The cameras on the shop floor already record video — NeuroVision video analytics turns it into occupational safety events. The neural network spots a person without a hard hat, vest, goggles or safety harness, entry into a danger zone and a fall, and immediately sends the area foreman the frame, the zone and the time. The system does not identify who committed the violation: this is control of protective equipment, not biometrics.
NeuroVision PPE control
We run the pilot on your cameras and recordings of past shifts — the accuracy is visible at your site, not in a presentation
Issuing a hard hat is not enough: the Labour Code requires the employer to monitor whether employees use protective equipment correctly. A safety engineer cannot stand at every crane, but the cameras are already there. Video analytics for manufacturing watches them on their behalf and reports only violations.
We connect to your IP cameras and video server. The pilot runs on recordings of past shifts and one or two areas.
At a scheduled walkround everyone is wearing a hard hat. Violations happen between walkrounds: someone took the hat off in the heat, stepped under a crane “for a second”, climbed the scaffolding without a harness. The neural network processes the camera stream all day and records every such episode.
The event reaches the area foreman while the person is still in the zone: go over, stop the work, hand out the missing protective equipment. The shift report shows where violations repeat and what to change — a sign, a barrier or a briefing.
For each zone you choose which protective equipment is mandatory there and at what hours. The neural network checks every person in the zone and flags a violation if it lasts several seconds — a stray frame does not turn into a false alarm.
Crane, assembly and loading zones and construction sites. A person in the zone without a hard hat, or with the hat in their hand — an event with a frame and a time. The rule can be enabled only for the hours the crane is operating.
Forklift lanes, the warehouse, in-plant transport routes, open-air yards. The neural network sees whether a person is wearing a vest with reflective stripes, from a distance at which the foreman cannot make it out.
Paint and welding stations, laboratories, food production. Control of masks and goggles where the instructions require them: a mask on the chin or goggles on the forehead — also a violation.
Scaffolding, roofing, steel structure assembly. If a person is in the work-at-height zone without a safety harness, the event goes to the person responsible for work at height.
A person under a load, next to a running machine, behind a barrier or in the path of a forklift. Entry into the zone is recorded regardless of whether the person is wearing a hard hat.
If a person has fallen or does not get up for a long time in a shop, a warehouse or a tank farm, the event goes to the foreman and security. Help arrives before colleagues notice.
Example — the finished goods warehouse of the Severny Profil plant: forklift lanes, racking and the shipping ramp, where a vest and a hard hat are mandatory.
The occupational safety service takes its own instructions and issue standards and maps them onto the camera frames: “Lanes — vest”, “Ramp — hard hat and vest”, “Charging room — goggles”. On each camera the zones and the hours when the rule applies are outlined.
Before alerts are switched on, the analytics is run over several days of archive. You see which events it found and together with us tune the threshold: what counts as a violation and what is just a person who took off a hard hat for a second by the site hut.
During the shift the neural network processes the camera stream. A person in a lane without a vest for more than a few seconds — the foreman receives an event: the frame, the “Lane 2” zone, the time 10:42. They walk over and decide on the spot.
The foreman marks the event as “resolved” or “false”. At the end of the shift and the month the occupational safety service sees a report: violations by zone, shift and type of PPE. False events teach the setup, and there are fewer of them over time.
Shop floor and overhead cranes
A hard hat in the crane zone, the danger zone under a load, a person at a running machine. Ceiling-mounted cameras see the whole bay at once.
Warehouse and logistics
A vest in the lanes, a person in the path of a forklift, a night shift without a foreman. The event goes to the shift supervisor and security.
Construction sites and work at height
A hard hat and harness on scaffolding and roofs, entry into the assembly zone. A camera on a mast or crane looks at the site from above.
Painting, welding, chemicals and food production
Goggles, a mask and a respirator at stations where the instructions require them. Here the violation is visible only up close — the camera is closer than the foreman.
Production safety with video analytics starts where the cost of a mistake is higher and it is harder for the foreman to keep track: large areas, many people and vehicles, work in shifts and at night.
To check contractors and workers before they are admitted to the site — see the page for industry and construction. Perimeter, entry and outsiders on the premises — video analytics for site security.
Occupational safety with video analytics relies on rules that are already in force: the employer issues PPE, monitors its use and does not allow people to work without it. Below is a summary of the rules as of October 2026: a guide for the occupational safety service, not a legal opinion.
