showcasing my project over the last years (again). The updated LC-Product works now pretty streamlined and it makes pretty interesting (and surprisingly good to compress) maps.
Stack is a collection of adjustable U-Nets that make predictions that are merged together to derive the final map.
Interesting read - there's a lot of progress in the open map styling scene and OSM maps with good data can make amazing maps and it has many different features [0].
It's also interesting to make maps directly from raw data (in this case that would be aerial imagery with mostly elevation data). E.g. if you find good training data you can easily make "land-cover" maps and although my example (of Austria) [1] does not have annotations (addresses, etc.), it's still nice to look at and has the benefit of being extremely compressible (compared to aerial raw imagery).
Senior Software Engineer specialized in Django, Pipeline-Development, GIS and applied machine learning in the context of earth observation, sensor mapping in general and remote sensing in particular.
Interesting. Came to similar conclusions when analyzing my (httpd) access logs for https://turmfalke.httpd.app/demo.html ... but so far nothing really out of the ordinary.
It's a nice start but what we also need would be an "OpenRasterMap" project - a project where we collect raster data (more and more data gets available) and store the entire planet (RGBN aerial imagery) in let's say 1m/Pixel.
It might be feasible to store entire planet's raster tiles in 10 years in a (beefy) desktop machine.
Sentinel-2 provides worldwide coverage in 10m resolution and is open data. It would be challenging to reach worldwide 1m coverage, even for the big players.
You may also want to have a look at OpenTopography, following a similar principle but for elevation data.
Yes, I am aware of Sentinel-2 10m data (RGBN), there are even commercial offerings. Actually the storage requirements are not that big for 10m, like ~4TB for the entire (non-water) world. Napkin math:
- 15000 Sentinel-2 scenes over land
- 10980^2 pixels per scene
- 50% lossless compression per scene
- 4 channels (RGB and infrared)
This results in 15000 * 10980^2 * 4 * 0.5 / 1024*4 ~ 3.3TB. This matches commercial offerings like in https://cloudless.eox.at/#data. Of course we have to add lower-scale imagery that is often used to provide XYZ tiling that would further increase the storage requirements.
Thanks for the link to OpenTopography, it's an interesting project and elevation data is extremely valuable.
That's actually a nice idea. I guess we will see more software in the source code dependencies analysis space. There's so much code and often it's nice to have some kind of metrics (LOC, PageRank, ...) to get a grasp of what's important in a codebase.
Wow never heard of GBIF, but you seem to offer useful data. Am I correct that you are in essence providing georeferenced animal/plant population data across the globe in mostly CC license?
showcasing my project over the last years (again). The updated LC-Product works now pretty streamlined and it makes pretty interesting (and surprisingly good to compress) maps.
Stack is a collection of adjustable U-Nets that make predictions that are merged together to derive the final map.
Happy to answer questions.
Happy holidays and happy hacking.