Picterra

Picterra, founded in 2017 and headquartered in Lausanne, Switzerland, provides a no-code/low-code geospatial AI SaaS platform that enables non-ML-expert users in geospatial organizations — GIS analysts, agronomists, environmental scientists, and field operations teams — to build, train, validate, and deploy custom machine learning models for automatic feature detection, change detection, and classification from satellite and aerial imagery without writing code.

Picterra's platform abstracts the ML workflow into a web-based interface where users draw training examples directly on imagery to teach a model what objects to detect (buildings, vehicles, crop types, tree species, mine tailings ponds, solar panels), then trigger automated model training, review detection results, and deploy models as a REST API or batch processing job against new imagery datasets. This approach dramatically lowers the barrier to using AI for geospatial analysis compared to building custom deep learning pipelines in PyTorch or TensorFlow.

The platform integrates with major commercial satellite data providers (Airbus, Maxar, Planet Labs) and drone/aerial imagery platforms, and supports standard GIS formats (GeoTIFF, GeoJSON, Shapefile) for interoperability with ArcGIS and QGIS workflows. Picterra's object detection capabilities have been applied to forestry inventory (tree canopy mapping), precision agriculture (crop disease detection, field boundary delineation), mining (tailings dam monitoring, stockpile volume estimation), and disaster response (building damage assessment).

Picterra raised approximately $10 million in venture funding from Swiss and European investors including Wingman Ventures, supporting R&D and sales expansion. The company competes with Descartes Labs, Scale AI, and Maxar's Radiant.earth platform in the geospatial ML tooling market, differentiating through its no-code user experience targeting domain experts rather than data scientists.