Baseera · Pitch Deck

Defence & Security · EO/IR & GEOSPATIAL

Baseeraبصيرة

Object and change detection across satellite, drone, and ground-sensor imagery — at the edge.

neuriya.ai — Sovereign, Arabic-first enterprise AI

Problem

Satellite, drone, and ground-sensor imagery is reviewed frame by frame, by eye

Analysts scan sequential satellite, drone, and ground-sensor imagery manually to spot structural, vehicular, or terrain changes, a process that doesn't scale with sensor volume.
Detection consistency depends on which analyst is on shift and how many frames they've already reviewed that day, not a calibrated standard.
Forward bases and remote sensor sites often have degraded back-haul connectivity, so imagery queues for review instead of being scored where it's collected.
Every frame gets equal attention in manual review, because there's no confidence score to tell an analyst which detections actually need eyes first.

Solution

Object and change detection across satellite, drone, and ground-sensor imagery — at the edge

Sequential imagery is compared automatically to surface structural, vehicular, and terrain changes an analyst would otherwise scan for manually.

One pipeline ingests satellite optical, drone thermal, and ground-sensor feeds and scores detections consistently across all three, regardless of who's on shift.

Detection models run on hardware at the sensor or forward base, so results are available even when back-haul connectivity is degraded.

Every detection carries a calibrated confidence score so operators can triage instead of reviewing every frame equally, with sign-off required before anything escalates to command.

Product

Baseera in action

How it works

From first touch to outcome

Connect

Satellite, drone, and ground-sensor feeds are connected once and normalized into a common detection pipeline.

Detect

Object and change-detection models score every frame, at the edge where connectivity allows it.

Review

Operators review flagged detections against confidence scores and sequential imagery before any escalation.

Escalate

Confirmed detections route automatically to the relevant command center, fully logged.

Market

Sizing the opportunity

SOM$16M

TAM$1.9B

SAM$230M

TAM — $1.9B. GCC EO/IR & geospatial-intelligence analytics market
SAM — $230M. Saudi defence & border-security computer-vision analytics spend
SOM — $16M. 3-year obtainable share via design-partner unit and edge deployments

Vision 2030's investment in satellite and drone sensing for border security and national infrastructure protection is increasing the volume of EO/IR imagery collected faster than manual analyst capacity can review it, making edge-deployable detection an easy operational case.

Business model

How Baseera makes money

Per-sensor-feed license

Priced by the number of satellite, drone, and ground-sensor feeds connected to the detection pipeline.

Edge-deployment subscription

Recurring fee for the deployed edge inference units and their ongoing model updates.

Implementation & model-calibration fee

One-time fee to calibrate detection and change-detection models to sensor types and terrain.

Revenue streams

Where the revenue actually comes from

Implementation & model-calibration fee

One-time

Charged per site or unit to calibrate detection and change-detection models to sensor types and terrain.

Per-sensor-feed analytics license

Recurring

The primary line: priced by the number of satellite, drone, and ground-sensor feeds under active monitoring.

Edge-hardware deployment subscription

Usage-based

Metered by the number of edge inference units deployed at sensor sites or forward bases.

Change-detection archive licensing

Licensing

Planned licensing of historical change-detection archives back to command for long-range planning, once sufficient sequential imagery has accumulated — not yet active.

Competition

Where we win

Global EO/geospatial-analytics vendors (Maxar, BlackSky analytics arms)

Established satellite and geospatial-intelligence analytics providers serving allied defence customers.

Priced and hosted for foreign cloud delivery, with no edge-deployable inference for connectivity-degraded forward sites.

Generic computer-vision platforms adapted from commercial security/retail use

Object-detection vendors repurposing commercial CV models for defence imagery.

No calibration for EO/IR sensor types or change-detection over sequential defence imagery, requiring custom model work per deployment.

Status quo: manual frame-by-frame imagery review

Most units today review satellite, drone, and ground-sensor imagery by eye, frame by frame.

No confidence-scored triage, no edge inference, and detection consistency depends entirely on which analyst is on shift.

Our edge

One pipeline normalizes satellite, drone, and ground-sensor feeds into a consistent detection standard, instead of a separate tool per sensor type.
Edge-deployable inference keeps detection running at forward bases and remote sites even when back-haul connectivity is degraded.
Operator sign-off is required before any detection escalates to command — the model flags, the person decides, with a full audit trail retained.
Founding team combines Saudi defence sensing experience with computer-vision engineering.

Traction

Early validation, built for pilot readiness

Full geospatial detection experience designed and built end-to-end: sensing overview, detections log, and detection case review.
Architecture pre-aligned to NDMO Classification, NCA ECC-2, ISO 27001, and edge-deployable inference.
Actively selecting design-partner units for a first edge deployment pilot cohort.

NDMO ClassificationNCA ECC-2ISO 27001Edge Deployable

Go-to-market

Land on a handful of sensors, expand to full coverage

1Land with 1-2 design-partner units, starting on a limited set of satellite, drone, or ground-sensor feeds.
2Prove detection accuracy and change-detection reliability over a pilot period against the unit's existing manual review process.
3Expand to full sensor coverage and additional edge deployment sites through the same unit's sensing infrastructure.

Team