search

wg-perception / reconstruction

3d reconstruction of objects from data in the DB
15 stars 17 forks source link

mesh and train problem #7

Open mcqrobot opened 8 years ago

mcqrobot commented 8 years ago

Hi: thanks for your reply for my last problem :https://github.com/wg-perception/reconstruction/issues/ and I modified the files arrcoding to your advice and then I continue to test. As I just a ROS freshman and just want to use your this great project. So when I finished modifying the files above, I test according the quick guide: http://wg-perception.github.io/object_recognition_core/quickguide.html#quickguide What I want to train and detect is the milk cartons. The capture and upload steps are successful. However, when I go on this mesh command: rosrun object_recognition_reconstruction mesh_object --all --visualize --commit. Still it throw some errors:

＊＊＊＊＊ERROR1:＊＊＊＊＊＊＊＊＊＊

 Traceback (most recent call last):
 File "/home/mcq/ws/src/ork_reconstruction/apps/mesh_object", line 190, in <module>
 simple_mesh_session(dbs, session, args)
 File "/home/mcq/ws/src/ork_reconstruction/apps/mesh_object", line 86, in simple_mesh_session
 raise RuntimeError("There are no observations available.")
 RuntimeError: There are no observations available.

I have to ignore this step and go on. The next step is trainning:rosrun object_recognition_core training -c rospack find object_recognition_tod/conf/training.ork --visualize,the output is just like: Training 8 objects. computing object_id: 508fefeabfa7a9aa1cc43b6c61008d82 Deleting the previous model da138258498efa9588050a10d500520f ..... After this, I keep going go to next step:detect. and I have launched the rviz and tpyed this command: rosrun object_recognition_core detection -c rospack find object_recognition_tod/conf detection.ros.ork --visualize expectedly, rviz which subsrible the "marker" topic can't see anything but worsely the teminal throw the many error such as: ＊＊＊＊＊＊＊＊ERROR2:＊＊＊＊＊＊＊＊＊＊＊

[WallTime: 1451702238.674595] Exception in your execute callback:            exception_type   CellException
[cell_name] = source1
[cell_type] = ecto::py::BlackBox
[function_name] = process_with_only_these_inputs
[type] = std::runtime_error
．．．．．

　Can you tell me how to solve these two problems .I really want to use your this work. Thx

saspelund commented 8 years ago

I am also having this first issue. I know that I am in the "correct" directory as the .bag files are located there. Would it be simple enough to see where exactly the program is searching for the files?

I tried tinkering around with the --db_collection option to no success.

saspelund commented 8 years ago

I was able to deal with this issue by making note of the session_ID displayed at the end of the upload sequence (I had to copy it manually from the output because I can't find it anywhere else) and then mesh each item one at a time using that session_ID.