drawzeropoint/hope

A ROS package for extracting horizontal planes given point cloud input. Proposed in the paper: Dong, Zhipeng, et al. "Hope: Horizontal plane extractor for cluttered 3d scenes." Sensors 18.10 (2018): 3214. (https://www.mdpi.com/343288)

hope's Issues

How to reproduce the same kind of results as stated in the illustrations ?

According to the repository, the objects belonging to different horizontal planes are colored with different colors and there is a clear differentiation between them like it is shown below :

I am trying to reproduce the same kind of results with a scene point cloud like this one below :

However, in my HoPE visualization I don't see such kind of clear differentiation in the colors (please see the image below) like that of the first illustration.

Can you please let me know how to get the same kind of results ?

How can we get the ros node working without a robot URDF ?

I do not have a URDF model with all the link relationship like base_link, vision_depth_optical_frame, etc. I am trying to use tf's static_transform_publisher to mimic the presence of a base_link. But the ros node would not detect point clouds. I did make sure that I changed the topic name of my point cloud in this line of the code.

The message I get continuously from the ros node is this :

[ros.hope]: Transform: Frame base_link does not exist.
[ros.hope]: Transform: Frame base_link does not exist.
[ros.hope]: Transform: Frame base_link does not exist.

But this is not true. I am sending static transform like this :

arun@arun:~/catkin_ws$ rosrun tf static_transform_publisher 0.0 0.0 0.0 0.0 0.0 1.0 base_link vision_depth_optical_frame 100
[ros.tf]: Sending transform from base_link with parent vision_depth_optical_frame
[ros.tf]: Sending transform from base_link with parent vision_depth_optical_frame

I also tried it the other way. It didn't work neither :

arun@arun:~/catkin_ws$ rosrun tf static_transform_publisher 0.0 0.0 0.0 0.0 0.0 0.0 1.0 vision_depth_optical_frame base_link 100
[ros.tf]: Sending transform from vision_depth_optical_frame with parent base_link
[ros.tf]: Sending transform from vision_depth_optical_frame with parent base_link
[ros.tf]: Sending transform from vision_depth_optical_frame with parent base_link

Code does not compile after recent updates

When I simply do a git clone of this repository after the recent updates related to dynamic reconfigure, the code does not compile. The code seems to be failing since the last 22 commits. On doing a thorough analysis, this commit (357246c) is found to be the most stable one. But it lacks important improvements. The later commits seem to have new feature additions/ improvements, but they do not compile.

I have attached a detailed report of the error messages from my catkin workspace.

error_log.txt

The publisher is not publishing on topic /plane_points

The publishers in general seem to have a problem. They are not publishing anything on their topics. The diagnostic message is below. This is the state on commit a10a145

arun@arun:~$ rostopic echo /plane_points 
WARNING: no messages received and simulated time is active.
Is /clock being published?

memory leak in ubuntu20.04

rt, in lib/plane_segment.cpp，memory leak at entering the loop for the second time.(after line 315 in my pic, running command in your guide with rgb and depth image in sample folder)