Mapflow turns recent satellite imagery into the three datasets RF planners work with: population grids, 3D maps of terrain, buildings and vegetation, and clutter packages ready for Atoll. Order a whole region, or update only what changed.
Talgar, Kazakhstan: 3,203 buildings and 14,240 trees and vegetation areas extracted by Mapflow. Terrain: Copernicus DEM GLO-30. Roads and river: © OpenStreetMap contributors. Coverage from a simplified propagation model, for illustration.
All three come out of the same 3D extraction, so clutter, buildings and population agree with each other and with the date of the imagery.
Population maps
Census or official population is spread across residential buildings by volume (footprint × height), then rolled up to an H3 hex grid of about 1.5 ha cells. The grid joins straight onto coverage predictions, traffic and site data.
Talgar, Kazakhstan
37,751 residents across 19,683 Mapflow buildings, on Copernicus terrain. Switch allocation models, tilt to 3D and hover any cell.
Estimate how many people each sector serves and add capacity before congestion shows up in your KPIs.
Rank candidate sites by the people inside their predicted footprint, not by area alone.
Report population coverage for license obligations and investment cases from the data your planners already use.
An H3 grid with residents and density per cell, per-building estimates, and two allocation models to compare.
Checked against the housing registry in Noginsk: building volume ranks apartment counts with a Spearman correlation of 0.80 across 476 apartment blocks.
3D maps
Every obstacle between the antenna and the subscriber, extracted by AI from one recent satellite image. No lidar flight or stereo survey to commission.
±3–4 mbuilding height RMSE in dense urban areas
≈1 min per km²of very-high-resolution imagery processed
Start from a terrain model of the area: Copernicus GLO-30 or the elevation data you already use.
Mapflow classifies land use and extracts every building and tree, with its height, from recent satellite imagery.
Load the layers into your planning tool and predict coverage on the city as it stands today.
BeforeAfter updateWhen new imagery shows a change, such as this construction site, Mapflow re-maps only that area and you recalculate the network.
Talgar, Kazakhstan. The site in step 4 is marked as active construction in OpenStreetMap; its new blocks are in the latest Mapflow buildings. Terrain: Copernicus DEM GLO-30. Coverage from a simplified propagation model, for illustration.
Atoll-ready datasets
We convert the 3D map into Atoll geographic data, mapped onto your clutter model and set in your project's coordinate system. Every delivery records the imagery dates and sources behind it.
Raster and vector formats Atoll reads, including GeoTIFF, BIL, Shapefile, GeoPackage and GeoJSON.
Example clutter classes
A complete dataset for a city, region or rollout corridor, built from the most recent imagery available. Mapflow searches basemaps and commercial providers for the freshest scenes over your area and alerts you when new ones arrive.
How Mapflow imagery search worksKeep your vendor dataset and refresh only what changed: new and removed buildings, grown vegetation and new roads, merged into your clutter. Days instead of a new purchase every one to three years.
Case study
We compared an existing Atoll dataset for 40 km² around Talgar with layers Mapflow extracted from recent satellite imagery, then converted the differences into Atoll-ready updates.
9,982
buildings missing from the existing dataset
97%
of them houses 3–10 m tall
12.7 km²
of vegetation mapped with heights, nearly a third of the area
A package per area, in your coordinate system and clutter model.
Review layers, run new areas and download results yourself.
Automate scheduled updates for the districts that keep growing.
On-premises deployment when your data can't leave your network.
Tell us the area and the datasets you need. We'll process a sample and show you the population grid, the 3D map and what changed against your current clutter.
