generating mesh throws error: Segmentation fault (core dump)

[HsiaoTsan]

When I am trying to use ORK Capture, and the inverted dot pattern to scan my own red Coke can in real world(not using the mesh file provided by tutorial), everything goes well until I run: rosrun object_recognition_reconstruction mesh_object --all --visualize --commit Then I get a very short one line error message: Segmentation fault (core dumped) If I ignore this error and continue to run training command, no error is raised:

 rosrun object_recognition_core training \
> -c `rospack find object_recognition_tod`/conf/training.ork \
> --visualize
Training 1 objects.
computing object_id: a42af78e71dc2de7b14fa138f6015685
[ INFO:0] Initialize OpenCL runtime...

And then if I continue to run the detecting command, rosrun object_recognition_core detection -c `rospack find object_recognition_tabletop`/conf/detection.table.ros.ork --visualize , I get a message:

......
[ INFO] [1540954146.810071742]: publishing to topic:/table_array
No models found for method "mesh"
[ INFO] [1540954146.873466802]: publishing to topic:/recognized_object_array

then I run rostopic echo /recognized_object_array, I get:

header: 
  seq: 184
  stamp: 
    secs: 1540955036
    nsecs: 346691541
  frame_id: "camera_rgb_optical_frame"
objects: []
cooccurrence: []

the objects array is empty, which means no object is found when I point my Kinect camera to the red Coke can.

I can successfully run the tutorial Object Recognition Using Tabletop to detect a physical red Coke can, using the mesh file provided by the tutorial.

Here's the workflow I am using:

1) roslaunch openni_launch openni.launch
2) rosrun object_recognition_capture capture --seg_z_min 0.01 -o silk.bag --preview
3) rosrun object_recognition_capture capture --seg_z_min 0.01 -o silk.bag
4) rosrun object_recognition_capture upload -i silk.bag -n 'Silk' milk soy silk --commit

5) rosrun object_recognition_reconstruction mesh_object --all --visualize --commit # error raises

6) rosrun object_recognition_core training \
-c `rospack find object_recognition_tod`/conf/training.ork \
--visualize
7) rosrun dynamic_reconfigure dynparam set /camera/driver depth_registration True
8) rosrun topic_tools relay /camera/depth_registered/image_raw /camera/depth_registered/image
9) rosrun object_recognition_core detection -c `rospack find object_recognition_tabletop`/conf/detection.table.ros.ork --visualize

Thanks in advance!

[Sivasankar-Ganesan]

Hi How did you solve the error? I am also getting this error.

[HsiaoTsan]

Hi How did you solve the error? I am also getting this error.

I didn't solve this error. If possible, try a larger CPU memory machine. Mine is 8GB.