Deep learning enhanced 3D sensor with powerful performance and continuously improving accuracy.
Integrated platform to collect and analyze a wide range of indoor or outdoor customer behavior data.
Robust feature set supporting your multiple applications and system integration.
TD-2000 embeds data capturing, stereo vision analytics and deep learning technology into one single device, which represents state-of-the-art technology and next generation video sensor for traffic intelligence purpose. Compared to other 3D sensors, its patented ‘DeepCount’ technology based on deep learning renders it superior in accuracy as it incessantly improves its count quality over time with the sensor being deployed over time.
Brick-and-mortar businesses and organizations can use the sensor for applications such as people counting and tracking, passing-by and dwell statistics, zone occupancy and hot spots visualization, distinguishing between adults and children, optimizing queue operation, etc.
Besides, the traffic intelligence data of the sensor with open interface can be easily integrated with your existing systems including ERP, CRM, Business Intelligence and the 3rd party applications such as mobile applications, digital signage etc.
Stereo vision analytics plus patented ‘DeepCount’ renders TD-2000 superior in accuracy rate up to 98% and above under typical applications.
(Up to 45°) Tilt installation enables TD-2000 to be applicable for complex scenarios like counting passing-by traffic flow.
Multi-sensor Joint Capability
TD-2000 can incorporate data from up to 9 sensors to cover a large area for traffic counting.
Multi-zone Traffic Counting
One single sensor can set up to 16 discrete zones for data collection, making it possible to serve multiple purposes.
With strong adaptability to the environment and low requirement on illumination, TD-2000 performs impeccably in both indoor and outdoor environments.
All-in-one platform embedding data capturing, deep learning and stereo vision analytics
Improving accuracy over time in actual scenarios based on patented “DeepCount”
Low requirement on lighting (≥ 2 Lux)
Distinguishing between adults and children
Filtering non-traffic objects like shopping carts, mannequins, shadows, reflections, etc.
Support of tilt mounting, flexible mounting brackets for unobtrusive installation
Supports scheduled streaming of digital video output remotely for count accuracy validation
Web portal for local and remote management
Open API for easy system integration
Up to 16 discrete zones for metric collection
Low bandwidth requirement by transmitting data through Ethernet
Internal storage stores configuration settings and at least 90 days of data
Data packaged in XML packets and delivered via HTTP/HTTPs, FTP/FTPs, or daily e-mail
Power over Ethernet or DC
*Up to 450. Tilt installation may degrade counting accuracy.