Geospatial Imagery Annotation

Kili Technology's geospatial annotation platform lets GIS analysts and domain experts label satellite, aerial and drone imagery to build production-ready training data for Earth observation AI. Dedicated Geospatial projects ingest GeoTIFF-based files (.tif, .jp2, .ntf), uploaded directly or synced as Cloud Optimized GeoTIFFs from AWS S3, Azure Blob Storage, Google Cloud Storage or S3-compatible buckets. Dynamic tiling keeps multi-gigabyte images responsive (files under 5 GB recommended). All bands of multispectral and hyperspectral imagery are preserved. Annotators can build true- or false-colour composites and band-math indices such as NDVI and NDWI, rescale high-bit-depth imagery such as SAR, and stack several images of one area (for example SAR and electro-optical) as layers of a single asset. Every annotation carries geo-coordinates. Labellers can switch coordinate reference systems, find a location by latitude and longitude, copy GPS coordinates, measure distances in metres and feet, and add context from XYZ base maps and WMS/WMTS servers.

Labelling tools include bounding boxes, polygons, points, lines, semantic segmentation with SAM 2 assistance, classification and transcription. Consensus, honeypot and multi-step review workflows control quality, and a map view lets managers assign and prioritise assets by region. Labels can be imported from shapefiles, GeoJSON or Kili format and exported as GeoJSON or Kili format, in the web app or programmatically through a GraphQL API and Python SDK. Kili runs as SaaS (Google Cloud in Belgium or Azure East US), as a managed application in the customer's own Azure subscription, or on-premise, including air-gapped. It holds SOC 2 Type II, ISO 27001 and HIPAA certifications, and Kili also offers managed geospatial labelling services.

Enabled Intelligence, a US geospatial AI provider, chose Kili after evaluating more than 35 platforms. It reports 95%+ accuracy on training data and 99%+ on ground-truth datasets, labelling SAR, electro-optical and hyperspectral imagery for military aircraft detection, naval vessel identification, agricultural monitoring, disaster recovery and land-use classification.