LiDAR imaging can monitor shopper numbers, but also where they browse and what interests them – all while maintaining their privacy
“Counting accurately was a nightmare – you can’t really count people, it’s so boring,” says Joe Morris, director of innovation at retailer TJ Morris, owner of Home Bargains. And it’s expensive: tasking a single member of staff to watch the door at 500+ stores for one hour a day would cost two million pounds annually. “Having someone outside for ten hours a day would be a massive cost,” says Morris.
Instead, for its Home Bargains stores, the company turned to LiDAR – a technology more frequently associated with driverless cars or digital 3D modelling – and has installed Hitachi’s 3D LiDAR Motion Sensor in 70 stores, with plans to roll it out across all 550 locations.
The technology is simple: a Hitachi sensor mounted on the ceiling beams out rays of light, tracking the time it takes for the laser’s light to bounce back, painting a rough picture of objects and people in the area. “By calculating the time of flight, you automatically know the distance and the shape of the body, which our software can recognise,” says Hideki Hayashi, product manager, Digital Media Group at Hitachi Europe. Paired with machine learning analysis, the system can accurately count how many people step over a virtual line or pass a specific point.
Accuracy was key for TJ Morris, not only for safety but for enabling more advanced use cases, such as business analytics. But while cameras have long been used to count foot traffic, Morris found after testing several that they failed too often – so until now the company hasn’t bothered to implement such systems. “LiDAR is quite a reliable technology,” says Morris. “It’s less susceptible to the Sun coming in and out of the windows, which causes low contrast on images.”
That said, the Japanese-designed LiDAR system needed a few tweaks to work with British shoppers. “Our sensor was optimised for Japan, so we found some issues,” Hayashi says. “Japanese people are typically very small and slight, and so tracking people of a different, larger build here in the UK was actually more difficult. Also in the winter in the UK, many people tend to wear very dark/black clothes – in Japan, this isn’t so much the case. Dark colours affect accuracy, especially on rainy days.” When it was rolled out at the first Home Bargain stores during damp weather, wet and dark clothing knocked accuracy down to 96 per cent, but after an update to the machine learning system, it’s now back above 99 per cent.
Beyond sensor accuracy, TJ Morris wanted to integrate the Hitachi system into its own bespoke management dashboard, which – among other tasks — automatically controls a traffic light at the entrance which signals green if would-be shoppers can safely enter the store, or red if they should wait for someone else to leave first.
And there are further uses: “The key thing we liked was the ability to work directly with Hitachi to develop this and look at other applications,” Morris says. “We see the sensor not just giving us counts of customers in and out, but potentially other information.”
For example, adding more sensors could allow Home Bargains managers to spot when a queue begins to form and alert staff to head to the tills, while precise visitor counts could be compared to the number of sale transactions for better business insights. “Our store managers have a huge amount to think about, and we give them a lot of autonomy in terms of how they run their store,” says Tom Morris, head of digital at TJ Morris. “So the more manual tasks that we can take off them, like counting how many people are in the store and how many more to let in, the happier they are.”
Indeed, while LiDAR is a useful tool for social distancing, this system wasn’t actually designed to be a people-counter, but for understanding how customers move through a store by building a heatmap, tracking where they linger, which items they’re drawn to and which they pick up and put back on the shelf – and even spot potential shoplifters. “We’re looking at turning our stores into a more integrated, sensor-driven environment, where sensors are aiding our store management to do their job, taking some heavy lifting off of them so they can do more of the value added stuff,” says Tom Morris.
That’s the sort of data an online store can collect as shoppers add and subtract items from virtual baskets — but with LiDAR there’s built-in privacy protections, as it can’t identify any individual. No images are captured, as with camera tracking, just a rough outline of a person. “It’s essentially a load of black and white dots. I couldn’t recognise you from it, I can’t see your face in that image,” adds Tom Morris.
And that means the data being collected isn’t a privacy risk. “We don’t have to put the layers and levels of security on this that we might otherwise need to for GDPR with traditional cameras,” Tom Morris adds. “We don’t want customers to feel that as we’re introducing new safety measures to protect them while they’re in store that they have to compromise privacy.
And safety is the core reason Home Bargains turned to Hitachi’s LiDAR in the first place. The technology may save money and offer additional business benefits, but the primary purpose is taking care of customers and staff. “We’ve scientifically decided what number to let in and are using technology to control that,” Joe Morris says. “There’s a lot of anxiety, and it is important that our customers and team members can see that we are controlling customers numbers in stores”.
–Modern life is saturated with data, and technologies are emerging nearly every day – but how can we use these innovations to make a real difference to the world? Hitachi believes that social innovation should underpin everything it does, so it can find ways to tackle the biggest issues we face today. Visit Social-Innovation.Hitachi to learn how Hitachi Social Innovation is Powering Good and helping drive change across the globe.