Founder of Presien, Kieran MacKenzie, was recently invited to speak with the HSEQ Academy about the opportunities for AI vision in a heavy industry environment. Kieran covered the basics of how an AI vision system works, the safety benefits and how the technology will change the sector. 

If you want to learn more about AI vision, the webinar is a PhD’s worth of knowledge compressed into 30 minutes. Tune in below or read on for Kieran’s take on what’s next. 

What exactly is AI vision?

Put simply, AI vision allows a machine to see something, understand it and take action. AI can be trained to detect any object you like - from people to kittens. Presien is yet to receive a request to build a custom model for the latter but heavy industry is getting creative with how they leverage AI vision. The technology used to do the detections is a Deep Convolutional Neural Network (D-CNN). 

The systems have become so good that they can identify human beings by recognising fingers on a steering wheel, even when a car is travelling at 100km/h.

‍

Improving safety with AI vision – the tech that trumps them all

Worldwide, there are around  374 million occupational accidents each year. Heavy industries disproportionality contributes to these statistics. Concerningly, the data doesn’t even take near misses into account. Yet, one study found 25% of safety incidents fail to be reported. 

There are many methods that OH&S teams are using to improve worksites, ranging from a reliance on hierarchy of controls (e.g. policies and procedures) to passive user devices (e.g. reversing cameras, mirrors) and alternative technologies (e.g. tags and beacons). These approaches rarely capture all incidents because they rely on manual, and generally paper-based, reporting.

Alternatively, offerings like Blindsight by Presien can collect, analyse, and report on safety automatically. At Presien, we streamlined safety reporting further with the Blindsight Index, the first automatically generated lead safety metric. 

The Blindsight Index, calculated by the number of near-misses per machine danger hour, automatically focuses on the fatal and serious injury people-plant scenarios rather than the less severe work injuries that dominate Accident Frequency Rate (AFR) and Lost Time Injury Frequency Rate (LTIFR) metrics. In short, it helps teams understand risk before an accident. 

‍

New perspective, new process?

With a raft of benefits, AI Vision is set to change the way we work. However, with any new innovation, there are challenges and questions to be asked. What will happen to the OH&S role? Who should be in charge of reviewing this new stream of data?

One thing Kieran knows for sure is that environments like heavy industry require solutions that are simple and seamless. 

In the coming months, Presien’s Blindsight system will have the functionality to ‘triage’ incidents by providing identification of the highest-risk interactions via a ‘risk score’. This feature will empower managers to focus on the highest on-site priorities and work towards improved behaviour. 

With the power of AI vision, improving operations doesn’t have to be a heavy lift. 

‍

What can and can’t the system do?

With heavy industry still in the process of educating itself about the possibilities of AI vision, there are a lot of questions. In the HSEQ Academy webinar, Kieran tackles a range of queries including the viability of Blindsight on tower cranes, the roll-out of Presien’s systems in partnership with OEMs, the adaptability of AI vision, the impact on situational awareness and more. 

Got a question that wasn’t covered? Let’s chat.