Closed IsaacTay closed 9 months ago
Hi, what are you running exactly? Binary, dataset, configuration file? Local changes? Can you share your configuration and how you installed PLVS?
I'm currently running PLVS using RosDocker on noetic. Almost all configuration is the default, with a local opencv version of 4.8.0. I also have libpcl-conversions-dev
installed to get Volumetric Mapping running.
I'm running a modified version of ./ros_xterm_d435, with Camera.type=pinhole
added to get it to work.
CUDA version is 11.4 with CuDNN 8
Hi, can you please share the YAML configuration file you are using? Did you test PLVS with or without lines, with our without dense mapping?
Run on the shell: ulimit -c unlimited
Then once a core is generated from the crashed app
you can run
gdb <application path> <core path>
(Ref: https://www.brendangregg.com/blog/2016-08-09/gdb-example-ncurses.html)
Could you please share the backtrace that gdb is reporting?
I found this issue may be related to the following ones depending on ORBSLAM3 https://github.com/UZ-SLAMLab/ORB_SLAM https://github.com/UZ-SLAMLab/ORB_SLAM3/issues/608 https://github.com/DavidPetkovsek/MORB_SLAM/issues/17
Pushed a new commit related to this issue. @IsaacTay let me know.
I can't replicate it reliably, seems to happen at random after a period of time.
Error message: