Skip to main content

Sensor Integration

Bring an IoT device — soil moisture probe, weather station, gateway, edge camera — onto the platform via the OGC SensorThings API.

What you need​

  • A working device that emits measurements.
  • A service account JWT (see Concepts · IAM).
  • An MQTT client for live streams; HTTP for batch / manual uploads.

The full lifecycle​

  1. Register a Thing. Describe the physical device with a name and a Location (GeoJSON Point).
  2. Add Sensors. Each Sensor declares its model and metadata.
  3. Define Datastreams. A Datastream binds a Sensor to an ObservedProperty and a unit of measurement. Optional but recommended: link the Datastream to a Region in the Farm API so observations are spatially grounded.
  4. Send Observations.
    • MQTT for high-frequency streams.
    • REST for batch updates and manual edits.
  5. Query. Use OGC-STA query syntax ($filter, $orderby, $top, $expand) for time-series analysis.

Payload conventions​

  • Use JSON-LD with an explicit @context for any metadata that should be cross-platform-readable.
  • Use AgroVoc URIs for ObservedProperty whenever a vocabulary entry exists.

Batch vs. streaming​

  • Streaming (MQTT) — sub-minute cadence; appropriate for live dashboards.
  • Batch (REST) — bulk uploads or manual edits; cheaper for low-cadence loggers.

Edge deployment​

The platform supports an edge agent pattern: a small process on a gateway that buffers observations during connectivity gaps and replays them once a link is available. Buffer is durable and uses the same JWT.

Reference implementations​

  • NaLamKI consortium — Raspberry-Pi-based soil-moisture sensors
  • BharatRohan drone telemetry (Telangana)

See it in action​