GUI is blank - Githubissues

introlab / rtabmap

RTAB-Map library and standalone application

https://introlab.github.io/rtabmap

Other

2.84k stars 786 forks source link

GUI is blank #1304

Open alphabet-lgtm opened 5 months ago

alphabet-lgtm commented 5 months ago

GUI is blank but may still work. I have tried [RTABMap-0.21.4-win64-cuda117.7z] and [RTABMap-0.21.4-win64-cuda117.exe] but got the same problem.

matlabbe commented 5 months ago

Can I assume the first example was the version without cuda from this link? https://github.com/introlab/rtabmap/releases/tag/0.21.4

I cannot see that problem on my Windows computer (Windows 11 Home). Which Windows version? Is that a computer with a dedicated or integrated GPU (though it should not matter)?

The only way to reproduce a similar issue is to close all panels from the main window: though the menu bar should still be there to reopen all panels (under Window menu).

alphabet-lgtm commented 5 months ago

Can I assume the first example was the version without cuda from this link? https://github.com/introlab/rtabmap/releases/tag/0.21.4

I cannot see that problem on my Windows computer (Windows 11 Home). Which Windows version? Is that a computer with a dedicated or integrated GPU (though it should not matter)?

The only way to reproduce a similar issue is to close all panels from the main window: though the menu bar should still be there to reopen all panels (under Window menu).

The version with cuda from the link(https://github.com/introlab/rtabmap/releases/tag/0.21.4).The version without cuda can work properly, but I want to use it with cuda. I used win10 with cuda11.7. Does this warning affect?

matlabbe commented 5 months ago

For the PATH error, it can be ignored if you use just the app (not using rtabmap as a library). I cannot reproduce the issue with the cuda version. You may try older versions like 0.21.0: https://github.com/introlab/rtabmap/releases

CUDA version won't improve the general speed of VSLAM. Use it only if you really need a feature that is provided only with CUDA build.

alphabet-lgtm commented 5 months ago

For the PATH error, it can be ignored if you use just the app (not using rtabmap as a library). I cannot reproduce the issue with the cuda version. You may try older versions like 0.21.0: https://github.com/introlab/rtabmap/releases

CUDA version won't improve the general speed of VSLAM. Use it only if you really need a feature that is provided only with CUDA build.

I have tried v0.21.0 and it just stuck and flashback. Next, I tried v0.20.16 and it also stuck like this.

I want to improve the speed of Loop closure dection by cuda because it worked so slowly(few seconds one frame). Do you have any suggestions?