Conservation

What is Conservation?

A camera-trap deployment produces tens of thousands of frames, most of them empty, and the ecological signal inside them is only as good as the review capacity available. Conservation identifies species with confidence and the frame attached, maintains a digital passport per animal so a history survives across encounters, and scores threats on a rubric tiered by evidence strength — so a threat asserted from a single blurred frame cannot outrank one built from repeated confirmed observation.

Problems it solves

• Identify species across large volumes of field imagery
• Keep a history for an individual animal across encounters
• Score threat with evidence strength made explicit
• Monitor habitat change over time
• Prioritise patrols and field effort
• Keep observation and inference visibly separate

Key capabilities

Species Identification

Identify species from imagery, with confidence and the frame that produced it.

Evidence Trail

Every figure carries the record it came from, so a reader can check the answer instead of trusting it.

Confidence Ceilings

Confidence is capped by the strength of the underlying basis, so a weak source cannot produce a strong-sounding claim.

Risk Scoring

Score a change, a lot or a habitat on a published rubric you can read component by component.

Change Detection

Compare observations over time and flag what changed, where and by how much.

Fact / Inference Separation

Keep what is known, what is inferred and what is alleged visibly distinct.

How it is built

Frame-linked identification

Every identification keeps the frame and a confidence value.

Animal Digital Passport

An individual's encounters, health and history stay together over time.

Evidence-tiered threat scoring

Weak evidence cannot produce a strong threat score.

How it works

How the work is structured, and what each step's output rests on.

1. 1

   Ingest field imagery

   Source

   Take camera-trap and survey imagery as captured.

2. 2

   Identify

   Observed

   Classify species with confidence and keep the frame.

3. 3

   Resolve individuals

   AI-inferred

   Attach encounters to an animal's passport where identifiable.

4. 4

   Assess habitat change

   Observed

   Compare against earlier observation of the same area.

5. 5

   Score threat

   AI-inferred

   Apply the evidence-tiered rubric, component by component.

6. 6

   Field verification

   Human-verified

   Record human confirmation distinctly from model output.

Example use cases

Each example has its own page you can read and share.

Ask Conservation

Illustrative examples. Results depend on your own data.

• “Identify the species in this set of trail-camera frames.”
• “Show this animal's encounter history.”
• “Where has habitat risk increased this month?”
• “What evidence supports this threat score?”

Security, governance and human control

Where the software deliberately limits itself.

Questions

How reliable is species identification?

Every identification carries a confidence value and the frame it came from, so a reviewer can confirm or overturn it. Confirmations are recorded as human-verified.

What is the Animal Digital Passport?

A per-individual record that keeps encounters, health observations and history together across deployments and time.