Our state of the art computer vision algorithms run directly as hardware-circuit in the FPGA. With more than five years experience on realtime computer vision we are able to accelerate resource-intense algorithms like dense stereo or dense optical flow to real-time speeds. Our special-purpose hardware processes the depth and flow maps as fast as the image sensor reads out the pixels.
Up to 10 Cameras
Our hardware platform is capable to read out up to 10 image sensors in parallel. All sensors are synchronised in hardware.
Each camera also contains an inertial measurement unit (IMU), making it optimally suited for algorithms fusing inertial and vision data, like visual inertial odometry (VIO).