Monitoring hundreds of camera views to identify incidents in real time would be impossible without video analytics. Our video analytics suite is an exception-based monitoring tool that uses the latest Artificial Intelligence technology and deep learning models to analyse footage and optimise the control room. Several features ensure smooth setup, ease-of use and accurate results. These include size filters, calibration, inclusion zones, image stabilisation, shadow suppression, and direction settings. Our background model, with accurate tracking, fixed lighting, travelling average and adaptive noise suppression, fine-tunes the calibration process.
Configure analytics on pre-recorded or live footage, so algorithms can identify and log incidents in a database for review. Retrieving evidence is easy.
Create a safe, efficient site by using video analytics to trigger automated actions. These include alarms, sending emails, and recording footage.
System reliability and low maintenance maximise up-time and cost-effectiveness. Real-time feedback on algorithms makes setup intuitive.
Our video analytics suite has many applications, from automated responses to motion in an area, to identifying objects and making sure they are where they should be. Using deep learning models, CathexisVision can classify a huge range of objects and set the conditions under which they will trigger actions. If a suitcase is left in an airport lounge, a chair blocks a fire escape, or a vehicle is abandoned in a parking lot, object detection can identify the stationary object and raise an alarm. If an object is removed, operators can be notified.
Rules can be added to object classification, so that the speed or direction of a person, vehicle or animal can be tracked without setting off false alarms. Other rules include picking up loitering, stopping in an area, or the enter and exit area algorithm, which could trigger events if an object enters and/or exits a zone.
- a person enters a restricted area
- an item is left unattended
- a vehicle stops on the shoulder of a freeway
- someone loiters next to an ATM
- a queue is too long
- a vehicle speeds
- someone attempts to scale a perimeter wall
- people enter or exit a shop
- blocked or authorised vehicles attempt to enter a premises
- a venue reaches capacity
CathexisVision uses deep learning models based on neural networks to simulate mental processing and differentiate between objects.
Choose from several algorithms to find the most effective solution, based on the scene and the site’s requirements..
- Link video analytics with integrations
- Review footage with SMART search
- Control cameras
- Archive video
- Investigate with a multi-layered map
- Database events and operator activities
- Assign priority levels and resources to events
Our analytics enhance the management of sites in any sector, optimising safety and efficiency. Be it in hospitality, residential estates, conservation, logistics, healthcare facilities, commercial properties, campuses or cities – the applications are vast.
In retail settings, the queue length algorithm can detect if a queue of people or objects is occupied for too long. For example, an audio clip calling staff to attend to a till can be played automatically when a queue of customers is too long. Line crossing can be used to count people entering or leaving a store or other establishment.
The CathexisVision line crossing algorithm can be easily adjusted to meet site specifications. Advanced line crossing can track a person if they cross a double or segmented line around a perimeter. CathexisVision’s other counting algorithms can count people crossing a line, from cameras mounted directly above or obliquely.
Managing vehicles is simplified with our algorithms. The speed rule provides alerts and recordings when vehicles travel over a desired speed limit. Working with a range of standard or specialised cameras, automatic number plate recognition uses character recognition to identify license plate details from regions worldwide.
We offer algorithms to implement health and safety measures automatically, such as mask detection to ensure compliance in public areas, people proximity for safe physical distancing, occupancy monitoring through people counting and object classification, and thermal camera integrations to read facial temperatures