ConductVision Enterprise
ConductVision marks visible flame and smoke in camera views of yards, sites and open land, and alerts staff with a still of the first frame.
ConductVision tracking on every frame of this clip.
Fire & Smoke Detection uses computer vision to identify visible flames and smoke plumes in video feeds, providing an early warning layer that complements existing fire alarm systems. The system is designed to flag the first visible sign of flame or smoke for human review. The models distinguish between smoke from fire, steam, dust, and vehicle exhaust to minimize false alarms. This capability is particularly effective in warehouses, forests, construction sites, and other environments where traditional sensor placement is impractical.
Outcomes
Planned burns and hot work run under permits with set hours and a set area.
When smoke drifts past the area or the hours, crews hear about it from a phone call.
After welding or cutting, someone has to watch the area for a set time in case something smolders.
When that watch is cut short, a smoldering fire can go unseen.
Storage yards, mulch piles and scrap heaps go unwatched overnight and on weekends.
A fire that starts after hours can burn until a neighbor or a passer-by sees it.
Steam, dust and low sun can look like smoke from some camera angles.
Staff who clear the same false alert every morning start to doubt the real ones.
What the model sees
ConductVision detections on this photo.
ConductVision detections on this photo. Photo: Slknight, CC BY-SA 4.0, cropped.
Endpoints
Each endpoint is defined before the pilot starts, with what is counted, its unit and how the results are broken down.
| Endpoint | How it is counted | Unit | Broken down by |
|---|---|---|---|
| PrimaryTime to alert | Seconds from the first frame with visible flame or smoke to the alert reaching staff. | Unit:seconds (median) | By:Camera, flame or smoke |
| SecondarySmoke events | Separate episodes of visible smoke, each counted once from its first to its last frame. | Unit:events per camera-week | By:Camera, hour |
| Fire-watch coverage | Share of closed hot-work permits whose full watch period was in camera view with no gap in the footage. | Unit:% | By:Area, week |
| After-hours detections | Flame and smoke detections between closing and opening time, counted once per event. | Unit:events per week | By:Site, class |
Each pilot measures ConductVision on your own footage against a sample checked by hand, before anyone relies on its numbers.
Alert precision
Share of flame and smoke alerts a reviewer confirms in a hand-checked sample of alert stills.
% · by camera, class
Missed events
Share of flame or smoke episodes a reviewer finds in sampled footage that ConductVision did not flag.
% · by camera, distance band
False alerts by cause
Alerts a reviewer rejects, counted by cause such as steam, dust, fog or glare, per camera per day.
alerts per camera-day · by camera, cause
How it works
Connect
Use existing surveillance cameras or add purpose-mounted cameras covering high-risk areas. PTZ cameras can be configured to sweep zones on a schedule, maximizing coverage area per camera.
Analyze
Dual-path models analyze both texture/color signatures of flames and the motion characteristics of smoke plumes. Temporal analysis across multiple frames reduces false triggers from transient events like reflections or fog.
Act
Confirmed detections trigger multi-channel alerts, sirens, SMS, email, and integration with building management systems. Each alert includes a snapshot, camera location, and confidence score for rapid triage.
What it detects
Fire
Visible flame in view, boxed in each frame
Smoke
Plumes and haze in view, boxed in each frame where they appear
Event start and end
The first and last frame of each fire or smoke episode, per camera
Smoke spread
Growth of the boxed smoke area from one frame to the next
Features
Use cases
High-rack warehouses have ceilings too tall for conventional detectors to respond quickly. Camera-based detection at rack level can identify fire in seconds, before smoke reaches the ceiling.
logisticsmanufacturingretail
Forestry agencies need early detection of fires in remote areas. Tower-mounted cameras with long-range lenses can scan the horizon continuously, detecting smoke columns before they become visible to ground observers.
governmentconservationutilities
Hot work, welding, and exposed materials create fire risk on construction sites. Visual detection can provide alerts when sparks or flames appear outside designated hot-work zones.
construction
Electrical faults in solar panels can cause fires across large, unmanned sites. Camera drones or fixed cameras can detect flames and smoke across the array, triggering shutdown protocols.
energyutilities
Chemical plants and refineries contain high-risk areas where fires can escalate rapidly. Video-based detection can provide a redundant layer alongside gas detectors and heat sensors.
manufacturingenergychemical
Heritage structures often cannot accommodate modern fire suppression systems without damage. Non-invasive camera-based detection can provide early warning without physical modification to the building.
governmenthospitalitycultural
The challenge
Traditional smoke detectors rely on particulate reaching the sensor, which can take minutes in large spaces like warehouses, atriums, or outdoor areas.
How ConductVision helps
ConductVision can detect visible smoke and flame in camera feeds, even in large open environments. Alerts can include the precise location and a video snapshot, enabling fire response teams to mobilize before conventional alarms activate.
manufacturinglogisticsconstruction
The challenge
During evacuations, incident commanders lack real-time data on which areas have been cleared, where people are stuck, and whether all exits are functional.
How ConductVision helps
The platform can provide a live occupancy overlay during evacuations, showing remaining population by zone, exit flow rates, and blockage points. This situational awareness can enable commanders to direct resources to areas that need assistance most urgently.
smart cityhealthcaremanufacturing
The challenge
Dangerous crowd densities develop rapidly at events, transit stations, and venues, and by the time security recognizes the danger, injuries may already be occurring.
How ConductVision helps
ConductVision can monitor crowd density in real time and trigger escalating alerts as density approaches dangerous thresholds. Early warnings can give security teams time to open additional exits, halt inflow, or make public address announcements before conditions become critical.
smart cityentertainmentinfrastructure
The challenge
Individuals who collapse or experience medical emergencies in public spaces may go unnoticed for critical minutes, especially in less-trafficked areas.
How ConductVision helps
The platform can detect people who collapse, fall, or remain motionless in unusual positions, triggering alerts to security and medical teams. Faster detection means faster response, which is critical for cardiac events, seizures, and other time-sensitive conditions.
healthcaresmart cityentertainment
Technical overview
Accuracy and speed are measured on your footage and cameras during the pilot, and reported with the conditions they were measured under.
Industries
Watches crews, trucks and equipment at plants, well pads and tank farms.
Tracks logging trucks, machines and wildlife on forest roads, and flags smoke.
Related
Finds people, vehicles and animals in thermal camera frames, day or night.
Watches water, wildlife, vessels and smoke at rivers, wetlands and shorelines.
Spots and tracks small drones in the sky over sites and perimeters.
Pilot
A pilot takes one prioritized outcome, runs on your cameras, reports the endpoints above and ends with a results review and a decision.
