Sunrise 3.0M-U(X3M-U) Datasheet

2022-05-24
■General Product Description:
●The Sunrise 3.0M-U (X3M-U) is a highly-integrated, high-performance and power efficient artificial intelligence System-on-Chip (SoC) powered by Horizon Robotics. Based on the dual-core BPU with Bernoulli-architecture, X3M-U supports real-time pixel-level video segmentation, structured video analysis and attention-based vision perception. Video streams (max. resolution up to 8Mpixel@30fps) inputs from MIPI CSI-2 Receivers (RX) or DVP are processed by the image signal processing (ISP) and video processing units before being passed to BPU for artificial intelligence inference computing such as video object detection, sentiment segmentation, scene parsing, etc.
●The X3M-U SoC also integrates a quad-core Cortex A53 CPU, an image signal processing (ISP) unit for wide/high dynamic range (WDR or HDR) camera input images, video signal processing units supporting noise reduction (3DNR), lens distortion correction (LDC), rotation and scaling, H.265/H.264/MJPEG video codecs, a DDR controller for DDR4/LPDDR4/LPDDR4X, one USB 3.0 host/device dual-role module, a Gigabit Ethernet MAC with RMII/RGMII interface, video bypassing or display output interface, two I2S ports for audio or speech, other integrated on-chip modules and rich off-chip sensors or connectivity peripheral interfaces, which enable greater flexibility and faster delivery for customers to implement their own products for different kinds of targeted applications.
●The X3M-U supports advanced perception capabilities and features that enable artificial intelligent Internet-of-Things (AIoT) functions, through precise detection, recognition and understanding of traffic participants and monitoring environments, including vehicles, pedestrians, human count, human shopping behavior statistics in store, face detection/recognition, car license plate recognition, etc.
●Except of video/image artificial intelligence detections, segmentation and recognitions, the X3M-U BPU can also support intelligent speech algorithms like keyword spotting, speech recognition, text-to-speech and so on to make AIoT device more complete.
●The advanced toolkit from Horizon Robotics provides a Linux-based training framework and development environment to assist customers in rapid application implementation and deployment. The BPU toolkit also supports floating point neural network model to fixed point neural network model conversion.

Horizon Robotics

X3M-UX3ME00IBGTMB-U

More

Part#

System-on-Chip (SoC)

More

AIoT applications ]Smart IP Camera ]Human-Machine Interface devices ]lnteractive Robots ]

More

Datasheet

More

More

Please see the document for details

More

More

FCBGA483;BGA;FCBGA

English Chinese Chinese and English Japanese

2022-05-10

Rev. 1.0

10.1 MB

- The full preview is over. If you want to read the whole 160 page document,please Sign in/Register -
  • +1 Like
  • Add to Favorites

Recommend

All reproduced articles on this site are for the purpose of conveying more information and clearly indicate the source. If media or individuals who do not want to be reproduced can contact us, which will be deleted.

Contact Us

Email: