Supervision of access roads to areas with a higher level of nature protection
The use of a camera system for monitoring movement has become a standard these days. But how to solve the situation when you need to monitor an area with reduced accessibility, without the possibility of connecting to a distribution network and with limited access to high-speed Internet? All in addition, in accordance with personal data protection and GDPR, when all processing and evaluation takes place on the device itself? It is clear that the conventional camera system is already short to meet these requirements.
The representatives of the Krkonoše National Park therefore decided to find out if such an "island" solution is even technically feasible and approached us with a request to see if we could offer a suitable solution as part of a PoC (Proof of Concept).
Our solution
From the very beginning, it was clear that solving such a complex task would require the involvement of modern technologies and the examination of the limits of the electrical energy consumption of each individual component. The use of our autonomous AI sensor viSEN (VisionCraft Sensor) proved to be the ideal basis in this case.
The sensor itself is characterized by extremely low power consumption and more than 50x lower carbon footprint than conventional CCTV cameras. However, to ensure full autonomy in the field of power supply, it was necessary to go even further. The uniqueness of the solution lies in the connection of the sound detector SED (Sound Event Detector) from the Pilsen-based company JALUD Embedded, which when identifying the sound of a motor vehicle (it uses machine learning technology for detection) "wakes up" our viSEN optical sensor, which then performs an AI analysis of the image and evaluates whether the vehicle is authorized to enter the given area of the national park or not. The result (not the video itself) is sent in the form of a few KB of data via the mobile network to the Invipo system (software for smart citiesand intelligent transport systems), where it is further processed and if necessary automatically communicated further.
The solution as a whole therefore consumes a minimum of electricity and data transmissions, which is crucial for the given application. The installed batteries will keep the system in operation for 14-21 days, depending on the traffic load and the number of vehicle detections.
Conclusion
We are certainly no strangers to nature conservation, and that is why we are happy that our sensorial AI technologies can find application even in areas where it is not so usual at first glance.
In addition, the topic of nature protection has attracted the attention of the public and the media for a long time, and it was no different in this case, when Czech Television broadcast a report on the installation of our AI solution for KRNAP in the main television news on ČT1 and ČT24.
