Automatic seat occupancy detection for public transport
About the Product
Die Seat occupancy detection of ViSenSys is an innovative concept that consists of a Multi-camera system, image processing software and Machine learning .
The video images are evaluated with the help of special software. The metadata are only recorded in the vehicle or at the station. The data such as train capacity or passenger numbers can be viewed in real time or historically.
The software never displays actual video material from the camera systems. It relies solely on metadata. The output that the user sees is a graphical reconstruction of the video images.
Software output from the user's point of view
The AI anonymously recognizes persons, luggage, bicycles etc. and assigns them to seats or standing room and evaluates them with a temporal integration.
With the help of AI and machine learning we can guarantee an unsurpassed accuracy in observation.
No personal data is stored! No manual counting necessary. More accurate statistics in real time for customers. VDV 457 zertifiziert.
Duration the occupancy of a seat from boarding to leaving the station (incl. time integration)
Anonym Survey of seat occupancy.
How does it work?
Hardware requirements | security standards
The system is based on existing camera installations, whether analogue or digital. We can also incorporate new camera systems and do not just rely on existing ones.
The transmission of all data required for evaluation is always encrypted and meets all necessary data protection requirements. No personal data of the passenger is stored or transmitted
Functionality of capacity counting
STEP B - Processing
Distortion in 4 levels
The digitized image is stored in the memory for max. 0.1 seconds and then irrevocably deleted.
In addition, the hardware that performs steps A and B is located in locked areas and is therefore only accessible with a key. Software access (hacker attack) is not possible because the hardware side ensures that there is no physical connection to the outside world (e.g. network/internet connection).
Processing using a single sample image as input
Levels 1 to 3 show the reduction in resolution and distortion. Since different levels of distortion are required, the number of images per level increases accordingly.
The images in level 4 are applied to each image in level 3 to determine the number of people in the distorted images. No shapes or colors are used that would allow a person to be identified. Each level is optimized for mechanical counting only.
Information about capacities
With the software for the Optical seat occupancy detection you can provide travellers with information about the capacity of a vehicle available. With this feature, passengers can choose where best to board the vehicle. The statistics can be visualised for the user in a dashboard and exported there.
Display option: mobile devices, apps, website
You can also provide information about the capacity of a vehicle available via mobile devices or on websites.
Certified according to VDV 457 and EN 50155
Companies that already rely on our solutions
Our team of experts from various fields is at your disposal. We guarantee highest quality for our products and services.
ViSenSys accompanies you through all project phases. We offer consulting, development, installation, calibration, training, upgrades, updates and much more.
We are also happy to develop individual concepts and the corresponding software together with you