No reviews yetBe the first to leave a review for Firewatch
Littlebird — The only full-context AI
The only full-context AI
Promoted
@berpj what a great and useful idea. A few questions for you:
1/ how are you analyzing the imagery to determine if there’s a fire? I’m guessing you’re using a CV API?
2/ are you planning to monetize the product? If so, how?
@asglenn thanks!
1/ All the Sentinel-2 imagery is available in this repository (https://registry.opendata.aws/se...). Basically, with Sentinel-2, the earth is divided into tiles (cf https://eatlas.org.au/data/uuid/...) and for each tile there are 13 images (each one corresponding to a specific spectral band: red, green, blue, infrared, etc)
Each time a tile is updated (it means one of the sat captured the area), my Python program downloads some of the bands and makes some processing (using Rasterio and Numpy) to determine if there are hot spots. It the image is clean (not too cloudy, etc) and hot spots were found, it's pushed to twitter.
2/ I don't plan to monetize this project
Report
@berpj thanks, I’m interested in your model. Still a really cool and useful product. Also, hope you’ll reconsider monetization at some point. There may be a way of doing it that still provides maximum benefit for free (advertisements or freemium services to certain users, etc)
Hey!
I'm the maker of Firewatch, a bot tweeting fires seen from space. It uses the images captured by Sentinel-2, a constellation of 2 earth-observation satellites.
I've also just added a dashboard with a map (link above).
Please share your feedbacks and questions. 😃
Deep Elon Musk
Deep Elon Musk