Guangjian Technology Launches Embodied AI Visual Perception Solution to Equip Robots with Vision

Bot Telegraph, July 20 — Guangjian Technology recently announced the official release of its embodied AI visual perception solution. Centered on AI binocular vision technology, the system integrates intelligent perception algorithms with a highly efficient computing architecture to equip robots with visual capabilities spanning core scenarios such as environmental perception, spatial understanding, navigation and positioning, and precise manipulation.

In real-world environments, complex surfaces like transparent glass, highly reflective metals, and low-texture walls impose stringent requirements on the completeness of depth information. Powered by AI binocular vision, Guangjian Technology’s solution maintains a robust depth recovery capability when confronted with these challenging materials, achieving stable reconstruction of millimeter-level details.

Guangjian Technology has introduced the Libra series of AI binocular vision modules. Among them, the Libra 1000 features a 2-centimeter baseline design tailored for wrist-mounted manipulation perception, making it ideal for scenarios such as robotic arm grasping and dexterous hand operations. The Libra 3000 utilizes a 7-centimeter baseline designed for head-level spatial perception, targeting applications like medium-to-long-range obstacle avoidance, SLAM positioning, and spatial mapping. Both products support high-speed GMSL output and are compatible with mainstream robotic platforms.

Source: Bot Telegraph China