Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 

RoboND

3D Segmentation in ROS on RGB-D data

This project performs 3D segmentation on data obtained from a RGB-D camera.

To run this project, copy "sensor_stick" directory to "src" directory of your catkin workspace (typically "catkin_ws"), then run catkin_make, source devel/setup.bash, then execute "roslaunch sensor_stick robot_spawn.launch".

The original project dataset looks like this: image

In the first step, RGB-D point cloud is downsampled using Voxel grid filter.

Then a passthrough filter is used to focus only on area above table (table desk included)

RANSAC is used to perform segmentation around table desktop plane.

The table is extracted as positive segments:

image

The objects are extracted as negative segments:

image

In the end, Euclidean clustering utilizing Kd-tree is used to partition object mesh into individual 3D objects and visualized in RViz: image