Marwiz Vision: AI-Based Security & Smart Surveillance Software

Smoking & No-Smoking Zone Detection: AI Compliance Monitoring for Workplaces

Last Updated: 26/10/2026
by Marwiz Vision
Turning Raw CCTV Footage into Boardroom-Ready Reports: How Marwiz's Reporting Engine Works

Almost every Indian workplace has a "No Smoking" board. Far fewer have a way to know whether anyone is following it. One compliance study of a tertiary health-care institution in a city that had been declared smoke-free visited 40 public places on its campus and found the overall compliance rate with the smoking ban was a mere 23%. A separate study in South Bengaluru found that among public places required to have designated smoking areas, only 19% complied. The rule is clear and the signs are up. What is missing is visibility. PubMed Centralnih

What Indian Law Requires of Workplaces

The Cigarettes and Other Tobacco Products Act (COTPA), 2003 is the base law. Its key points for anyone managing a premises:

  • Section 4 bans smoking in public places. In 2008, the prohibition was extended to workplaces, and the rules specified the size and design of the No Smoking board and defined "designated smoking area." PubMed Central
  • The duty sits with the person in charge. Under Rule 3 of the 2008 rules, the owner, manager, or person in charge of a public place must ensure that nobody smokes there, and no matchbox or ashtray should be provided. Nyaaya
  • A named person must handle complaints. The rules require the owner or manager to designate a person to whom complaints can be made. Tobacco Control Laws
  • Exceptions are narrow. Smoking areas are allowed at airports, hotels with 30 or more rooms, and restaurants with seating for 30 or more, but not for ordinary workplaces. Tobacco Control Laws
  • The fine is small. A Section 4 violation carries a fine of up to ₹200 and no imprisonment. Section 172 BNSS

The low fine is often read as a sign that enforcement does not matter. Commentators note that a ₹200 fine is too small to deter habitual smokers and few officers are designated for enforcement. For an employer, the picture is different. The legal duty to keep the premises smoke-free rests with management, and the practical costs of a violation, such as fire risk, complaints, insurance questions, and audit findings, are far larger than ₹200. The Global Adult Tobacco Survey for India, cited in one study, found that around 26% of adults are exposed to second-hand smoke at the workplace. Dr. Abhishek Gandhinih

Why Signs and Patrols Don't Work

  • Smoking is brief and hidden. A cigarette takes a few minutes and often happens behind a building, in a stairwell, or in a storeroom, exactly where guards are not.
  • Patrols are predictable. People learn the route and the timing.
  • No record. Even when a guard catches someone, there is rarely evidence, a timestamp, or a way to see whether the same spot is a repeat problem.
  • Nobody owns the hotspot. Without data, management cannot say which zone, shift, or gate has the problem.

The stakes are higher in some environments. In a warehouse holding packaging or chemicals, a fuel yard, a pharmacy store, a paint shop, or a textile godown, a lit cigarette is an ignition source, not only a health rule. In hospitals, schools, and offices, the concern is second-hand smoke exposure and duty of care.

How AI Smoking Detection Works

The AI runs on the IP cameras already covering your zones. You mark the areas where smoking is prohibited, and the system watches them continuously.

  • Zone setup: Define no-smoking zones on the camera view, such as a loading dock, stairwell, fuel yard, or ICU corridor. Designated smoking areas, where the law allows one, can be marked as permitted.
  • Behavior and object detection: The model looks for cues such as a cigarette or beedi in hand or mouth, the hand-to-mouth motion, and visible smoke.
  • Event logging: A confirmed detection creates a timestamped record with camera, zone, and an image or short clip.
  • Alert and review: Security or the safety officer receives an alert, and a person confirms the event before any action is taken.
  • Reporting: Events are aggregated by zone, shift, and week so management can see patterns and target the hotspot.

Signs and Patrols vs. AI Monitoring

Factor Signage + Guard Patrols AI Smoking Detection
Coverage Only where a guard happens to be Every covered camera, 24x7
Response time Depends on the next patrol Alert while the event is happening
Evidence Verbal report, if any Timestamped image or clip
Hotspot visibility Anecdotal Data by zone, shift, and week
Consistency Varies by guard and shift Same rule applied everywhere
Deterrence Fades once the routine is learned Sustained, because monitoring is continuous

Where This Matters Most

  • Warehouses, godowns, and logistics hubs: Stacked packaging and combustibles make an ignition source a real hazard. Detection helps management catch behavior before a fire, and it supports what your fire safety audit expects.
  • Fuel, chemical, and pharma sites: Any area with vapors or flammables needs a strict no-ignition rule, and the log shows the rule is monitored.
  • Hospitals and healthcare campuses: Smoking near wards and ICUs harms patients and staff. The health-care study above shows how easily compliance slips even inside a smoke-free campus.
  • Offices, schools, and colleges: Stairwells, terraces, parking corners, and washrooms are typical hidden spots. Detection helps the administration act on complaints with evidence.
  • Hotels and restaurants: These are the exceptions where designated smoking areas are allowed. Detection helps confirm that smoking stays inside the designated zone and out of lobbies, corridors, and guest rooms' surrounding areas.

What to Check Before You Deploy

Check What to Ask Your Vendor
Real-world accuracy A cigarette is a small object. How does the model perform at your camera distance, at night, and with vapes, incense, or steam that can look like smoke? Ask for a pilot on your own cameras.
Camera fit What height, angle, and resolution are needed for reliable detection in the zone?
Human verification Can an operator review each alert before any action is taken?
Zone flexibility Can you set different rules for no-smoking zones and designated smoking areas?
Reporting Can you see violations by zone, shift, and period, and export them for audits?
Privacy and retention How long are images and clips stored, who can view them, and how does this fit the Digital Personal Data Protection Act, 2023?
Policy readiness Do you have a written policy and a communication plan so staff know monitoring exists?

Using It Fairly

Detection works best when it supports a clear, communicated policy and not a surprise. Tell staff and visitors that monitoring is in place, use the first weeks to raise awareness, and keep a human in the loop to confirm events. Where a designated smoking area exists, make sure people know where it is. The aim is fewer violations, not more penalties.

From Signage to Evidence

A No Smoking board states a rule. Detection shows whether the rule is being followed. Management gets a clear picture of where and when violations happen, a record for audits and insurers, and a faster response in the places where smoking is most dangerous.

Marwiz Vision runs this kind of detection on the same platform as its restricted-area monitoring, PPE detection, and fire detection. A smoking alert, a restricted-zone breach, and a fire alert can then appear in one dashboard, with events logged and reported in one place.

Frequently Asked Questions

Under COTPA and the 2008 rules, smoking is prohibited in public places, which the rules extended to workplaces. Narrow exceptions exist for airports and for certain hotels and restaurants with designated smoking areas. Confirm current rules with your legal team.

The rules place the duty on the owner, manager, or person in charge to ensure nobody smokes in the public place. The individual who smokes can also be fined.

COTPA provides for a fine of up to ₹200 for a Section 4 violation. Employers may also have their own disciplinary policies.

This is a real limitation. Cigarettes are small, and vapor, steam, and incense can be confused with smoke. Accuracy depends on camera placement and lighting, so test it on your own site and keep a human reviewer in the loop.

Usually not. It can run on existing IP cameras if the angle, height, and resolution suit the zone. A site survey will confirm.

Detection can work at the event level, flagging a violation in a zone without naming a person. If you link events to identity, for example through access control, make sure your policy and data handling allow it.

A time-stamped log of monitored zones and violations gives auditors evidence that the rule is actively enforced. Whether it meets a specific audit standard depends on that standard, so check with your auditor.