Digua Robot Unveils Open Source X-Lens for Real-Time Depth Estimation with 40 Million Parameters
- Published
- Source
- leaderobot.com
Digua Robot has introduced the X-Lens depth estimation model, which features 40 million parameters, significantly fewer than traditional models. This innovative solution can process mixed inputs from both fisheye and pinhole cameras, producing dense depth maps with real-world scale. On the Digua Xuri S600 chip, it achieves over 20 FPS with six-camera input and 41 FPS with a single fisheye camera.
The importance of X-Lens lies in its ability to provide absolute depth measurements, unlike many existing models that only offer relative depth. This capability is crucial for robotic applications, enabling precise interactions with objects in the environment. By translating each pixel into a light ray direction, X-Lens overcomes the challenges posed by camera distortion, allowing for accurate depth estimation across different camera types.
Looking ahead, the adaptability of X-Lens to various camera inputs without requiring extensive model adjustments is noteworthy. This flexibility could lead to broader applications in robotics and automation. No further timeline was disclosed at the time of publication.
Original report
Open Source X-Lens by Digua Robot: 40 Million Parameters for Real-Time Depth Estimation
leaderobot.com · LeaderobotThis briefing is an independently written summary based on publicly available reporting and is provided for industry information and news discovery. The original report and source publication are credited and linked where applicable. RobotToday does not claim ownership of third-party source material.
Rights concerns? Contact [email protected] with the relevant URL and details. We will review the matter and take appropriate action where warranted.
More related news
- 1
- 2
-
3
ROBOTTODAY Weekly October 19 – 23, 2026 RobotToday · 1 week ago
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
Related Suppliers
ImmerVision
Canada
ImmerVision develops wide-angle and panomorph lens technology and image processing software for 360-degree cameras used in automotive, security, and robotics.
ArgosVision
Korea, Republic of (South)
Korean CISS spinoff whose ArgosVue fisheye-stereo sensors give robots a 200 x 140 degree depth view, pitched as a cheaper LiDAR alternative.
Videre Design
United States
US stereo vision camera maker; STH camera head & Small Vision System for robotics depth perception; used by NASA, MIT, RoboCup teams.
Luxonis
United States
US maker of OAK spatial AI cameras and DepthAI SDK, combining stereo depth sensing with on-device neural inference for robotics and vision applications.
ACE Robotics 大晓机器人
China
Shanghai embodied-AI startup founded 2025, chaired by a SenseTime co-founder; builds the open Kairos world model and the A1 robot brain module.
Fireworks AI
United States
Redwood City AI infra startup (2022) serving open-source LLMs like Kimi K3 via API, used by Cursor, Vercel and Notion.
MagicLab 魔法原子机器人技术（无锡）有限公司
China
Wuxi-based Chinese startup founded January 2024 building the MagicBot humanoid (42 DOF, 5-hr runtime) for multi-robot collaborative industrial manufacturing.
Xense Robotics (Qianjue Robotics) 千觉机器人
China
Shanghai startup founded in 2024 by an SJTU tactile-sensing researcher, selling optical tactile sensors and the XTac UMI G1 data-collection gripper.
Goyo Optical Inc 御用光学株式会社
Japan
Asaka, Saitama, Japan lens maker founded 1976; produces industrial machine-vision and CCTV lenses from 2.8mm to 1100mm focal length.
Stereolabs
United States
San Francisco stereo-vision company founded in 2010; became a wholly owned Ouster subsidiary in February 2026 after shipping 90,000-plus ZED cameras.
DreamVu
United States
Philadelphia firm founded 2017 that captures real venues with its own single-shot 360-degree stereo camera and sells world-model and robot training data.
micropsi industries GmbH
Germany
Berlin AI vision software firm; MIRAI system lets industrial robots learn from human demos to handle variance—no CAD, no camera calibration required.
Robotics needs a service framework.
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
Join the RobotToday community on LinkedIn
Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.
