeoTIFF into COG:

gdal_translate -of COG -co COMPRESS=DEFLATE -co PREDICTOR=YES in.tif out.tif

Turn a RGB GeoTIFF into COG with JPEG compression:

gdal_translate -of COG -co COMPRESS=JPEG in.tif out.tif

Merge GeoTIFFs:

gdalwarp -of COG -co COMPRESS=DEFLATE in*.tif out.tif

Clip GeoTIFF by mask

RGB:

gdalwarp -of GTIFF -co TILED=YES -overwrite -dstalpha \
    -cutline mask.gpkg -crop_to_cutline \
    in.tif /tmp/tmp_clipped.tif
gdal_translate -of COG -co COMPRESS=JPEG /tmp/tmp_clipped.tif out.tif

Single-band:

gdalwarp -of GTIFF -co TILED=YES -overwrite -dstalpha \
    -cutline mask.gpkg -crop_to_cutline \
    in.tif /tmp/tmp_clipped.tif
gdal_translate -of COG -co COMPRESS=DEFLATE -co PREDICTOR=YES \
    -b 1 -mask 2 /tmp/tmp_clipped.tif out.tif

Sum GeoTIFFs (eg sparse LiDAR point clouds)

Change the boundaries of a GeoTIFF (upper left x, upper left y, lower right x, lower right y):

gdalwarp -of GTiff -co COMPRESS=DEFLATE \
    -t_srs "EPSG:5514" \
    -te -602000 -1166000 -594000 -1155000 \
    -r near in.tif out.tif

Merge two GeoTIFFs with the same boundaries, taking max value from each:

gdal_calc.py -A in1.tif -B in2.tif --outfile=out.tif \
    --calc="maximum(A,B)" --format=GTiff --co="COMPRESS=DEFLATE"