3D Vision Camera

High Definition 3D Stereo Camera with a high performance
on-board graphics unit and a depth sense capture

Depth sense - visual ability to perceive the world in 3D

A very compact camera that can measure distances out to 50m and works not only indoors but outdoors, the S4SVision™ camera is a strong candidate for applications such as large-scale architectural scanning and obstacle detection. Stereo cameras are passive devices and work by comparing two images taken by cameras several inches apart. The software then looks at the distance in pixels between similar features in each image and uses that to estimate the depth or distance from the camera to objects.

Real-time vision processing

The system is designed with the capability to run algorithms intrinsically and process data for real-time tasks: object and pedestrian recognition, movement detection and tracking, environment modelling

High Performance On-Board Processing

The core processor is built around the revolutionary NVIDIA Tegra® K1 SoC and uses the same NVIDIA Kepler™ computing core designed into supercomputers around the world. This gives you a fully functional platform and APIs for quickly developing and deploying compute-intensive systems for computer vision, robotics, medicine, and more.

Optimised for Robotics

Advanced visual localisation, obstacle detection
and path planning for autonomous vehicle

3D Point Cloud

S4SVision™ provides industry standard Point Cloud data for robotics applications. Working in an unstructured or dynamic environment the point cloud data can be used to detect objects and their positions. In the field of service robotics 3D point clouds are mostly used for creating maps of household environments and for reliable object grasping and manipulation. An efficient tool for 3D point cloud processing comes in a form of c++ library called PCL (Point Cloud Library). The system registers and incorporates point clouds of an object and removes scanning errors with two methods for registering: feature based registration and Iterative Closest Point algorithm (ICP).



8mm, f/2.2 (38mm equivalent)

Sensor type

1/1.8" @ 1.3M per sensor

Stereo Baseline





720p (20 FPS) and 480p


350x175x50mm and 900g


800mA, 12V


NVIDIA Kepler "GK20a" GPU with 192 SM3.2 CUDA cores (up to 326 GFLOPS)


NVIDIA "4-Plus-1" 2.32GHz ARM quad-code Cortex


2 GB DDR3L 933MHz




USB 3.0, USB 2.0 and HDMI

OS Requirements

Linux Ubuntu 1.4 or MS Windows 7 (x32/64)

Office in London, UK
Office in Kaunas, LT
Office in London, UK
Office in Kaunas, LT

Office in Kaunas, LT

Donelaičio g. 62, Kaunas, 50000


Office in London, UK

45 King William Street, London, EC4R 9AN