Skip to content
BytezTech
  • Deployed on NVIDIA Jetson edge devices
  • Multiple live CCTV feeds
  • 150+ projects since 2016

AI surveillance · Smoking detection

Smoking Detection AI for CCTV Cameras: Real-Time Alerts for No-Smoking Zones

BytezTech builds smoking detection AI that runs on the CCTV cameras you already have. A custom YOLOv8 model, trained on your footage, recognises smoking behaviour in restricted zones, processes several camera feeds at once on an NVIDIA Jetson device on site, and sends an alert with a snapshot to your security team. Pilot on one or two cameras in 2-4 weeks.

150+
applications delivered
8
countries served
44
client reviews
3
AI products in production

The problem we solve

No-smoking signs do not watch the cameras

A lit cigarette near a fuel pump, a paint shop or a stairwell is a fire risk the moment it happens, yet most sites only find out from a complaint or a report after the shift. Guards cannot watch forty screens. Smoking detection AI watches the no-smoking zones on your existing CCTV and alerts the right person within seconds, with a snapshot to act on.

Where is smoking detection used?

  • Fuel stations and chemical sites

    A lit cigarette near pumps or storage is an immediate alert, not a report after the shift.

  • Factories and warehouses

    No-smoking zones near stock, paint shops and packaging lines are monitored on existing cameras.

  • Hospitals and campuses

    Entrances, stairwells and smoke-free grounds are checked around the clock.

  • Hotels and residential towers

    Smoking in lifts, corridors and fire exits is flagged with the camera and time.

  • Public transport and depots

    Platforms, bus bays and workshops with fuel are covered from the control room.

  • Hot-work compliance

    The same model can be extended to open flames and welding outside permitted areas.

How it works

How smoking detection works on CCTV

A YOLOv8 model trained on your footage runs on an NVIDIA Jetson next to the NVR and raises alerts only inside the zones you mark.

Camera to alert in under a second, on site

  • Zone rules per camera

    Only the areas marked no-smoking raise alerts; smoking areas are ignored.

  • Snapshot, camera, time

    Each event goes to email, WhatsApp, Telegram, webhook or your VMS.

  • Confirm or dismiss

    A review screen feeds confirmed events back into the model to cut false alerts.

  • Same device, more rules

    PPE, hot-work, intrusion and ANPR can run on the same Jetson.

Deliverables

What you get

Everything in the list ships with the project. No add-ons, no surprises after the pilot.

  1. 01A smoking detection model trained on footage from your own cameras, not a generic demo
  2. 02Processing on an NVIDIA Jetson device on site, so video never leaves your premises
  3. 03Alerts with a snapshot, camera name and timestamp to email, WhatsApp, Telegram or your VMS
  4. 04Zone rules per camera: only the areas you mark as no-smoking raise alerts
  5. 05A review screen to confirm or dismiss each event and improve the model over time
  6. 06Daily and weekly reports by site, zone and shift
  7. 07Works alongside PPE, intrusion and ANPR detection on the same device

Why it works in the real world

Built for live sites, not lab demos

  1. 01

    Trained on your cameras

    The model learns your angles, lighting and uniforms, so a hand near a face or steam from a vent does not become an alert.

  2. 02

    Runs on site

    Inference on NVIDIA Jetson at the edge: alerts in seconds, several camera feeds per device, and no footage leaving the premises.

  3. 03

    Reports that change behaviour

    Daily and weekly reports by site, zone and shift show where signage, training or controls are needed.

Our development process

How does a smoking detection project work?

  1. Step 01

    Camera survey

    We review the angle, lighting and resolution of each camera and pick the zones to cover.

  2. Step 02

    Pilot in 2-4 weeks

    A model is trained on your footage and run on one or two cameras with live alerts.

  3. Step 03

    Rollout

    More cameras are added per Jetson device and alerts are wired into your existing tools.

  4. Step 04

    Tuning and support

    Events your team confirms or dismisses feed back into the model to reduce false alerts.

Our expertise

Our technology stack

YOLOYOLOv8PythonPythonOpenCVOpenCVNVIDIANVIDIA JetsonNVIDIADeepStreamEdge AI

Technical details

Input
Existing IP or analogue CCTV via RTSP or your NVR
Processing
On site, NVIDIA Jetson edge device; cloud optional
Detections
Smoking behaviour; optional hot-work, PPE and restricted-zone rules
Alerts
Email, WhatsApp, Telegram, webhook, VMS or dashboard
Pilot
2-4 weeks on one or two cameras

Where it is used

Smoking detection by site type

  • Fuel stations and chemical sites

    Pumps, tank farms and storage: an immediate alert instead of a report after the shift.

  • Factories and warehouses →

    Paint shops, packaging lines and stock areas monitored with the PPE rules on the same cameras.

  • Hospitals and campuses →

    Entrances, stairwells and smoke-free grounds checked around the clock.

  • Hotels and residential towers

    Lifts, corridors and fire exits flagged with camera and time.

  • Transport depots →

    Platforms, bus bays and workshops with fuel covered from the control room.

