Ask HN: Machine Learning and Carbon Sequestration?
Hey HN,
I'm a software engineer and have been recently inspired by the https://www.trilliontrees.org/ project.
On a recent consulting project, my friends at https://deepai.org/ were tasked with estimating the number of palm trees in a given region based on satellite imagery. Although the satellite imagery provided was not of the best quality, we were able to come up with a machine learning model that could estimate the quantity of palm trees reasonably well.
We think we could extend this model to track other types of trees and foliage, and perhaps extrapolate the amount of carbon sequestration occurring on a city-by-city/year-over-year basis. It could be a useful tool for academia as well as a nice visualization for the public via a web-based dashboard. Throw in some gamification and a leaderboard to show which cities are planting the most trees would be pretty doable as well.
Viable idea? Already been done?
Would love some suggestions from folks in the knowhow since my knowledge of climate change right now doesn't go much further than the fundamentals. It would be useful if you could somehow estimate the health, species and growth rate of those trees.