Hi,

I am testing the code on ROS Melodic. Everything runs fine except that the markers are not published on the rivz. Only the rtab map is being built with yolo running in parallel. Do I need to change something in the config file?

Thanks

How can I run this using a stereo camera (ZED2), and without turtlebot?
Im trying to modify RTAB-Map to read from a stereo camera, but obj_positioner fails to position the detected objects on the map.
I have modified the nodes to read from ZED stereo camera, but still it cant work with RTAB-map.
Can you please provide more details on using different RGB-D cameras other than Kinect?
Thanks

Hello,

I am trying to build this repo using catkin build in Ubuntu 20.04 & ROS -Noetic.
All other packages build fine except for the 'Auto' package.

The error seems to be related to pcl-conversions/ pcl-ros but I am unable to resolve it.

I have attached the entire error message.
PCL-conversions' error.odt

Any help would be greatly appreciated!

First of all, great work and thanks for open source it! I tried to run the pipeline but kept having problem with rtabmap_ros. I am using Ubuntu20.04 + ROS Noetic and installed rtabmap_ros followed their github repo. Still getting this error but not sure why, any tip would be very helpful!
COMMAND: roslaunch auto obj_positioner.launch
... logging to /home/xuefei/.ros/log/d1775444-0b96-11ee-8d47-954732257d68/roslaunch-xuefei-ThinkPad-X1-Extreme-77201.log
Checking log directory for disk usage. This may take a while.
Press Ctrl-C to interrupt
Done checking log file disk usage. Usage is <1GB.

started roslaunch server http://xuefei-ThinkPad-X1-Extreme:40353/

SUMMARY

PARAMETERS

• /object_positioner/boxes_topic: /darknet_ros/boun...
• /object_positioner/camera_cx: 472.65
• /object_positioner/camera_cy: 268.7
• /object_positioner/camera_frame: /kinect2_rgb_opti...
• /object_positioner/camera_fx: 537.0
• /object_positioner/camera_fy: 536.95
• /object_positioner/classes: ['door', 'bench',...
• /object_positioner/detection_flag_topic: /darknet_ros/flag...
• /object_positioner/global_frame: /map
• /object_positioner/max_proj_distance: 8.0
• /object_positioner/pointcloud_topic: /camera/depth_reg...
• /object_positioner/projection_method: 1
• /object_positioner/quiet_mode: False
• /object_positioner/robot_frame: /base_link
• /rosdistro: noetic
• /rosversion: 1.16.0

NODES
/
filter (auto/filter.py)
object_positioner (auto/object_positioner)

auto-starting new master
process[master]: started with pid [77231]
ROS_MASTER_URI=http://localhost:11311

setting /run_id to d1775444-0b96-11ee-8d47-954732257d68
process[rosout-1]: started with pid [77264]
started core service [/rosout]
process[object_positioner-2]: started with pid [77271]
process[filter-3]: started with pid [77272]
[ INFO] [1686845289.008300536]:
CAMERA FRAME: /kinect2_rgb_optical_frame
[ INFO] [1686845289.008956735]:
MAX PROJECTION DISTANCE: 8.000000
Traceback (most recent call last):
File "/home/xuefei/catkin_ws/src/3d-object-semantic-mapping/auto/src/filter.py", line 14, in
from rtabmap_ros.msg import MapGraph
ModuleNotFoundError: No module named 'rtabmap_ros'

For running this code, how can I record my own .bag data to replace the sequence3_kinect.bag on my own robot base?
Is the procedure like this：

1. turn on my robot base，and publish the /odom topic
2. turn on laser sensor, and publish the /scan topic
3. turn on kinect sensor, and publish the /kinect2/qhd/image_color_rect/compressed and /kinect2/qhd/image_depth_rect/compressed topic
4. Start recording .bag data

Can you please provide more details?
Very thanks.

The links provided are not providing anything to download.
Also how can I isolate the run of placing doors on map, where I already have a map? Can you please provide a more elaborate description on nodes and their publisher subscriber only for object placement on maps.

Hi Verlab,

I am new to ROS. I have follow the step by step that you provided. However, on rviz, only tf odom appear. 2D map not generated. I am using ubuntu 18.04, ros Melodic, cuda 10.2.

Hi,
I find your research very interesting. I have some question about the implementation.
Did you guys process everything on your laptop ?
So you recorded videos using your platform and then used that video to run rtabmap and yolo on your pc ? Did i understand correctly ?
and you guys used two differents cameras: Microsoft Kinect and Orbbec Astra

thank you for your work !