OrionAI reads satellite imagery, terrain and water on your map

OrionAI 2.10.0Released 24 October 2026

OrionAI in Map Studio now works with grids of values as well as features. It makes cloud-free Sentinel-2 and Landsat images, calculates about 280 spectral indices such as NDVI, maps change between dates, classifies land cover, reads slope and terrain, and traces streams and watersheds.

Read the announcement: OrionAI now reads the land itself

Changes since 2.9.0.

#New

  • Satellite imagery. Ask for vegetation, water or built-up land anywhere and OrionAI makes a cloud-free image from Sentinel-2 (10 m, from 2017) or Landsat (30 m, from 1984) for your area and dates, then calculates the index you need. About 280 spectral indices are available, from NDVI and NDWI to NDBI and burn ratios.
  • Change over time. Compare an index between two dates (for example, vegetation change in the Mau Forest between 2019 and 2024), track it month by month, or compare land cover between years.
  • Land cover. Classify land cover around a place from satellite imagery, or summarise the 10 m global land cover map by area.
  • Terrain. Slope, aspect, hillshade, ruggedness, contours, what can be seen from a point and distance to features, from the Copernicus 30 m elevation model.
  • Water. Where water flows, the stream network with stream order, the watershed that drains to a dam, town or gauge you name, wetness and height above the nearest drainage line, a first screen for flood exposure.
  • Values by area. Population, elevation, rainfall or land cover summarised for each area, or read at each point, so you can map people per county or the share of each district under cropland.
  • Surfaces from points. Interpolate readings such as rain gauges into a surface, and turn surfaces into contours, shapes or points.
  • Image results on the map. Grids of values are drawn on your map as image layers. Only you can see them until you publish.

#Improved

  • Long steps keep going. Imagery and hydrology steps on large areas can take several minutes; OrionAI waits for them and shows progress instead of giving up.
  • Places by name. OrionAI finds a named dam, town or gauge and uses it, for example as the outlet of a watershed.
  • Shapes cleaned up. Simplify shapes, repair invalid ones, close gaps and overlaps between neighbouring areas, and check a layer for errors.

#Fixed

  • Accurate Sentinel-2 values. Reflectance from recent Sentinel-2 scenes is no longer corrected twice, so NDVI and other indices are not inflated.