  • Construction sites →

    Smoking and hot-work outside permitted areas, alongside helmet and vest checks.

Our successful stories

We build and deploy production-ready AI systems

Real systems delivered for real clients: custom AI development, NVIDIA Jetson edge deployment and automation that is live today. Watch the demos.

Project 01

Real-Time Smoking & Safety Detection

BytezTech built a real-time computer vision system that detects smoking behaviour and safety violations from live camera feeds. Designed for edge deployment on NVIDIA Jetson: no cloud dependency, no data leaves the premises.

  • YOLOv8
  • Python
  • NVIDIA Jetson
  • OpenCV
Problem
Manual monitoring of large facilities for smoking and safety compliance is inconsistent, delayed, and impossible to scale across multiple camera feeds.
Solution
A custom YOLOv8 model trained to detect smoking behaviour and safety violations in real-time video. Deployed on NVIDIA Jetson edge hardware, the system processes camera feeds locally and triggers instant alerts without sending footage to the cloud.
  • Real-time detection across multiple simultaneous camera feeds
  • Edge-deployed on NVIDIA Jetson: zero cloud latency, full data privacy
  • Proof-of-concept validated for industrial and facility safety environments
Read the case study

Project 02

ANPR Smart Gate: Number Plate Recognition for Societies, Parking and Factory Gates

Smart Gate reads Indian number plates at the gate in real time on an NVIDIA Jetson edge box and logs every entry and exit automatically. The demo shows live gate-camera footage from RIO IT Park, Surat, then the Setup studio, the dashboard, guard mode, the incident desk and plate-as-ticket paid parking.

  • NVIDIA Jetson
  • Object detection
  • OCR
  • Edge AI
  • Python
Problem
Guards keep a handwritten vehicle register, unknown vehicles slip through at busy hours, and nobody can say who entered, when, or which cars have overstayed.
Solution
Smart Gate reads plates from standard IP cameras, including Indian-made cameras, on a small NVIDIA Jetson box at the gate with no cloud AI API. It handles Indian plates, two-line plates and BH series plates for cars, trucks and auto rickshaws, keeps reading when the internet drops and syncs later, and gives security a dashboard, guard mode, blacklist alerts and paid parking with cash and UPI.
  • Live plate reads and automatic entry and exit logs on a real gate camera
  • Dashboard for vehicles inside, visitors, overstays and open alerts, with reports and multi-site management
  • Blacklist alerts tracked from first sighting to resolution; the handwritten register becomes a searchable digital record
Read the case study

Case studies

Free CCTV feasibility check

Send a short clip from one camera covering a no-smoking zone. An engineer tells you whether smoking can be detected reliably from that view and what the pilot would involve.

Request it

Enquiry

Talk to an engineer about Smoking Detection AI for CCTV

Tell us what you want to detect or automate, and which cameras, documents or systems you use today.

  1. 1An engineer reads your message and replies within one working day.
  2. 2A 30-minute call to look at your cameras, documents or workflow.
  3. 3A written pilot proposal: scope, timeline and what we need from you.

Send your requirement

An engineer replies within one working day. We use your details only to answer this enquiry.

Frequently asked questions

What is smoking detection AI?

Smoking detection AI is computer vision software that watches CCTV video for the posture, hand-to-mouth motion and cigarette that indicate someone is smoking, and raises an alert when it happens inside a zone you mark as no-smoking. It replaces manual monitoring of many screens.

Is this the same as a smoke detector?

No. A smoke detector senses smoke particles in the air and reacts late and only indoors. Smoking detection AI sees the person smoking on camera, indoors or outdoors, and alerts within seconds with a snapshot of who and where.

Do we need new cameras?

Usually not. Most 1080p CCTV cameras work if the zone is in view and reasonably lit. We check your footage before the pilot and tell you if a camera needs repositioning.

How are false alarms handled?

The model is trained on your own footage, zone rules limit where alerts fire, and every event can be confirmed or dismissed by your team. Dismissed events are used to retrain the model.

Does video leave our site?

No. Inference runs on an NVIDIA Jetson device on your premises. Only the alert and its snapshot are sent to the people and systems you choose.

Can it detect PPE violations or intruders as well?

Yes. Smoking detection is one model on the same edge device. PPE, restricted-zone, loitering and ANPR detection can run on the same cameras.

Related

  • AI Surveillance Systems →

    BytezTech builds AI surveillance systems on your existing CCTV: intrusion, PPE, smoking and plate detection, with alerts to WhatsApp, email and your VMS.

  • PPE & Safety Detection →

    Real-time PPE, smoking and hot-work violation detection from existing CCTV, running on NVIDIA Jetson at the edge. Alerts within seconds across multiple cameras.

  • NVIDIA Jetson Edge AI Deployment →

    We deploy vision models on NVIDIA Jetson devices next to your cameras: model optimisation, multi-stream pipelines, updates and monitoring, video stays on site.

  • AI Vision Engine →

    Real-time object detection and video analytics on NVIDIA CUDA and TensorRT. ANPR, PPE, smoking detection and sports analytics on any RTSP IP camera.