Drone-mounted view over snowy terrain, with a bounding box marking the vehicleVehicle 75%
scaleout
Tactical CV Network

From the model
to the mission.

Operational computer vision for cameras, vehicles, and drones. Tactical CV Network combines application models, tools, and integrations, with a local learning loop to improve models as conditions change.

Discuss an evaluation Explore the workflow
The operational challenge

Conditions change.
The operational task remains.

Counter-UAS, ISR, and range surveillance place different demands on detection. The task, sensor feeds, target hardware, and operating environment define what a useful model needs to do.

Different observations

Terrain, weather, viewpoints, and sensor inputs can differ from the examples used to develop a model.

Real device constraints

Compute, power, memory, and available connectivity affect how models are operated and updated.

An operational workflow

Detections need a place in the team’s existing systems, with clear responsibilities for reviewing model changes.

Tactical CV Network / What’s included

A workspace for operation
and improvement.

Application models, operational tools, and deployment integrations brought together for your computer vision task, built on Scaleout Edge.

Illustrative Scaleout Edge Ground Node workspace showing a drone view, vehicle detection, and tools for operation and model improvement.
Scaleout Edge
StreamsDrone Station
Scaleout EdgeLabel Studio▸ Training●   Connected
Vehicle 75%
Illustrative interface. Explore the underlying Scaleout Edge platform.

Application models and workflows

Models, class sets, and workflows selected for the operational task, such as drone detection.

Tools for operation and improvement

Connect field observations, annotation, retraining, evaluation, and human approval in the local workflow.

Deployment and integration

Scaleout engineers help connect devices, sensor feeds, and operational systems. Engineering support and knowledge transfer help your team or chosen partner establish the workflow, within an agreed scope.

Adaptive AI / The workflow in practice

From useful frames
to an approved model update.

Run models on field devices, work with observations at the local site, and review the next model before it returns to operation.

Observe

Select useful frames.

AI CompanionGround Node

The Edge AI Companion runs models on devices and retains observations. At the Vision Ground Node, review useful frames and add or correct their labels.

Improve

Compare the candidate.

Ground Node

Use the Vision Ground Node to retrain a candidate on local data. Compare its detections and evaluate it against relevant labelled examples.

Deploy

Approve the next version.

Scaleout EdgeAI Companion

Your team reviews the results and approves a model version. Scaleout Edge distributes it to selected devices, including Companions, when connectivity allows.

New observations inform the next improvement
Where the software runs

On the device. At the site.
Across your network.

Tactical CV Network brings connected deployment roles together on the Scaleout Edge platform, from onboard detection to local model improvement and coordination across sites.

Drone operating above snowy terrain
On the device

Edge AI Companion

Device-side deployment bundle

Runs application models on the device’s compute. Retains observations and telemetry during disconnection, then synchronises with the Ground Node when links return.

Onboard or device-side compute
Example rugged computing hardware for local-site operation
At your site

Vision Ground Node

Ground-side deployment bundle

Provides the local workspace for reviewing and annotating observations, training candidates, evaluating results, and approving updates for field devices.

Local computing infrastructure
Scaleout Edge
Across participating sites

Shared coordination

Scaleout Edge platform role

Coordinates federated learning and model distribution across sites. Sites exchange model updates while operational footage stays local.

Infrastructure selected for the deployment

Images show example deployment hardware. Hardware configuration and supply, supported functions, and integrations are agreed for your deployment.

Modules within the deployment

VisionKit supports the computer vision workflow. DroneKit provides onboard autonomy and surveillance runtime functions for drone deployments. Ground Nodes and Companions combine the relevant software with hardware configurations suited to their deployment role.

Connected to your operational system

Sensor feeds, models, target compute, and operator interfaces are scoped around the application. This can include TAK/C2 integration where required.

Define an evaluation

Start with your task.
Agree what useful looks like.

Bring your operational task, devices, and field conditions. Work with Scaleout engineers to define a focused evaluation around your deployment.

Discuss an evaluation
  1. 01

    Define the task

    Agree the detection task, representative data, target hardware, success criteria, and responsibilities.

  2. 02

    Evaluate the workflow

    Test detection and a local model update, including review, approval, and operation with limited connectivity.

  3. 03

    Plan deployment

    Use the results to plan field validation, a pilot, or broader deployment, with integration, support, and training agreed for the next stage.

A shared delivery. Scaleout provides software and engineering support for the agreed workflow. Your team brings domain expertise, data access, equipment, and operating requirements; chosen partners can support hardware and integration.

Your team stays in control. You decide which model updates are approved and how they are used operationally. Data access, trained-model rights, and ongoing responsibilities are agreed for your deployment.