2

I'm curious if anyone has suggestions for viewing and processing really large imagery. I'm pretty experienced with desktop apps like ENVI and am comfortable with python and GDAL but I'm analyzing dozens and dozens (soon to be hundreds) of DigitalGlobe images that are around 40 gigs per file and I'm struggling to effectively do data exploration because it takes so long to even view them. And then I'm not sure what the best environment is to handle hundreds of images this size.

Does anyone know any cloud based viewers and analysis sandboxes that you recommend? I'm getting access to some beefy machines over the next week or two and I'd prefer to do it locally. Overall I just wanted to get an idea of suggested pipelines for analyzing this much imagery.

  • 1
    Don't expect anything over the cloud to be fast, work locally! What format is the raster currently in? Have you built pyramids? What software do you want to consume it in (yes, it makes a difference, different packages have different format support)? Are you using this data scene-by-scene or are you trying to use as a mosaic covering your area of interest? – Michael Stimson Apr 23 '18 at 22:10
  • Don't built pyramids with a very large Imagery !!! it will cause your computer to freeze when you do that... – PROBERT Apr 23 '18 at 23:33
  • You have a point @PROBERT though the definition of very large is changing constantly; I recall when 4MB was considered large but recently I am using 300+GiB rasters and have worked with TiB+ sized rasters which built nicely. The process of building pyramids does depend on your software and settings though, using GDALADDO is reasonably quick even with rasters up to terrabyte size, Esri isn't too bad if you skip the first level though it would take about the same time to convert the format to ECW. Without pyramids I would not attempt to load these images into software like QGIS or ArcGIS. – Michael Stimson Apr 23 '18 at 23:57
  • @MichaelStimson - I didn't realized his tag was about gdal so that might be different compare to ArcGIS ! so I was just suggesting...but I would be very interesting to know if gdal can handle that big. – PROBERT Apr 24 '18 at 0:05
  • You might want to consider loading them loading them into an array database like rasdaman – nmtoken Apr 24 '18 at 6:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.