Mazzap is a proof-of-concept open-source utility for generating agro-ecological digital twins, using publicly available data (or at least readily collectible data).
For my purposes, a digital twin is essentially a critical component of a VEIL, which is a key component of my vision for Eden Falls.
You can play with a toy version of Mazzap right here on my site--just click the link in this sentence!
What's the point?
Obviously you can pretty readily visualize lidar point clouds with tools like
potree or cloud compare. You can view
SSURGO data in QGIS (and indeed any geospatial data you can imagine). So in a sense there's nothing "new" here.
But in another sense there's a lot new going on:
- Tools like potree and cloud compare just visualize point clouds. There are no meshes there. Meshes are important for things like collisions but most importantly, segmentation. If you wanted to use a lidar capture as an RL enviornment for a robot, for example, collisions and segmentation comes in handy. Also, if you want to build a tool to model tree/plant growth and how they respond to changes in the environment over time, segmenting the individual trees is critical.
- Visualizing all of the relevant data in 3D in one canonical model is kind of new! At least, in the open source world. Obviously I'm cheating a bit because I get to define what the relevant data is but it is relevant to me and I'm sure it will be relevant to at least one other person.
- Having all of the relevant data in one canonical model let's you build semantic representations of your data, which will be critical for GAIA and therefore critical for the VEIL.
At any rate, we're still early, but you'll see. This is gonna be really cool when I'm done.