Sensor setup guide - koide3/glim GitHub Wiki

Sensor setup guide

Recommended sensors

Livox MID360 (for most use cases)

  • Equipped with IMU.
  • A very wide field of view that ensures robust estimation.
  • Applicable to both indoor and outdoor scenarios.

Microsoft Azure Kinect, Orbbec Femto Bolt (for indoor mapping)

  • Equipped with IMU.
  • High quality point cloud measurements.
  • Best for mapping in narrow indoor scenes (Note: Cannot be used in outdoor scenes).

Setup guide

See also Important parameters.

All sensors

  • Recommended point cloud FPS: 10 ~ 20 FPS
  • Recommended IMU FPS: > 100 and points_fps * 10 FPS
  • Change config_ros.json: glim_ros/(imu|points)_topic (image_topic is not necessary).
  • Change config_sensors.json: sensors/T_lidar_imu

Livox MID360

  • Change glim_ros/acc_scale to 9.80665 (Livox drivers uses [g] instead of [m/s^2]).

Microsoft Azure Kinect (Orbbec Femto Bolt)

  • Reduce point cloud FPS and resolution (fps=15 and depth_mode=WFOV_2X2BINNED) to reduce the bandwidth pressure.
  • Set sensors/global_shutter_lidar=true to disable motion distortion correction.

Velodyne VLP16

  • Set preprocess.json: preprocess/k_correspondences=20.

Step-by-Step Tutorial with MID360