[gdal-dev] Best way to GTiff options for hyperspectral data

[Scott]

You might create a Cloud Optimized GeoTiff (COG) from ENVI for 
frequently use band combinations, such as:

gdalwarp -f COG -co COMPRESSION=DEFLATE -t_srs EPSG:4326 \
    -src_band 23 -dst_band 1 \
    -src_band 130 -dst_band 2 \
    -src_band 420 -dst_band 3 \
[envi file] new.tif

Now, you have a COG that can be efficiently accessed over the web for 
visualization/analysis using software like Leaflet and 
georaster-for-leaflet.

You can can style (analyze) band(s) pixel values using 
georaster-for-leaflet by defining custom palettes. Note, the more bands 
and the larger the data type will mean more data. So, a single band, 
byte value will always load fastest. Doing something like a 3 band Int16 
isn't painfully slow. I've got tons of examples for COG (and FGB) at 
CloudNaitveMaps.com

https://leafletjs.com/
https://github.com/GeoTIFF/georaster-layer-for-leaflet


On 4/23/24 09:15, Jose Gomez-Dans via gdal-dev wrote:
> Hello,
> 
> I have quite a lot of very large hyperspectral data (>600 bands). The 
> data is in ENVI BIL format, but I would like to use GTiff for it, and 
> possibly access the data via HTTPS for processing and visualisation. I 
> notice that in the STAC world, it's one file per band, which gets a bit 
> tiring quickly ;-) Are there some recommendations for this kind of data? 
> I will need to look at a few bands "spatially" (e.g. build a false 
> colour RGB), but also quickly access the whole spectrum.
> 
> Thanks!
> Jose
> 
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/gdal-dev