There is a developing situation in the factory. A forklift has accidentally touched a production machine, which is now leaking oil. This causes several potentially dangerous circumstances. Oil on the floor makes the surface slippery. The damage to the machine has caused it to start overheating. In efforts to assess the incident, workers nearby are taking other routes than usual.

At a high-rise construction site, a crane lifts a pallet of bricks from a truck. As it continues pulling up the cargo, it turns 90 degrees to reach the 4th floor where a bricklayer is building an exterior wall. Due to a sudden gust of wind, the load swings slightly too fast and nearly hits the worker. This time, fortunately, the bricklayer and the crane driver both come out with a fright.

At manufacturing plants, in warehouses and at construction sites, accidents and near-accidents happen all too frequently despite the range of safety equipment and rules that have been put in place over the years.

And they are very costly. Think of the health and well-being of employees and their families, possible damage to property, potential delays in deliveries, insurance costs, investments in safety equipment and procedures, and the risk of reputational damage to an employer brand, or indeed a consumer brand.

Nobody's perfect

In some cases we don’t even know what the right safety measures and margins should be. Human error is nearly always a factor. Some people are more rule-abiding or careful than others, but everyone gets tired at times or absent-minded, even for a few critical seconds.

Accidents are difficult to predict. They are unique almost by definition. And yet there are patterns of circumstances and behaviors that, if we just knew them and knew how to recognise them, could help us prevent accidents.

Advian’s Intelligent Safety solution with video-assisted Edge AI can help prevent accidents in real time.

Our sophisticated artificial intelligence software analyses fused data streams from environmental sensors as well as video data.

Sensors can constantly monitor things like temperature, humidity, air or water pressure, altitude, motion speed, sound, smoke, chemical composition, radiation, etc. With our artificial intelligence algorithm, camera’s can “recognise” people, animals, vehicles and other objects, calculate how fast and in what direction they move, and the distance between them.

Data to learn from

Our solution can alert workers of gas leaks, fire, and smoke. Or when a place turns exceptionally hot or cold, a floor surface is unusually slippery, or a colleague has fallen from a structure. If it detects a potentially dangerous situation, it can even take action itself. It could, for example, remotely put the brakes on a moving forklift.

It can learn to identify patterns between people’s behavior, what (safety) equipment they are carrying, how and where machinery moves, the strength and safety of building structures, the location of potentially dangerous materials, weather conditions, and so on.

accident-prevention-solutions

The more data we feed into the system, the more accurately it will recognise circumstances that could be problematic.

Intelligent Safety can play a role in designing the layout of factories and warehouses, and the planning of logistics. For example, video-assisted Edge AI can help us analyse how, where, when and at what speed trucks and forklifts move on the premises. The artificial intelligence can learn about safe speeds and safe distances between machines and people, and how to optimise routes.

Safety managers regularly take safety walks around the premises to inspect for safety and report any issues. The manager might check that exits are closed, fire equipment is in place, scaffolding is safely installed, and no potentially dangerous material is left unattended.

At the cutting edge

With video-assisted Intelligent Safety and sensor fusion, many of these checks can be performed automatically without the need for the manager to enter a potentially unsafe situation. Also, from the analysis of video data during normal factory or construction operations, the AI can surface anomalies in activities and behavior that could pose safety risks. For example, it might detect when people at a plant tend to accidentally misplace fire equipment or forget to close exits.

We also take data from observing the wider environment into the equation. In a freight harbour, straddle carriers lift and transport sea containers to and from ocean vessels. The process is increasingly automated. If another vehicle or a person approaches the route of the carrier, we can detect this with cameras in different places on the quay. Real-time AI can inform all vehicles and people involved when there is a risk of collision.

With Edge AI, we put the computing power to the edges of the network, close to where the data are collected, because sending it back and forth for analysis in the cloud or on a centralised server would not be feasible. It would simply be too slow for real-time alerts and both the data communication and central computing capacity would be prohibitively expensive.

Edge AI data analytics is about to take a huge leap. For those interested in finding out more we’ve put together an eBook, ‘An introduction to Intelligent Safety: Preventing accidents, cost-efficiently, with video-assisted Edge AI’. Download it here and learn:

💡 How video-assisted AI at the edges of digital networks can significantly improve accident prevention

💡 Which Edge AI use cases may benefit your organisation

💡 How to start your Edge AI journey by taking your first steps today.