AI safety monitoring turns the CCTV cameras you already have into a safety system. It checks that people are wearing the right PPE, warns when someone walks into a danger zone or near a moving forklift, and spots smoke, fire and fallen workers, then alerts your supervisors and keeps a record for audits.
AI PPE detection uses computer vision models to watch camera feeds and recognise people and the protective equipment they are wearing. When someone is missing a required item in a zone that needs it, the system raises an alert and records the event.
The same video analytics can watch for other unsafe situations: a person entering a marked danger zone around a press, robot or conveyor, a forklift and a pedestrian getting too close, smoke or flame in a large open floor, or a worker who has fallen and not moved. Each rule is set per camera and per area, so a food line can require hairnets while a loading bay requires vests.
Because it watches people, it has to be designed for privacy. We process video on a computer at your site rather than in the cloud, store event snapshots rather than continuous footage where possible, and set retention limits, in line with the Personal Data Protection Act No. 9 of 2022.
We check which existing cameras have a usable view, where extra cameras are needed, and which rules apply in each area.
Each camera view is divided into zones with their own rules: required PPE, no-entry areas, forklift lanes, smoke watch.
AI models running on a computer at your plant analyse the video streams continuously. Video does not need to leave the site.
A violation triggers a siren, screen alert, supervisor message or email, with a snapshot so the supervisor sees exactly what happened.
Compliance rates, near-miss counts and event images by area and shift show where training, layout or guarding needs to change.
Deep learning detectors, of the kind used in modern YOLO-family models, locate people, helmets, vests, gloves, masks, hairnets, forklifts and vehicles in each frame.
Polygons drawn on each camera view define where a rule applies, so a person is only flagged for a missing hairnet inside the food zone, or for crossing into a press guarding area.
People and vehicles are tracked across frames, which allows proximity warnings, man down detection when someone stops moving on the floor, and counts of people in an area.
Models trained on smoke and flame watch large spaces, warehouses and stores, where a camera can see smoke long before it reaches a ceiling detector.
Inference runs on an industrial computer with a GPU at your site. It keeps working if the internet goes down and keeps video under your control.
Two example project scopes showing how the station, the checks and the outputs are defined. Every plant is different, so the final configuration is confirmed after a feasibility study on your own parts.
The situation: A food manufacturer must show buyers and auditors that GMP rules are followed on every shift, and relies on supervisor spot checks.
The situation: A warehouse runs forklifts in aisles shared with pickers and has had near misses that nobody reported.
Industrial cameras, lenses, line-scan cameras and code readers are available through our machine vision catalog, quoted with delivery to Sri Lanka. We select the hardware for your application during the feasibility study.
Browse the hardware catalogEvery product, surface and line behaves differently under a camera, so every project starts with evidence from your own material.
A site visit to understand your product, line speed, quality criteria and where the check should sit.
We image your good and defective samples and show what the system can reliably detect. This is the go or no-go.
One station on one line, running beside your inspectors until results match and your team trusts them.
More lines, integration with your PLC and systems, and local support from our Sri Lankan engineering team.
Automatic counting of parts, garments, cartons and packs by camera, sensor or weight, with counts sent to your PLC or ERP. Designed and built in Sri Lanka.
Camera measurement of every point of measure against the spec sheet, plus seam inspection for skipped stitches, puckering and open seams. For apparel QC.
Unmanned weighbridges with number plate recognition, load photos and automatic gross and tare matching, for tea factories, cement plants and quarries.
Sort fabric rolls, cut panels and washed garments into buyer shade bands with CIELAB measurement and 555 codes.
Every machine vision application we design and build, in one place.
Usually yes, for cameras with an IP video stream and a clear view of the area. Detecting small items such as gloves or hairnets needs enough resolution on the person, so some areas may need an extra camera or a closer view. We survey the cameras first.
No, it does not need to. The AI runs on a computer at your site, so video stays on your network. Only alerts and snapshots are sent to supervisors, and storage can be limited to events rather than continuous footage.
Workplace safety monitoring can be done lawfully, but the Personal Data Protection Act No. 9 of 2022 applies to images of identifiable people. Good practice is to tell employees what is monitored and why, process on site, avoid face recognition unless there is a clear need, and set retention limits. We design the system around those principles.
It does not have to. PPE and zone monitoring can work on anonymous detections, reporting compliance by area and shift. Linking events to individuals is possible but is a policy decision for you, with privacy implications.
Typically within a few seconds of a violation being seen, depending on the number of cameras and the computer used. Rules can include a short grace period so someone adjusting a helmet is not flagged.
Yes, as an early warning alongside your fire detection system, not as a replacement for it. Cameras can see smoke in large open floors and warehouses before it reaches ceiling detectors, which is where it adds most value.
Tell us what you need to check and share some good and faulty samples. We will show you what a camera can reliably see before you commit to anything.
Photographs show representative plants and equipment, not Cerox installations or client sites. Diagrams by Cerox Engineering. Photo credits, via Wikimedia Commons: Construction site in Vienna by Yair Haklai, CC BY-SA 3.0; Workers in hard hats and high visibility clothing by Acabashi, CC BY-SA 4.0.