📄️ Detections, events and observations
These three words appear all over the app and they mean different things.
📄️ How a project is organised
A project stores your data in a hierarchy. Two of the words in it are new to most people, site and deployment, and those two are what the map, the trap nights and the rates are built on.
📄️ How trap nights are counted
A trap night is one camera working for one day. It is how you turn raw counts into a rate you can compare between sites.
📄️ Confidence and verification
The AI gives two confidence scores, both between 0 and 1. The detector says how sure it is that the box holds an animal. The species model then says how sure it is about the species. You can see both on a marked-up image: this box is 90% animal and 98% coyote.
📄️ How labels get cleaned up
The label you see is not always what the model said. Between the model and your screen, AddaxAI cleans up the results in three steps. This page explains what they do, so a label that surprises you still makes sense.
📄️ Dates, times and timezones
Camera trap analysis lives or dies on time. This page explains where AddaxAI gets the time, and what it does when the time is missing.
📄️ Species names and taxonomy
How AddaxAI names the animals it finds, how those names are organised, and how to add one it does not know.
📄️ The charts explained
A project turns your results into charts. The Dashboard gives you the overview, and the Insights pages go deeper, each with its own filters.