NODAR has released its stereo vision algorithms as standalone, licensable software products, allowing original equipment manufacturers and developers to deploy long-range 3D perception technology without being tied to specific camera hardware or integrated development kits.
The move marks the first time NODAR’s patented stereo vision technology has been made available as an independent software offering. The new NODAR software development kit (SDK) enables customers to integrate the company’s perception algorithms into existing compute platforms and custom camera setups across a range of autonomous and automation applications.
See also: STEER Tech and KINEXON Integrate Autonomy Software With RTLS for Industrial Vehicles
The SDK includes two main components. Hammerhead, NODAR’s core stereo vision engine, generates dense, long-range 3D point clouds with centimetre-level accuracy by processing stereo image pairs. An optional module, GridDetect, builds on this depth data to provide 3D object detection, size estimation and motion tracking for use cases such as advanced driver assistance systems, collision avoidance, robotics and industrial monitoring.
NODAR said the software is designed to operate reliably in challenging environments, including dust, fog, rain, glare and low-light conditions. The platform supports deployment on x86 and ARM systems equipped with NVIDIA GPUs and integrates through common interfaces such as ROS2 and ZeroMQ.
See also: PlusAI Partners With Japan’s T2 to Deploy Level 4 Autonomous Trucks
By decoupling its perception technology from dedicated hardware, the company said system designers can optimise cost and performance while reusing existing camera and computing investments. NODAR said the approach is intended to support applications beyond automotive autonomy, including agriculture, heavy equipment, rail, marine, aviation and industrial inspection.
