Large-scale machine learning for point cloud processing

1 Like

Those ETH folks do all the coolest stuff!
I was out there last September and got to see the crazy field gantry system they had set up for crop research. They use one of those football stadium “spidercam” cabled camera rigs but they have it installed over a farm with a huge sensor payload on it with LiDAR and cameras and the whole kit.

They were just about to launch a new sensor platform with 13 cameras and hyperspectral on it.


That’s amazing! That type of system was part of my (failed/rejected) first thesis. Glad to see it in application, and to know I’m not actually nuts haha

I’m sure you’re not the only one! The entire field of plant phenomics is pretty much a decade of everyone seeing what seems possible and then finding out how every single part of the actual implementation turns into a nightmare of complexity :slight_smile:

I’m not sure I’ve been to any big phenomics facility where there is much of any alignment between what they originally envisioned and what is actually working at scale, (if at all!)