Skip to main content

Stereo Depth Estimation

The D-Robotics stereo depth estimation algorithm takes stereo image data as input and outputs the disparity map and depth map corresponding to the left view. The algorithm draws inspiration from the IGEV network and adopts a GRU architecture, offering good data generalization and high inference efficiency.

IP Address Explanation

Hardware Model Description

  • Stereo Camera Model: GS130WI
  • Image Sensor Model: SC132GS

Model Versions

The model versions are listed below. The default version used is 2.4_int8.

PlatformAlgorithm VersionQuantization MethodInput SizeMax Frame Rate (fps)Model Description
RDK X5V2.4_int16int16640x352x3x215Current main version, high-precision depth estimation
RDK X5V2.4_int8int8640x352x3x223Current main version, high frame rate depth estimation

Functional Architecture Diagram

Functional Architecture Diagram

Startup Steps

  1. Use the following command to confirm the camera's I2C signal is normal.

    i2cdetect -r -y 4
    i2cdetect -r -y 6

    Execution Result Image

  2. Use the following command to run the camera. Access [IP]:8000 via a web page to view the camera image and depth estimation results.

    cd /userdata/magicbox
    source /opt/tros/humble/setup.bash
    source app/ros_ws/install/setup.bash
    bash app/ros_ws/src/magicbox/hobot_stereonet/script/run_stereo.sh --mipi_rotation 0.0 --stereonet_version v2.4_int8 --camera_info_topic /image_combine_raw/right/camera_info

    Execution Result Image

Reference Materials