Highly integrated, plug and play

The SeerSense® DS80 is the industry's only intelligent sensing module integrating a binocular depth engine, a time-of-flight (TOF) engine, a visually scalable local area mapping (VSLAM) engine, an AI engine, and a video encoding engine. Its design leverages the strengths of passive binocular vision and time-of-flight (TOF) depth perception, achieving complementary capabilities. The localized VSLAM engine, binocular depth and time-of-flight (TOF) depth calculation engines, and AI inference engine eliminate system computing power dependence, enabling plug-and-play operation. The feature-rich VSLAM SDK can handle complex operational scenarios. The CNN engine, based on an independent hardware unit, adapts to a variety of local video input sources, including fisheye cameras, time-of-flight (TOF) cameras, and RGB color cameras, enabling local real-time AI inference and computation, and supports the deployment of multiple models for parallel processing. Combined with its spatial perception capabilities, it supports scene recognition and spatial semantic segmentation, making it an ideal AIOT terminal device. Hardware-based H.265 video compression significantly reduces transmission bandwidth. Rich image processing capabilities and comprehensive multi-camera intrinsic and extrinsic calibration support a wide range of visual sensing applications. Compatible with the Open CV and Open VINO open source communities, this toolchain allows users to quickly deploy offline trained models locally on the module. Hundreds of models from the Open Model Zoo can be leveraged to address a wide range of application scenarios, significantly shortening development cycles. Connecting to a host device via a USB Type-C port, expansion ports allow for integration with other sensors such as LiDAR and multi-machine cascading, providing high scalability.

The SDK supports Windows, Linux, and Android operating systems, and provides various wrapper APIs for application development corresponding to different products, such as ROS/ROS2, Unity, C/C++, Python, and Java JNI.

This product has passed CE, FCC, RoHS certifications, as well as CB and FDA Class I laser safety certifications, and is suitable for various product markets at home and abroad, including industrial, medical, security, education, office, entertainment, and home use.

Product Specifications

Spatial computing, multi-sensory fusion

The SeerSense® DS80 integrates three core spatial perception capabilities: VSLAM, depth perception, and AI reasoning. It enables device-level localized deployment, significantly lowering the barrier to entry for user development and use.

1) Visual SLAM (VSLAM) is a vision-based SLAM (Simultaneous Localization and Mapping) system. It uses image sensors to acquire spatial feature information, generating incremental maps while simultaneously determining its own position and posture. This multi-view geometric computing application places high demands on algorithms and computing power. It is a foundational technology in robotics and AR/VR, and has consistently attracted significant industry attention. In 2018, Xvisio Technology pioneered the industry's integration of its long-developed algorithms with high-performance chips, launching the eXLAM80, a VSLAM sensor module based on the Intel Movidius VPU. The module received a 2018 CES Innovation Award. Subsequently, Intel Realsense launched a similar product, the Realsense T265, in 2019, becoming the only two device-level VSLAM solutions on the market and ushering in the VSLAM device 1.0 era.

The SeerSense® DS80, developed by Xvisio Technology, is an upgraded version of the eXLAM80. It utilizes the Intel Movidius Myriad X VPU, significantly improving the VSLAM processing capabilities of the previous generation. In addition to high-speed, high-precision standalone VIO mode, it also supports multiple operating modes, including SLAM, CSLAM, end-to-end cloud deployment, and real-time loopback detection. It also adds high-speed QR code-assisted positioning and spatial anchoring capabilities, fully leveraging the advantages of multi-sensor fusion.

2) Depth perception is a fundamental capability for spatial perception. In addition to its basic binocular vision depth engine, the SeerSense® DS80 is equipped with an industry-leading Sony VGA-resolution iTOF depth sensor. TOF depth detection is based on the time-of-flight of photons, resulting in higher ranging accuracy than binocular vision depth engines. The SeerSense® DS80 implements module-side depth resolution and filtering. Operating at a 940nm wavelength, it effectively suppresses interference from near-infrared wavelengths, achieving a frame rate of up to 30fps and an error of less than 1% within a range of 0.2-4 meters. It can output depth maps, spatial 3D point clouds, and IR grayscale images. Single-frequency or dual-frequency modes are available, depending on the operating range. Single-frequency mode has an effective operating distance of 1.5 meters, while dual-frequency mode reaches up to 4 meters. It can be used for depth detection, 3D object measurement, and 3D environment modeling.

For outdoor environments with strong light and interference from multiple cameras, the binocular depth engine in the SeerSense® DS80 can be used. This engine features high resolution, a wide field of view, and a high frame rate. It offers an effective horizontal field of view of up to 110°, a frame rate of up to 60 fps, a resolution of 1280 x 720, and a working distance of up to 5.5 meters. Within the effective operating range, the error is less than 3%, though its accuracy is slightly lower than that of a Time-of-Flight (TOF) sensor. The following is a comparison of passive binocular ranging performance between the SeerSense® DS80 and the Intel RealSense D435 (the two products have different fields of view, with the SeerSense® DS having a wider field of view).

Both depth engines can work simultaneously with the VSLAM engine, enabling real-time detection of object depth while tracking and positioning in space. This effectively enables 3D modeling, navigation, positioning, and obstacle avoidance.

3) The localization of the AI engine effectively leverages the local capabilities of multiple camera inputs and, in combination with the VSLAM and depth engine, deploys a variety of intelligent functions. These include 3D gesture recognition and tracking, human skeleton extraction, 3D face anti-counterfeiting recognition, 3D object recognition and spatial segmentation, and scene recognition. Users can use the OpenVino offline model training tool to train and import scene-specific models, enabling plug-and-play convenience on the SeerSense® DS80.

Series of products, full coverage

SeerSense® DS80 also has two simplified versions: SeerSense® DR80L for facial recognition applications and SeerSense® S80 for pure location tracking.

Product Series

Competitive Comparison

Empowering industries and widely used

Seer Sense® DS80 can be widely used in a wide range of fields, including industry, security, healthcare, logistics, retail, office, film and television, and is the basic sensing unit for Metaverse and AIOT applications.

Understanding and reconstructing real space is the first step toward the metaverse. Since its inception, Xvisio Technology has adhered to the integration of algorithms and hardware, building a comprehensive development path from underlying sensors to product platforms, and developing core technologies and product deployment capabilities across the entire supply chain. The Xvisio Technology team shares the shared vision of "innovating machine vision perception capabilities and surpassing the limits of human vision." As a high-tech company leading the global spatial perception and interactive technology, Xvisio Technology continues to forge ahead and will soon release a series of new products. Please follow the "Xvisio Technology" official account for the latest news.

Please contact us for product details and pricing.

www.xvisiotech.com | contact@xvisiotech.com

Xvisio Technology is a high-tech startup focused on core XR spatial perception and interaction technologies (high-speed VSLAM technology) and XR product solutions. Founded in Silicon Valley in 2016 and based in Shanghai, Xvisio currently has R&D centers in Silicon Valley and Europe. It boasts a distinguished domestic and international R&D team and over 20 core technology IPs covering algorithms, hardware, software, systems, and applications. Its key products include the SeerSense® perception module and the SeerLens® series of XR glass, as well as core technology licensing and customized product development services. Its clients include many industry-leading companies, and its products are sold throughout Japan, the United States, Europe, and other regions.