`sensor_msgs_py.point_cloud2` functions treat numpy arrays as having shape (width, height) about common_interfaces HOT 3 OPEN

Yes, the fix is simple enough so I'll go ahead with a PR. It will probably be sometime next week.

from common_interfaces.

So I originally thought this was just a relatively harmless swapping of axes, but now I'm seeing it's creating issues with the point order in organized clouds. Specifically, I'm trying to read point clouds with the following structure:

PointCloud2:
  height:  1544
  width:  2064
  fields:
     sensor_msgs.msg.PointField(name='x', offset=0, datatype=7, count=1)
     sensor_msgs.msg.PointField(name='y', offset=4, datatype=7, count=1)
     sensor_msgs.msg.PointField(name='z', offset=8, datatype=7, count=1)
     sensor_msgs.msg.PointField(name='normal_x', offset=16, datatype=7, count=1)
     sensor_msgs.msg.PointField(name='normal_y', offset=20, datatype=7, count=1)
     sensor_msgs.msg.PointField(name='normal_z', offset=24, datatype=7, count=1)
     sensor_msgs.msg.PointField(name='curvature', offset=32, datatype=7, count=1)
  is_bigend:  False
  is_dense:  True
  point_step:  48
  row_step:  99072
  num_bytes:  152967168

When I call read_points, I get an array of shape (2064,1544) as expected. However, when I try to index the points in a 2D region, I find they are out of order. Switch the argument order in the call to reshape inside read_points fixes this and lets me access a contiguous region of the organized cloud.

from common_interfaces.

I agree that it seems transposed. Would you be able to open a PR with the fix and maybe a unit test so that we can prevent regression in the future?

from common_interfaces.