Deciphering the coordinate values in a LiDAR pcd file

Vignesh_Sundar · July 13, 2020, 9:53pm

I am trying to work on the aspect of removing the ground reflection and detection in a pcd file of a LIDAR data. How will I get the values of the x, y, z 3D points in the data and the intensity values in that particular set be obtained ? Basically I am trying to get the lowest z values possible in the data set and get the intensity of that and subtract the value from all the other ground points. Is this the right way to do it ?

holger-motional · July 17, 2020, 1:54am

Take a look at how we load lidar data:

github.com

nutonomy/nuscenes-devkit/blob/master/python-sdk/nuscenes/nuscenes.py#L550


:param pointsensor_token: Lidar/radar sample_data token.
:param camera_token: Camera sample_data token.
:param min_dist: Distance from the camera below which points are discarded.
:param render_intensity: Whether to render lidar intensity instead of point depth.
:return (pointcloud <np.float: 2, n)>, coloring <np.float: n>, image <Image>).
"""
cam = self.nusc.get('sample_data', camera_token)
pointsensor = self.nusc.get('sample_data', pointsensor_token)
pcl_path = osp.join(self.nusc.dataroot, pointsensor['filename'])
if pointsensor['sensor_modality'] == 'lidar':
    pc = LidarPointCloud.from_file(pcl_path)
else:
    pc = RadarPointCloud.from_file(pcl_path)
im = Image.open(osp.join(self.nusc.dataroot, cam['filename']))


# Points live in the point sensor frame. So they need to be transformed via global to the image plane.
# First step: transform the point-cloud to the ego vehicle frame for the timestamp of the sweep.
cs_record = self.nusc.get('calibrated_sensor', pointsensor['calibrated_sensor_token'])
pc.rotate(Quaternion(cs_record['rotation']).rotation_matrix)
pc.translate(np.array(cs_record['translation']))

Then you can simply access the matrix as pc.points. To get the lowest z values you can do np.min(pc.points[2, :]).

Vignesh_Sundar · July 24, 2020, 11:28pm

Thanks a lot. I shall try the approach.