Sprinx - co.exist :: people & vehicle mobility analysis
Exploiting the powerful neural network, co.exist is the Sprinx solution dedicated to multimodal mobility analysis in the city, logistics, transport and public sectors.
How it works
Sprinx as an Independent Software Vendor (ISV) is focused on developing software platforms that can be installed, both server-side and at-the-edge, on market hardware and belonging to different brands.
co.exist combines Artificial Intelligence algorithms, a user-friendly interface and Data Intelligence to monitor people and vehicle mobility by generating alarms and statistical data without installing additional equipment. The software platform, equipped with its front-end, can be easily integrated into more complex system architectures, public and private, through the activation of standard communication protocols or custom ones.
In the case of server-side solutions, the co.exist is entirely agnostic to the camera brand, requiring a standard Onvif-RTSP stream for analysis. The co.exist can be installed on a standard PC platform, using just the CPUs, thanks to the AI inference engine based on the Intel® OpenVINO™ toolkit, or Nvidia GPUs, according to the system requirements.
As for the at-the-edge solutions, Sprinx has been undertaking a verification activity for some time now aimed at selecting and certifying devices capable of guaranteeing performance levels according to Sprinx standards. As a result, the list of compatible CCTV camera brands and models is constantly updated; for this reason, it is suggested to request the latest updated list in force.
Activating professional algorithms directly onboard the camera represents an extremely interesting and cost-effective solution in distributed architectures, such as urban ones. Moreover, onboard analytics allow not worrying about the network infrastructure.
Using for the analysis of the Raw image coming directly from the processor also allows to keep the camera performance constant, not even decreasing the flows and profiles available on the camera.
The use of neural networks trained for behaviour analysis applications, integrated with a 3D object tracking approach, allows for being particularly adaptive in pre-existing CCTV systems ensuring good performance even in unsuitable installation and environmental conditions.
Additionally, the co.exist web user interface displays a dashboard and an event journal for tuning and maintenance purposes. Mobility data and alarms can be forwarded to third-party systems using standard protocols and displayed on the Sprinx centralized management platform, dragon.
Key Features
- Field proven AI
More than 15.000 cameras around the world turned into intelligent sensors by Sprinx software since 2009. - Hardware Agnostic
Compatible with any IP camera or video encoder that supports RTSP video streaming protocol. This scheme facilitates system maintainability while enabling limitless evolution. - Mobility insight
Centralized upper-level user experience empowered by Dragon. - Plug&Play
Fast calibration and easy configuration, ready to interact with third-party CCTV & mobility platform. - Adaptive & cost effective
Decision-making tool for a smarter mobility, compatible with even existing CCTV cameras. - At-the-edge plugin
Installable directly onboard the Open Platform of video cameras from major CCTV manufacturers on the market. - CPU only
The AI inference engine based on Intel OpenVINO does not require any additional hardware or GPU.
Stopped vehicle
Vehicle classification
Slowdown & Queue
Pedestrians
Wrong way
Crowd
Traffic Data Collection
- Counting of people, bicycles, and vehicles in transit
- Origin Destination matrix at intersections/roundabouts
- Classification of vehicles in transit (4 classes: motorcycle, car, van/commercial vehicle, truck/bus)
- Average speed of vehicles divided by vehicle class (km/h per class)
- Traffic density, calculated on the basis of the integration of multiple data