NavLive, a UK-based construction technology firm, has selected NVIDIA’s Jetson Orin NX system-on-module to power its latest handheld scanner, enabling advanced real-time processing in a compact, portable device.
The scanner, designed for on-site use, integrates five sensors, including three cameras, LiDAR, and an inertial measurement unit, all streaming data simultaneously. Previous models required multiple microcontrollers to process sensor inputs, but the adoption of Jetson Orin NX consolidates these functions onto a single platform.
The Jetson Orin NX module, which combines CPU, GPU, and dedicated deep learning accelerators, allows the scanner to run all processing on-device, eliminating dependence on Cloud computing. Lower power consumption extends battery life to around one hour, while high-bandwidth interfaces support multiple cameras and sensors.
NavLive collaborated with Connect Tech to adopt the Boson carrier board, which accommodates three cameras – a configuration not supported by other off-the-shelf boards. The scanner uses the platform to perform simultaneous localisation and mapping (SLAM) and object detection for building features such as windows and doors. It also has potential for future features, including anonymisation tools to support regulatory compliance.
NavLive engineers have developed a custom camera driver that leverages NVIDIA’s software ecosystem, including Linux for Tegra, Vision Programming Interface, and the Multimedia API. Heavy computer vision tasks, such as debayering and image undistortion, are handled by specialised hardware, freeing up CPU and GPU resources for other functions.
David Wisth, Chief Technology Officer and Co-founder at NavLive, said the move reduced the device’s weight and power requirements while enhancing performance. “Customers immediately notice that the device is easier to hold, yet it delivers much more in terms of capability,” he said.
