Lane_Extractor

A Project to extract a lane from the pointcloud map and save it back as a pcd file.

Point Cloud -> extract -> Vector_lane

Execution screen

👆 <0sec ~ 30sec>

👆 <30sec ~ 60sec>

Preparations

Available in environments with [PCL & ROS kinetic] installed.

http://www.pointclouds.org/downloads/linux.html [PCL download]
http://wiki.ros.org/kinetic/Installation/Debian [ROS kinetic download]

Map processing

$ rosrun lane_extractor intensity_cloud_saver [Original pcd file path] [Intensity Processing pcd file name] [min_intensity]  

You can set the desired intensity coefficients.
If you put 450 in [min_intensity], create a pcd file, leaving only the point with an intensity value of 450 or more.
Search speed is extremely fast when processing applying.

Visual debugging

Run rviz and open the conforming file stored in the rviz folder in the Lane_extractor package.

Run

1. run lane_extractor_node

There are two ways to run the Lane Extractor.

First,

$ rosrun lane_extractor lane_extractor_node  

Running without giving a argument behind it will be executed by loading the pcd file in the path set to Default in the code.

Second,

$ rosrun lane_extractor lane_extractor_nod [pcd file path]  

When you pass the path of a pcd file to a factor, you use a pcd file that is passed to a factor.

2. run rosbag file

$ rosbag play [ndt_pose path] or  
$ rosbag play -r [speed rate]  [ndt_pose path] 

When you run the command, the line_extractor node receives the message ndt_pose and saves it to the queue.

3. run /find_lane rosservice to get lane point

User can set a value.

$ rosservice call /find_lane "rad: [value] minIntensity: [value] maxIntensity: [value]" 

If you have run all of the robag files, use the ros service to have the lane_extractor node find lane.

4. run /save_lane rosservice to get pcd files

$ rosservice call /save_lane "num : [int]" 실행

Save the lane points found so far to a file. (File saved in the location of the lane_extractor node.)

Contact

If you interested in my project. Send me an email.

Thanks!