How Computer Vision and Deep Learning improve Forklift Safety

Every year, there are close to 100,000 forklift accidents across the United States. A fatal forklift accident occurs approximately every 4 days. The sad reality is that most of these accidents fall into a small number of repetitive themes:

  • Pedestrians being hit or crushed by forklifts.

  • Drivers being crushed by forklifts tilting over or falling off of ramps/ledges.

  • Drivers being struck by falling materials.

The fact that many of these accidents are similar in nature means that the lessons from them are not being shared across facilities.

What if your safety managers and drivers had the ability to learn from the root causes of every forklift accident across the country? What if they had a system in place that warns them ahead of time that an accident is about to happen based on indicators from all of those other accidents? What if they could then take proactive measures to prevent the incident from occurring in the first place?

That is the promise behind Computer Vision and Deep Learning, also known as Artificial Intelligence (AI). Together, these technologies can greatly enhance forklift safety and make humans safer and more efficient.

Computer Vision refers to software processing images coming from a camera. While the concept of Computer Vision has been around for a long time, recent advancements in Deep Learning allow us to detect, recognize, and predict high-level concepts from images like never before and with oftentimes super-human accuracy and reliability.

OneTrack Computer Vision is powered by Deep Learning

For example, a system with this technology can recognize a person in an image and distinguish it from a rack or a box. The same software can also recognize if an operator on a forklift is distracted or not following standard procedures, such as driving without their hands on the wheel. When you combine the power of Computer Vision, Deep Learning, and Telematics, warehouse managers can gain unprecedented insights into their operations that tell them not just what is happening, but where, how and why.

To provide a simple example, imagine an accident occurs where an operator collides their forklift with a support column, causing damages to the forklift, product, and potentially injuring the driver. A system powered by Deep Learning not only detects that the incident occurred, but it also has the ability to learn why the incident occurred. The data captured allows it to recognize the same pattern again in the future and provide a warning when it thinks a similar incident is about to happen again.

One example of a company that has decided to implement this technology is Holman Distribution. Holman is a leading provider of Third Party Logistics (3PL) and Supply Chain Management services with base operations in the Seattle area of the Pacific Northwest and warehousing and manufacturing logistics operations all across the United States. Given the company's large network of facilities, their processes, requirements, and workflows can vary greatly from site to site. As Holman Distribution was looking for a way to boost their safety, accuracy and efficiency, they turned to OneTrack due to its unique ability to adapt, learn, and evolve.

“There is nothing more important than the safety of our team and those that we interact with in the conduct of our operations. I am confident that the Computer Vision tools will help us to keep our priority on safety, as well as positively impact the safety performance of our industry. We are excited to partner with OneTrack to implement this cutting-edge technology.”

Brien Downie, President, Holman Logistics

What makes the implementation of a Deep Learning powered system so compelling is its ability to continuously improve. The system continues to learn while it is deployed, actively preventing incidents, and getting better at it as well. With every device that goes online, our customers' forklifts become smarter, faster, and more accurate.

That means that accidents that occur at a 3PL in New Jersey can enhance forklift safety at a manufacturing plant in Oregon.

As the system continues to learn, drivers become safer and more effective, and the number of accidents can be effectively reduced over time.

OneTrack has taken the lead in deploying Computer Vision and Deep learning in warehouses. Within our fleet, we have tracked hundreds of thousands of hours of forklift operations. Not only do we provide our customers with an unmatched level of visibility, but we help them prevent safety issues from occurring in the first place. All day, every day.

Reach out to us to learn more or for a demo!