search

PRBonn / lidar-bonnetal

Semantic and Instance Segmentation of LiDAR point clouds for autonomous driving
http://semantic-kitti.org
MIT License
959 stars 206 forks source link

ROS interface for DarkNet53Seg #23

Closed kame-hameha closed 5 years ago

kame-hameha commented 5 years ago

Hi, I would like to use the segmented point clouds in ROS1...

I just ran your train and inference scripts with the pretrained DarkNet53 and am wondering if you already released the ROS interface? I read that you would do it after the ICRA deadline. J

Cheers, Kai

Hi,

Not a shot in the dark at all.

I thought of providing this functionality as I do in bonnetal with an infer_video.py and infer_img.py, but unfortunately, images are a lot more standard than point clouds to deal with.

So, I quickly realized that to make an infer_scan.py script I would have to specify a format, and if we decided on our own format, then our inference script was already good enough.

I could provide a script, for example, for .ply, .stl, or .pcd files, or something standard like that, but since I am going to release a ROS interface soon after ICRA deadline I let this side project go for now.

So, long story short, putting the data in the kitti format is the easiest way for now.

Originally posted by @tano297 in https://github.com/PRBonn/lidar-bonnetal/issues/3#issuecomment-522912374

kame-hameha commented 5 years ago

Update: I am now able to use the model and do the inference within ROS. Moreover, I get the predictions and will now put them back into the original pointcloud. Cheers, Kai PS: Great work with DarkNet53Seg and the Semantic Segmentation of point clouds using range images!

akouri-dd commented 4 years ago

@kame-hameha are you going to contribute your ROS wrapper back to the project?

akouri-dd commented 4 years ago

I am going to consume PointCloud2s and run inference on them, and then need to store predictions as well. If this code is already written it would save me some time. If not, I will write it and then contribute back. Let me know.