Just fill in the form and we will contact you shortly
On-premise license request
Just fill in the form and we will contact you shortly
3DiVi Edge AI Hardware edge computing equipment
It is intended for neural network processing of video streams directly at the places where they are received. It eliminates the delay associated with the transmission of information, increases the speed of decision-making, and reduces network bandwidth requirements.
The computing unit is designed to build vision systems on vehicles and UAVs, as well as on any stationary geographically distributed objects. The unit contains the neural processor with a performance of up to 32 TOPS and video processor with support for parallel processing of up to 32 HD video channels. The unit is made in a compact aluminum housing with passive cooling and IP65 protection, contains all the necessary interfaces for building embedded systems. During 2024, it will be added to the register of Russian industrial products (according to RF PP 719).
Specifications;3DV-EdgeAI-1684x
CPU;8 ARM Cortex-A53 cores, frequency up to 2.3GHz
NPU;32TOPS (INT8), 16TFLOPS (FP16/BF16), 2TFLOPS (FP32)
VPU;32 channels H.265/H.264 1080p@25fps video decoding<br> 12 channels H.265/H.264 1080p@25fps video encoding
RAM;16GB LPDDR4x
Storage devices;64GB eMMC (OS)<br> up to 512GB SD (optional)<br> M.2 NVMe 2280 (optional)
Basic interfaces;1GB Ethernet<br> USB3.0<br> RS232, RS484, CAN
Wireless interfaces (optional);Wi-Fi IEEE 802.11 ia/b/g/n/ac, Bluetooth 5.x module (M.2 key A+E)<br> 3G/LTE/4G module (M.2 key B)
Navigation;GLONASS/GPS module
Supply voltage;DC 9…27 V
Power consumption;up to 30 watts (excluding the consumption of external devices via USB)
Dimensions;215х218х68 mm
Weight;2,3 kg
Temperature range;-20...+60 ОС (-40...+85 ОС is possible)
Protection class;IP65
Software;The Linux operating system, 3DiVi developer tools, ready-made detectors (face and object detection, face verification and identification, face tracking, face attribute recognition, image quality assessment algorithm (QAA), skeletal models of human poses, activity type recognition, ADAS and DSM components), the ability to create detectors and models on the customer's specification.
Areas of application:
Control of drivers of public and freight transport
Unmanned systems
Unmanned systems
Smart city (control of human and automobile traffic)
Smart city (control of intersections)
Smart city (control of law and order in public places)