Occupational safety video analytics answers the question “is this person in this zone wearing a hard hat”, not “who is this person”. That is why it is easier to deploy: no biometric consents or lists of faces are needed.
The system sees a person and the item on them, but does not compare the face with a database of employees. Who stood in the lane without a vest, the foreman finds out for themselves when they walk over.
An event neither punishes nor stops the line by itself. The foreman walks over, finds out the reason and decides: hand out protective equipment, stop the work or suspend the employee under the procedure adopted by the company.
The foreman sees their own areas, the occupational safety service — the whole plant, security — fall events. The archive and frames go no further than those who need them.
The cameras are not used to count who worked how long and when they went on a break. The module is set up only for protective equipment, zones and falls — what occupational safety requires.
Pilot — 1 month free
Cloud in Russia or your own server without internet
Your cameras and archive
The price depends on the number of cameras, the types of PPE and where processing takes place — in the cloud or on your server. Start with a free pilot in one area.
One or two cameras and an archive of past shifts, zone marking, threshold tuning, going through the events with the occupational safety service.
on request
Processing on servers in Russia, events and reports in the dashboard, API and alerts.
on request
Installation next to the plant’s video server, without internet access, a licence by the number of cameras.
Cameras, the video server and installation are not included in the price — that is your equipment or the integrator’s work. Video analytics is not in the price matrix : we calculate the price per project.
Terms from instructions, Rules 766n and conversations with an integrator.
The project «Development of a multimodal system using artificial intelligence "Neuromed" for managing the diagnostic processes of a clinic on the basis of the "Neurovision" computer vision platform» is supported by the Innovation Promotion Foundation within the federal project «Domestic Solutions»
The Neurovision computer vision platform is registered with Rospatent (certificate No. 2023661508) and included in the Russian Software Registry (entry No. 19972 of 13.11.2023)
The right to use the Platform is granted under a licence agreement concluded by Neurovision with its customers
Tell us which areas and protective equipment need to be monitored, how many cameras there are and whether there is an archive, and whether you need a server without internet access — we will offer a pilot on recordings of past shifts.
Can't put your request into words? Book a call and our team will contact you at a convenient time
Book a callThe company is included in the register of IT companies and the register of personal data operators of the Russian Ministry of Digital Development.
Address: Moscow, Presnenskaya Embankment 12 — Moscow City business centre, Federation Tower. Nearest metro stations: Delovoy Tsentr and Mezhdunarodnaya.
Materials on face recognition, algorithm accuracy and protection against deepfakes — in our blog.
It is a neural network that processes video from cameras in a shop, warehouse or site and flags occupational safety violations by itself: a person without a hard hat or vest, entry into a danger zone, a fall. The foreman receives an event with a frame, a zone and a time instead of watching all the screens.
A walkround shows what is happening at the moment of the walkround. A camera watches the whole shift, including nights and weekends. Walkrounds remain — video analytics suggests where to go first and gives a report on where violations repeat.
A hard hat, a high-visibility vest, protective goggles, a mask and respirator, a safety harness at height. Separately — entry into a danger zone, a person in the path of machinery and a fall. For each zone you choose what is mandatory there and at what hours.
Usually not: already installed IP cameras that see the whole person are suitable. For goggles and masks you need a camera closer to the station — we will choose it during the pilot on recordings of your shifts.
No. PPE control by camera does not compare faces with a database of employees: the event contains a frame, a zone and a time. Who stood without a hard hat, the foreman sees for themselves when they walk over.
The Labour Code allows the employer to use remote video recording of work processes to monitor safety (Art. 214.2 of the Labour Code) and requires employees to be informed about it (Art. 214). Consent to biometrics is not needed: PPE detection does not establish identity. Set out the purposes and procedure in an internal act.
That is a different scenario — access control. Entry by face works through access control with face recognition under the rules of Federal Law 572‑FZ; the PPE control module does not recognise faces. For how video analytics works with faces in general, see the page video analytics.
Go over and find out the reason: hand out the missing protective equipment, stop the work in the danger zone. An employee who does not use the issued PPE must be suspended from work by the employer (Art. 76 of the Labour Code) — under its own procedure; a video analytics event helps to notice the violation, the decision is made by a person.
The first month of the pilot in one area is free. After that we calculate the price by the number of cameras and types of PPE, separately for the cloud and for installation on your server without internet access.
