Rtabmap + ZED2i TF Error

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Rtabmap + ZED2i TF Error

rishikandimall
This post was updated on .
Hi, I am trying to connect ZED to Rtabmap for vslam, but keep getting this error even after trying many variations of settings. Does someone know why this might be happening and what the error means?

[ERROR] (2026-02-25 19:12:19.447) MsgConversion.cpp:2182::convertRGBDMsgs() TF of received image for camera 0 at time 1772064739.025150s is not set!

I am running:
ros2 launch rtabmap_launch rtabmap.launch.py \
   rtabmap_args:="–delete_db_on_start Rtabmap/TFBufferSize 500" \
   rgb_topic:=/zed2i/zed_node/rgb/color/rect/image \
   depth_topic:=/zed2i/zed_node/depth/depth_registered \
   camera_info_topic:=/zed2i/zed_node/rgb/color/rect/camera_info \
   depth_camera_info_topic:=/zed2i/zed_node/depth/camera_info \
   imu_topic:=/zed2i/zed_node/imu/data \
   odom_topic:=odom \
   frame_id:=zed2i_camera_link \
   approx_sync:=true \
   wait_imu_to_init:=false \
   wait_for_transform_duration:=0.02 \
   always_check_imu_tf:=false \
   rgbd_odometry:=false \
   approx_sync_max_interval:=0.02

ros2 launch zed_wrapper zed_camera.launch.py \
   camera_model:=zed2i \
   camera_name:=zed2i \
   node_name:=zed_node \
   publish_tf:=false \
   publish_map_tf:=false \
   publish_imu_tf:=true \
   imu_enabled:=true \
   depth_mode:=DEPTH_MODE_PERFORMANCE \
   enable_depth:=true \
   pos_tracking_enabled:=true \
   use_sim_time:=false
Reply | Threaded
Open this post in threaded view
|

Re: Rtabmap + ZED2i TF Error

matlabbe
Administrator
I could not reproduce the TF error.

Here what I am trying (tested with a zedm and using zed odometry)
ros2 launch zed_wrapper zed_camera.launch.py \
   camera_model:=zedm  \
   camera_name:=zed2i \
   node_name:=zed_node \
   publish_tf:=true  \
   publish_map_tf:=false \
   publish_imu_tf:=true \
   imu_enabled:=true \
   depth_mode:=DEPTH_MODE_PERFORMANCE \
   enable_depth:=true \
   pos_tracking_enabled:=true \
   use_sim_time:=false

ros2 launch rtabmap_launch rtabmap.launch.py \
   rtabmap_args:="--delete_db_on_start Rtabmap/TFBufferSize 500" \
   rgb_topic:=/zed2i/zed_node/rgb/color/rect/image \
   depth_topic:=/zed2i/zed_node/depth/depth_registered \
   camera_info_topic:=/zed2i/zed_node/rgb/color/rect/camera_info \
   depth_camera_info_topic:=/zed2i/zed_node/depth/camera_info \
   imu_topic:=/zed2i/zed_node/imu/data  \
   odom_topic:=/zed2i/zed_node/odom \
   frame_id:=zed2i_camera_link \
   approx_sync:=true \
   rgbd_sync:=true \
   approx_rgbd_sync:=false \
   wait_imu_to_init:=false \
   always_check_imu_tf:=false \
   visual_odometry:=false
Without zed odometry, it is also working.
ros2 launch zed_wrapper zed_camera.launch.py \
   camera_model:=zedm  \
   camera_name:=zed2i \
   node_name:=zed_node \
   publish_tf:=false  \
   publish_map_tf:=false \
   publish_imu_tf:=true \
   imu_enabled:=true \
   depth_mode:=DEPTH_MODE_PERFORMANCE \
   enable_depth:=true \
   pos_tracking_enabled:=false \
   use_sim_time:=false

ros2 launch rtabmap_launch rtabmap.launch.py \
   rtabmap_args:="--delete_db_on_start" \
   rgb_topic:=/zed2i/zed_node/rgb/color/rect/image \
   depth_topic:=/zed2i/zed_node/depth/depth_registered \
   camera_info_topic:=/zed2i/zed_node/rgb/color/rect/camera_info \
   depth_camera_info_topic:=/zed2i/zed_node/depth/camera_info \
   imu_topic:=/zed2i/zed_node/imu/data \
   frame_id:=zed2i_camera_link \
   approx_sync:=false \
   wait_imu_to_init:=true \
   always_check_imu_tf:=false

I also tried latest from source and current humble binaries. Also using latest zed ros2 package. What is your system and which rtabmap/zed versions are you using? Also are you using zed with nitros? You can disable nitros to debug here.

I get however these kind of warnings (using latest rtabmap_ros):
[WARN] [1773610929.524029124] [rtabmap.rgbd_odometry]: Dropping image/scan data with stamp 1773610929.475528 (delay=0.048479). Something is wrong because the clock difference with the previous topic received (0.001596s) is much lower than the expected one (0.100247s) estimated from the topic stamps (previous stamp=1773610929.375281). If you are processing a large bag with flaky replaying delay, consider setting parameter "always_process_most_recent_frame:=false" to avoid aggressively dropping data.
Could be related to my computer/USB cable, here rqt_runtime_monitor with only zed wrapper running (no subscription, no rtabmap), it is dropping more than 30% of frames! That would be the reason why I get many warnings alike above.