FLOAM_ssl is a LIdar depth point cloud acquisition and location project based on ROS+RealSense L515.

📦 original project repo: github.com/wh200720041…

I understand the project structure

  • launch/: Run the startup file, which is configured herenodeandparam, which controls the node services and super parameters started at runtime

Switch between official lab video data or RealSense L515 real-time data. The first two lines of the launch file decide which part of the data to use

<! -- Lab video data -->
<node pkg="rosbag" type="play" name="rosbag_play" args="--clock $(env HOME)/Downloads/L515_test.bag"/> 
<param name="/use_sim_time" value="true" />

<! -- RealSense L515 real-time data collection -->
<include file="$(find realsense2_camera)/launch/rs_camera.launch">
  <arg name="color_width" value="1280" />
  <arg name="color_height" value="720" />
  <arg name="filters" value="pointcloud" />
</include>
Copy the code
  • rviz/: Used to determine the layout of the page and the adjustable parameter adjustment section
  • src/ – laserMapping – laserProcessing – octoMappingHow do octrees display on THE ROS visualization platform –octoEstimation
  • CMakeLists: Compile output file decision.launchand.cppThe nodes in the

Integrate point cloud triangulation to reconstruct the original project

How to modify and run

After modifying the SRC/octomAppingNode. CPP file, switch to catkin_wk/ and run the FLOAM_ssl/run

Integration point cloud triangulation code

Put FLOAM_ssl/cloud_visualization. CPP in the SRC/octomAppingNode.cpp file and recompile and run it

Basic steps

  1. Convert the original PointXYZRGB format to PointXYZ (temporarily removing the color data)
  2. Normal Estimation
  3. Concat cloud-data with normal data
  4. Surface Reconstruction
  5. Visualization of reconstructed results | Visualization

Existing problems

The current code is equivalent to using the acquired data of the first frame for surface reconstruction once the radar is opened, but the problem is that the surface reconstruction visualization window conflicts with the original FLOAM window. After the surface reconstruction visualization window is closed, the original FLOAM window and the surface reconstruction visualization window will refresh the data of the next frame

The above problems have not yet begun to solve