[英]How can I label velodyen points in Kitti dataset point per point?
I am trying to train my neutral network using velodyen points in kitti dataset. 我正在尝试使用kitti数据集中的velodyen点训练我的中性网络。 The dataset has around 450 frame as bin file .
数据集有大约450帧作为bin文件。 Each frame has around 120k row of x,y,z,re as shown below:
每帧都有大约12万行x,y,z,re,如下所示:
However, I want to add another column that has the labels for each point . 但是,我想添加另一列,其中每个点都有标签。
Any thought how I can handle this ? 有没有想到我该如何处理?
Thanks in advance 提前致谢
The KITTI Vision Benchmark Suite does not provide per-point labels so if you want that information you would have to manually annotate it or find other dataset to work with (like: http://www.semantic3d.net/ ). KITTI Vision Benchmark Suite不提供每点标签,因此,如果您想要该信息,则必须手动对其进行注释或查找其他可使用的数据集(例如: http : //www.semantic3d.net/ )。
I guess that you could use the 3D bounding box information ( http://www.cvlibs.net/datasets/kitti/eval_object.php?obj_benchmark=3d ) to label all the points inside the bounding box with the corresponding class, but you will still have missing per-point labels for the rest of the scene. 我想您可以使用3D边界框信息( http://www.cvlibs.net/datasets/kitti/eval_object.php?obj_benchmark=3d )来用相应的类标记边界框内的所有点,但是您场景的其余部分仍将缺少按点标签。 So maybe kitti is not the right dataset for what you are trying to do (use per-point labels)
因此,kitti可能不是您想要做的事情的正确数据集(使用每点标签)
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.