Zarr in the Browser — Fast, Flexible, and Surprisingly Powerful for Big Geo Data by Tobias Ferreira, Jul 1, 2025 medium.com/p/zarr-in-th...

Free multi-modal remote sensing model predictions for Europe

Copernicus Land Monitoring System (CLMS) High-Resolution Layers (HRL) Croplands (2017–2023) geoserver.vlcc.geoville.com/geoserver/HR...

Grasslands (2017–2021/3) geoserver.vlcc.geoville.com/geoserver/HR...

Data viewer wekeo.copernicus.eu

Thinkpad X1 Carbon Gen 8 touchpad right button gradually stopped working. Saved ~200 € by mapping a mostly unused nordic key to right-click

#Requires AutoHotkey v2.0
$SC01A::Click "right" ;Å = right-click
^SC01A::Send "å" ;Ctrl + Å = å
^+SC01A::Send "Å" ;Ctrl + Shift + Å = Å

AlphaEarth Foundations & the Satellite Embedding dataset

Presentation by Valerie Pasquarella at GDG AI for Science - Australia, 2025-08-06

www.youtube.com/watch?v=1oiy...

Or... Shrink each tile by ½ the minimum same-UTM-zone overlap of 9780m. Then split overlaps between tiles from different UTM zones at the mean central meridian of those UTM zones. (TODO: split remaining tiny same-UTM-zone overlaps, absorb tiny triangles in polar areas.)

Real Sentinel2 tiles, an even worse case.

Or maybe not, sometimes such splitting would create multiple regions from the same tile.

I think this could be improved on by choosing the tile the boundary of which is the farthest away from the point that we want. Testing the concept using two tiles in the same coordinate system. The split between these tiles would be at the zero contour.

From which Sentinel 2 tile to load data for a given location? Spherical Voronoi partition based on tile centroids. (Engel & Körner 2024)

github.com/meengel/Sent...

JAX implementation github.com/hamk-uas/lin...

My 1st ML paper! "Per-channel autoregressive linear prediction padding in tiled CNN processing of 2D spatial data" arxiv.org/abs/2502.12300

Our linear prediction padding method gives slightly better CNN super-resolution results than zero or replicate. It better extrapolates activations spatially.