Checklist
AI camera monitoring checklist for fluid processes.
Use this checklist to screen wastewater, foam, scum, oil-film candidates, river CCTV, and other fluid-process scenes before designing a PoC. The goal is to confirm whether existing footage can support practical review, not to promise universal detection.
Start with business fit, then validate footage after NDA when needed.
Screening questions
A practical pre-PoC checklist
1. Can a human see it?
If an experienced operator cannot see the target cue in footage, AI validation will be difficult.
2. Is the camera fixed?
A stable view is easier to validate than a moving or heavily vibrating camera.
3. Is the target large enough?
The target anomaly must occupy enough of the frame to support review.
4. Are lighting issues manageable?
Glare, night conditions, rain, fog, shadows, and lens contamination should be checked.
5. Are normal and abnormal clips available?
After NDA, short examples help define what should count as a review candidate.
6. What action follows?
The PoC should connect visual detection candidates to a human review, patrol, notification, or record workflow.
Scope control
Start with one asset, one anomaly, one decision.
Broad PoCs tend to fail because they mix multiple cameras, anomalies, and workflows. A narrow PoC makes it easier to judge whether camera AI creates operational value.
Want to screen a camera scene?
A first discussion can happen without footage. We clarify the target facility, visible anomaly, camera availability, and operational decision.