The new SAM3 model from Meta is blowing my mind Shown here: detecting putting greens, pools, and cars in Scottsdale from simple text prompts via Mapbox imagery R, Shiny, mapgl for the UI; Python backend via Qiusheng Wu's segment-geospatial package (thanks Qiusheng!)

I'll share how I did this at some point - perhaps blog, R package, workshop, or some combination It's clear to me though that the potential of SAM3 for geospatial is enormous

Very nice! Thank you for extending SamGeo to R

@Meta @Mapbox @giswqs You might find my PIMINTO app interesting.

@Meta @Mapbox @giswqs How good is the model generating a yardage book for that golf course?

@Meta @Mapbox @giswqs curious how this compares

@Meta @Mapbox @giswqs This is incredible! The text-to-object detection with SAM3 opens up so many possibilities for geospatial analysis. Love seeing R + Python working together here with Shiny and segment-geospatial 🔥

@Meta @Mapbox @giswqs Is this open sourced? I want to check out the UI code. Thanks

@Meta @Mapbox @giswqs I'm working on something similar but am still awaiting access to Sam3 on huggingface. How long did it take for you to gain access?

@Meta @Mapbox @giswqs I got it in minutes

@Meta @Mapbox @giswqs What's the tiles server 🤯 such a high resolution

@Meta @Mapbox @giswqs Is the aerial imagery up to date? Can I use this for roofs of residential properties in Florida?

@Meta @Mapbox @giswqs yes with a fresh update for many areas!

@Meta @Mapbox @giswqs This is huge. If I wanted to manually look at the aerial data and then have the ai find the solar on a zip code of houses, how would I go about doing that?

@Meta @Mapbox @giswqs Mind-blowing demo. Curious how you handle imagery freshness and keep detections robust as cities change.

@Meta @Mapbox @giswqs That’s awesome. Wondering how this can be integrated to do analysis on large drone collected photogrammetry sets. Would love to explore if you are open