The modern retail market has to be incredibly competitive to be able to survive in the age of online shopping and mega-supermarkets. An important part of this survival is gathering reliable customer information, so retailers can keep ahead of the trends and make business decisions to stay relevant and pull people into their ‘bricks and mortar’ stores. Centrum Schalkwijk a Shopping Mall in Haarlem, the Netherlands, is using Hikvision cameras, to help with that. The cameras, installed by Cocon Security – based in Breda in the Netherlands, are delivering accurate people-counting statistics to provide crucial information for their business strategy and management.
Traditional methods of counting people going in and out of the mall were not effective for the shopping centre. They involved counting the numbers of people connecting to the centre’s Wi-Fi. Since people can walk around the mall without entering, and may not have switched their Wi-Fi on, this was not very accurate. These methods are also limited in the information they can provide.
“People counting is very important to shopping centres because the retailers always want to know what ‘footfall’ they are likely to expect,”. “This information helps the larger chains, for example, to make a decision on whether or not to invest in the centre and move in.”
Individual shop managers can see when they are likely to need more employees to cope with the busy times, helping to create shift rotation, for example.
The real estate owners can also use the information to calculate optimum rental prices for the retail spaces in the centre.
Cocon Security installed eight Hikvision Dual-lens DeepinView People Counting cameras (iDS-2CD6820F/C) around all of the entrances to the shopping centre. The cameras, armed with Deep Learning technology, are also equipped with binocular stereo vision, 3D people detection and tracking capabilities, and height filtering – all contributing to their amazing 97% accuracy rate.
These cameras were picked for their accuracy, which had to be higher than 95%, and their Deep Learning capability.
Deep Learning technology gives the camera a lot more ‘intelligence’ to filter out ‘false alarms’, or inaccurate data. For example, it can distinguish between different people, even if they enter in a group close together. There’s also the ability to track individuals and filter for ‘loitering’, which helps with the overall data analysis.
Of course, every solution will have its differences. That’s why the cameras were adjusted to meet the needs of Centrum Schalwijk and tailored to the environment. For example, tweaks were made to the cameras to discount prams and trolleys from the count.
Bringing the whole solution together in Centrum Schalwijk’s Security Centre is a Hikvision Blazer Express (v1.3), an easy-to-use and highly efficient intelligent video management software (iVMS) station, designed specifically to meet the needs of small- to medium-sized retail surveillance applications.
The data from the cameras can be analysed, displayed and reported on, with a function to export to the centre’s own software platform. There, the data can be displayed and reformatted into spreadsheet reports, for example. The system can integrate 16 cameras in total and simultaneously bring data from all of them to analyse. In fact, this is the only suitable solution that can bring so many cameras together and is designed to make access to data and analysis a lot easier.
Reports from the system can be configured for: Number of people entering/exiting, number of people entering OR exiting and data from a single camera or multiple cameras. There is also an auto-sending function which automatically sends reports to designated recipients, streamlining the process and getting accurate information to the people who need it quickly.
Centrum Schalkwijk: We are amazed with the accuracy of the new solution – it is leaps and bounds above our previous people-counting methods. This will be incredibly useful to provide centre management with footfall information that they can use to make business decisions in the future.