Rusu and Eiten,

A bag file with 100 messages (topic: /tilt_cloud_shadowfiltered and /tf) and
the launch file is attached; 

I found that a single tilt_cloud_shadowfiltered message to ground_removal
node contains point cloud one full scan of laser, i.e., ranging from min_ang
to max_ang. So if these points are captured when the laser scans the ground,
they are located  along a line. So I's suspecting that this is the reason
why each time the fitted plane is not a horizontal plane. And I am also
guessing if two or more lines of laser scan are combined before fed to
ground_removal node, the fitted plane should be a horizontal one. So in the
source file sac_inc_ground_removal_standalone.cpp, I added the above
assemble procedure before executing ground_removal; unfortunately the result
is not satisfactory - part of the ground is still left in the result. 

best regards
xiaojun 


http://ros-users.122217.n3.nabble.com/file/n1784882/cloudin_tf.bag
cloudin_tf.bag 


http://ros-users.122217.n3.nabble.com/file/n1784882/ground_plane.xml
ground_plane.xml 
-- 
View this message in context: http://ros-users.122217.n3.nabble.com/Navigation-stack-issue-about-ground-plane-detection-tp1707477p1784882.html
Sent from the ROS-Users mailing list archive at Nabble.com.