IRMVLab / SNI-SLAM

[CVPR'24] SNI-SLAM: Semantic Neural Implicit SLAM
100 stars 1 forks source link

No mesh files in output folder #5

Closed Jayyy0311 closed 1 week ago

Jayyy0311 commented 2 months ago

Hello,

I'm reaching out for assistance as I've encountered an issue while running python3.7 -W ignore run.py configs/Replica/room1.yaml After the tracking progress bar indicates completion at 100%, the terminal seems to freeze without any additional output. I've checked the output folder and found that the mesh folder and tracking_vis folder is empty. 2024-09-10 14-31-29屏幕截图

lizhiteng111 commented 2 months ago

Hello, I also encountered this issue. May I ask how you resolved it?

Jayyy0311 commented 2 months ago

Hello, I also encountered this issue. May I ask how you resolved it?

Yes, I set the meshing resolution parameter to 0.05, and after running it and waiting for a while, the mesh file appeared as expected. But the tracking_vis folder is still empty. And I've noticed that the command below is not working either. python visualizer.py configs/Replica/room1.yaml --output output/Replica/room1 --top_view --save_rendering

zstsandy commented 2 months ago

mesh saving time is relatively long, requiring approximately 10 minutes of waiting. Visualization is saved in mapping_vis folder. Before run python visualizer.py, you need to set mesh_freq to 40 (or larger) and save mesh at specific frequency while running the code.

Andrewuqing commented 2 months ago

网片保存时间相对较长,大约需要等待 10 分钟。可视化效果保存在 mapping_vis 文件夹中。在 run 之前,您需要设置为 40(或更大),并在运行代码时以特定频率保存网格。python visualizer.py``mesh_freq

Hello, this command has no effect. How should I modify it python visualizer.py configs/Replica/room1.yaml --output output/Replica/room1 --top_view --save_rendering

zstsandy commented 1 month ago

you can set mesh_freq to 40 or larger in https://github.com/IRMVLab/SNI-SLAM/blob/3860456545d3cf7b2358dd3c056ef8d5a2623caf/configs/SNI-SLAM.yaml#L50 and rerun python -W ignore run.py configs/Replica/room1.yaml, which may take quite a long time to save mesh during mapping process. Then, run python visualizer.py for visualization.