FLOW's core concepts

Get an idea of how flexible and useful FLOW really is thanks to its way of working with trajectories.

DataFromSky avatar
Written by DataFromSky
Updated over a week ago

Trajectories and operators

FLOW is all about working with trajectories and processing them with operators

Positions of traffic objects like cars, pedestrians, etc. can be observed in many ways, for example through a camera on an intersection or with GPS transmitters. FLOW uses these positions to build trajectories—spatiotemporal data, from which it's possible to calculate much more about the traffic object—velocity, acceleration, orientation, etc. 

When you have thousands of trajectories to analyze, you may want to select only some of them that meet certain criteria, like those that belong only to ambulances. That’s where operators come in.  

Operators receive trajectories and do a certain operation with them. They're the basic blocks with which you build your custom analytics. Most of them then output a set of trajectories, with which further operations can be done. The types of operators are: 

  • Filters, which investigate trajectories and output only those that meet a certain condition. For different filters, the conditions can be various: E.g. you could set up a Speed Filter that selects only trajectories that have exceeded the local speed limit. These trajectories can then be further processed, e.g. their license plates stored in a database.

  • Group operators, which tie trajectories into groups based on certain criteria, like being on a collision course with each other.

  • Set operators, which treat collections of trajectories as sets and perform set operations on them. 

Operators can be chained arbitrarily to finally select just the trajectories you need. Then you can view various information about them or have them received in your own application.

Independent and scalable

Trajectory extraction from video feeds, video files, and other sources can be done on a wide range of devices from AI cameras to powerful servers. All processing can be distributed across a dynamic number of devices according to current needs. Thanks to its high processing speed, FLOW is suitable even for time-critical tasks like controlling traffic lights based on lane occupancy.

Trajectory data can be shared and merged across many devices. This enables tracking of vehicles across multiple monitored intersections, or even the whole city if the coverage is sufficient.

Easy access to your data

Access all your FLOW assets in one place through FLOW Insights, a client application that helps you view and manage all traffic data sources, analytics, and their results.

Learn more about FLOW Insights here.

 We will be happy to answer any questions you may have. Contact us!

Did this answer your question?