...

Ant Group Subsidiary Robbyant Launches LingBot-Depth, a Spatial Perception AI Model

LingBot-Depth

Robbyant recently announced the official open-source release of its LingBot-Depth spatial perception model. This high-precision tool helps robots understand 3D environments more effectively. It specifically targets challenges found in complex real-world settings. This subsidiary of Ant Group aims to enhance how machines perceive depth. The LingBot-Depth model focuses on improving the accuracy of 3D environmental mapping for various robotic platforms.

LingBot-Depth Technology and Performance Highlights

The new technology is a solution to problems with reflective and transparent surfaces. Conventional sensors have difficulty detecting glass or polished metal surfaces. LingBot-Depth employs Masked Depth Modeling to recover lost visual information. It examines texture and object contours to fill in gaps. This process creates much sharper 3D depth maps. It allows robots to navigate safely without changing their existing hardware.

Robbyant also formed a strategic partnership with Orbbec for this launch. They will integrate the model into next-generation depth cameras. This collaboration uses chip-level raw data to boost performance. The LingBot-Depth model achieved a 70% reduction in relative error during indoor testing. It also improved task success rates on major industry benchmarks.

“Robbyant’s work in spatial intelligence models and algorithms complements Orbbec’s expertise in 3D vision chips and robotic vision systems,” said Len Zhong, Head of Product Management of Orbbec.“In this collaboration, the chip-level depth data provided by the Gemini 330 delivers a stable, high-fidelity, and physically grounded data foundation for the LingBot-Depth model.”

The company plans to open-source a massive dataset soon. This includes 2 million high-quality RGB-depth pairs. These samples cover extreme and ambiguous environmental conditions. Such data will help the global robotics community innovate faster. Robbyant continues to focus on “hand-eye-brain” integration for future robotic companions.

Explore IT Tech News for the latest advancements in Information Technology & insightful updates from industry experts!

News Source: Businesswire.com