Skip to Content

AI Smoke and Fire Detection Model - ITRI Edition

Site Edition
Inquiries welcome. Please contact sales for a complete solution and quotation.

Terms and Conditions
Stock and lead time vary by product; the formal quotation prevails

AI Smoke and Fire Detection Model · ITRI Edition

The ITRI edition smoke and fire detection model analyzes on-site camera video continuously, recognizes smoke and flames and marks where they occur; after confirmation it pushes LINE and Microsoft Teams alerts in real time and can trigger on-site strobe warning lights, so site managers, safety officers and engineering supervisors know of an anomaly in the earliest stage of a fire, buying time for evacuation and initial firefighting. The model has been tested with simulated video and is tuned continuously based on test feedback.

AI smoke and fire detection: marking smoke and flame positions

Real-life photo; detection boxes are illustrative.

Use case: smoke appears in a warehouse and staff evacuate on alarmIllustrative scene generated with AI; not footage of the actual product or a customer site.

Key Features

AI Video Analytics
2 classes
Detection Targets
Smoke, flame
24/7
24/7 Monitoring
Uninterrupted automatic detection
3 channels
Synchronized alerts
LINE / Teams / warning light
ONVIF
Existing Cameras
Compatible with mainstream IP cameras

Simulated Detection Test Video

Field test

4S360 smoke and fire detection model test (small fire version): taken from a public fire-safety awareness video, simulating an indoor fire scenario for detection; actual results vary with on-site lens, lighting and environment. Test results continue to feed back into model tuning. Watch on YouTube

Core Detection Features

Smoke detection

Recognizes smoke in the image and marks its position, warning early before a fire grows, unlike spot smoke detectors that react only once smoke drifts to them.

Flame detection

Marks where the flame occurs, helping managers quickly judge the ignition point and the situation on site.

Multi-channel real-time alerts

After an anomaly is detected, alerts are pushed to the LINE official account (with event snapshot), the Microsoft Teams channel and the on-site warning light at the same time.

Warning light linkage

Alerts also drive on-site sound and light warnings to remind workers to evacuate quickly or take response measures.

Automatic video evidence capture

When an event triggers, footage from before and after is captured automatically, giving a complete timeline for later investigation and occupational safety audit evidence.

Continuous tuning from field test feedback

Various fire scenarios are tested with simulated video, and unusual patterns found are folded into model tuning, gradually getting closer to on-site detection needs.

Detection and Alert Flow

1
Camera streams

Video from existing IP cameras (ONVIF) is fed continuously into the AI for analysis.

2
Smoke and fire recognition

The ITRI edition model recognizes smoke and flames and marks where they occur.

3
Alert events and video evidence

Generates an alert event and captures video from before and after the event automatically.

4
Instant notification and on-site linkage

Pushes LINE and Teams alerts and drives on-site warning lights at the same time, so site managers and safety officers know immediately.

Applications

Multi-industry use
Construction sites
Lumber yards / processing plants
Chemical plants
Warehouse and logistics centers
Petrochemical / energy facilities
Basements / parking lots
Electrical rooms / substations
Waste treatment sites

Technical Specifications

Core TechnologyDeep-learning video analytics, smoke and flame detection
Video SourceExisting IP cameras, compatible with the ONVIF standard
Monitoring Method24-hour continuous automatic detection
Alert ChannelsLINE official account (with event snapshot), Microsoft Teams, strobe warning light
Linkage ControlWarning lights driven automatically; third-party device I/O can be added
Video EvidenceVideo from before and after the event is captured automatically for later investigation, safety audits and evidence
DeploymentEdge computing, cloud server or hybrid architecture
System IntegrationSupports API integration with existing site management platforms and the 4S360 smart site platform

Want to know how AI smoke and fire detection is deployed on site?

We provide site surveys, camera placement advice and alert integration methods; a site assessment comes first.

Compatible Hardware