Monitoring people movement intensity
Fortunately, gone are the days when an "army" of students with manual counters was hired for similar measurements. Likewise, it is no longer necessary to build a complex (and expensive) data collection infrastructure. Using modern technologies and AI, it is possible to achieve the desired result easily, quickly and automatically - so with minimal effort.
But how do you deal with a situation where the event is held in a publicly accessible space where hundreds to tens of thousands of people move every day? It's simple - first you need to get reference data for the monitored area, i.e. the number of people who are usually traversing the area at a given time of day. Furthermore, it is advisable to "clean" the data of various external influences, such as the weather. If we let our AI-based viSEN sensor work for some time, it gathers all the necessary reference data. Subsequently, nothing prevents us from being able to reliably measure the number of visitors to an event held for a given area as we can cross reference the difference from the average base rate and the peaks that happen during the measured activity.
Our solution
To monitor the intensity of people's movement, the city of Brno chose one of the most frequented areas - náměstí Svobody (The Freedom Square). When designing the implementation of the contract, it was considered to use either a viBox (VisionCraft Box), i.e. a "smart" box that connects to the city's already existing camera infrastructure, or our AI autonomous sensor viSEN (VisionCraft Sensor). In this particular case, the latter was deemed a better fit.
The advantage of the viSEN sensor is its easy installation and quick data collection right during the install date. We usually mount the sensors in cooperation with a partner (here Technical Networks Brno) on a public lighting pole, where we only need LTE signal and a power cable with 3-4 hours of electricity. Everything else is taken care of by the sensor itself. The first data are available online within a few hours after the installation is completed and are presented either on a clear dashboard or sent directly to the city's information system. The City of Brno uses both of the mentioned options, where the dashboard with data presented in real time is accessible to anyone at a public URL address. The viSEN sensor is installed on the border between Masarykova Street and Namesti Svobody, thanks to which it is able to track the movement of people in both directions.
As is customary with the solutions we provide, the entire monitoring is carried out with maximum emphasis on privacy protection. All image data processing takes place in the sensor itself, which then only sends anonymized metadata about the monitored objects - no photos or videos leave the sensor.
Thanks to this solution, the city of Brno has a comprehensive overview of the intensity of the movement of people in one of the most frequented places in the center, as well as the number of visitors to the events that take place on Náměstí Svobody throughout the year.
Conclusion
Helping companies and institutions make better decisions based on data, not just guesswork - that's the goal we set for ourselves at VisionCraft, and it's great to see that there are cities that realize the importance of data for their operations and efficient management of public money. If the city also decides to make the data available online, it is an excellent example of how it can operate transparently for its citizens today.
Statement of the contracting authority
Measuring the intensity of pedestrian traffic on Brno's busiest street has long been a problem for the city. In the past, these statistics were frequently requested by entrepreneurs, event organizers or authorities who wanted to estimate, for example, the decrease in people on the streets during the COVID-19 pandemic. VisionCraft edge AI sensor, which has been operating in the largest Brno square for several years, has two major advantages. First, the numbers are available in real time, including historical statistics. Secondly, we do not have to deal with the protection of personal data, because we only receive metadata about the objects, not image data. During the period of operation, the numbers have already been used several times for various estimations and evaluations of city center attendance.
Mgr. Jan Zvara, head of the data, analysis and evaluation department, Brno
