Edge Compute · DataBlade

Turn any Hesai LiDAR into real-time decisions.

The Metrolla DataBlade processes 3D perception on-site, so your site reacts instantly — without cloud bills, privacy risks, or months of integration. People counted. Vehicles tracked. Gates opened. Alerts fired. All in under 80 ms.

Metrolla DataBlade ruggedized edge compute unit running 3D LiDAR perception on-site

Why edge perception?

Cloud pipelines turn sensor data into reports. Edge perception turns it into action.

React in milliseconds

Under 80 ms from LiDAR frame to event. Open gates, trigger alerts, or reroute traffic while it is still happening.

Protect privacy by design

No cameras, no faces, no PII. Only anonymous 3D objects leave the device, so you can deploy where video cannot go.

Kill the cloud bill

Perception runs on the blade. Stop paying to stream point clouds to a datacenter just to get a person count back.

Count once, everywhere

Blades mesh so the same person or vehicle keeps one ID across sensors. Accurate journeys, no double counting.

What you get with DataBlade

Live decisions instead of dashboards. Coverage instead of cameras. Outcomes your ops team can actually use.

Act on events in under 80 ms

Fast enough to trigger alerts, open gates, or reroute traffic while it's still happening — not minutes later in a report.

No cloud bill, no GPU rack

Perception runs on the blade itself. You stop paying to stream point clouds to a datacenter just to get a person count back.

Deployable where cameras aren't allowed

Only anonymous 3D objects leave the device — no faces, no video. Get coverage in privacy-sensitive spaces without a legal review cycle.

One site, one identity per person

Blades mesh so a person or vehicle keeps the same ID across sensors. You get accurate journeys and zone counts instead of double-counted blobs.

Install once, leave it outside

Weather-rated for pole, ceiling, or trailer mounts. Survives the seasons so your team isn't babysitting hardware.

Plugs into the stack you already run

Structured events over REST, MQTT, or webhooks — drop them into your VMS, BI tool, or control system without a custom integration project.

What ships in the box

Form factor
Ruggedized edge compute, pole / ceiling / trailer mount
Perception
3D detection, classification, multi-object tracking with persistent IDs
Latency
<80 ms LiDAR frame → tracked event
Networking
Wireless mesh + Ethernet, streams REST / MQTT / webhooks
Environment
Weather-rated for outdoor deployment
Compatibility
Sensor-agnostic — pairs with all Hesai LiDAR models Metrolla distributes
  • Zero-touch onboarding
  • OTA firmware & model updates
  • Fleet health monitoring
  • Works with any Hesai LiDAR we sell
FAQ

Evaluating the DataBlade

The questions we get most from teams sizing an edge perception deployment.

What does the Metrolla DataBlade actually provide?

A ruggedized edge compute unit that runs Metrolla's 3D LiDAR perception software on-site. Point any Hesai sensor at it and you get real-time detection, classification, and multi-object tracking with persistent IDs — streamed out as structured events, not raw point clouds.

How is this different from just plugging a LiDAR into a server?

The DataBlade ships pre-tuned for Hesai sensors, handles calibration, and runs perception locally. No GPU server rack, no cloud pipeline, no ML team required. Deploy it on a pole, ceiling, or trailer and it starts producing tracked-object events immediately.

What latency and throughput should we expect?

End-to-end latency is under 80 ms from LiDAR frame to tracked event on the wire. Each DataBlade handles one sensor's full frame rate; multi-sensor sites scale by adding blades that mesh together.

How does the wireless mesh work across multiple units?

DataBlades form a self-organizing mesh so objects keep the same ID as they move between sensor coverage zones. That means accurate zone counts, hand-offs, and journey analytics across a site without stitching feeds together downstream.

How do we get data out of the DataBlade?

Structured events stream over REST, MQTT, or webhooks. You choose whether to send counts, zone events, or full tracked-object streams into your dashboard, VMS, BI stack, or control system.

Is video or PII ever transmitted?

No. Perception runs on-device and only anonymous 3D objects and events leave the unit. No faces, no video frames, no re-identifiable data — which makes DataBlade deployable in privacy-sensitive environments.

Which Hesai LiDAR models pair with the DataBlade?

Every Hesai sensor Metrolla distributes is supported — JT, XT, QT, AT, Pandar, FT, OT, and ETX series. We help you match sensor coverage and DataBlade count to the site during the quote.

How do we evaluate the DataBlade for a project?

Request a quote with your site layout and target use case (people counting, vehicle tracking, perimeter, etc.). We spec sensor placement plus the number of DataBlades needed, and can arrange a pilot deployment before you commit to a full rollout.

Deploy perception without the cloud bill

Get a DataBlade + Hesai sensor bundle spec'd for your site. One quote covers hardware, perception software, and integration support.