search

NVlabs / FB-BEV

Official PyTorch implementation of FB-BEV & FB-OCC - Forward-backward view transformation for vision-centric autonomous driving perception
Other
608 stars 40 forks source link

vis_occupancy results problem #24

Open Chris91gp opened 8 months ago

Chris91gp commented 8 months ago

Thanks for great work，how to use tools/ vis_occupancy.py to show the result? it looks like has bug. I don't know where the problem is. the inputs shape is 200 200 18

image

KarlLi5 commented 6 months ago

Hello, I encountered the same issue as you. May I ask if you have successfully resolved it?

pisquilo commented 6 months ago

Hi, I also have a similar issue. Did anyone manage to solve this?

WMCh commented 5 months ago

@Chris91gp @KarlLi5 @pisquilo Hello, the voxels with the prediction label equal to 17 (free) can be filtered out when drawing. Add the following codes

fov_voxels = fov_voxels[fov_voxels[:, 3] != 17]

before https://github.com/NVlabs/FB-BEV/blob/832bd81866823a913a4c69552e1ca61ae34ac211/tools/analysis_tools/vis_occupancy.py#L187

gaopeng91 commented 1 month ago

thanks, I already tried, it works. question closed