For anyone building robotics or spatial perception projects on Linux, Banana Pi has a new all-in-one option worth a look. The BPI-OM7 integrates the company's low-profile RK3588-based single-board computer with an ORBBEC Gemini 2 3D depth camera, all mounted on a tripod and ready to run edge AI workloads out of the box. The Rockchip RK3588 brings eight CPU cores (four Cortex-A76 at 2.4 GHz, four Cortex-A55 at 1.8 GHz), a Mali-G610 MP4 GPU, and a 6 TOPS NPU for on-device inference.

The depth camera side uses ORBBEC's MX6600 depth engine with active stereo IR sensing at 850nm, covering a range of 0.15 to 10 meters (0.5 to 32.8 feet) with depth resolution up to 1280x80 at 30fps and a wide 91-degree horizontal field of view. A 1080p RGB camera and built-in 6-axis IMU round out the sensor package, with hardware-level depth-to-color alignment handled on-chip. The SBC itself ships with 8GB LPDDR4x and 64GB eMMC by default, with options up to 32GB RAM and 128GB storage. Connectivity includes dual 2.5GbE ports, Wi-Fi 6, Bluetooth 5.2, HDMI 2.1 output up to 8K60, and an M.2 slot for NVMe storage. A 40-pin Raspberry Pi-compatible header is included for expansion.

On the software side, Banana Pi is preparing an Ubuntu 24.04 image with a Docker-based setup for the Orbbec SDK and the RKNN toolkit to leverage the NPU. The getting started guide walks through running YOLO5 object detection and multi-stream depth demos, and a separate point cloud reconstruction toolkit on GitHub provides a C++ pipeline for capture, denoising, registration, and visualization. Users who prefer the Armbian ecosystem have an additional option, as Armbian builds for the underlying BPI-M7 SBC are available and current as of early 2026. The Orbbec SDK itself has been fully open-source since version 2.0.0 (released in late 2024), and ORBBEC publishes an OrbbecSDK ROS2 wrapper with support for ROS2 Foxy, Humble, and Jazzy on ARM64 Linux, making the Gemini 2 a practical sensor for robotics stacks that are already ROS2-native. The BPI-OM7 is available now at $740 (EUR 680).