Skip to content

Latest commit

 

History

History
153 lines (122 loc) · 6.84 KB

mapora_readme.md

File metadata and controls

153 lines (122 loc) · 6.84 KB
Raw

Mapora

Mapora is an open-source georeferenced point cloud generating tool. Basic idea is putting each individual point cloud onto the pose which are matched with time in nanosecond precision.

One of the important point for exporting tightly coupled point clouds with Mapora is having high accurate GNSS/INS positions. To provide that, it is highly recommended to use a post-processing software for GNSS/INS positions.

The other important issue is having precise LiDAR-IMU calibration angles. If not, the possibility of having bad features in map is very high.

Some of the views of the point clouds can be seen below.

mapora_building mapora_cars mapora_environment mapora_map

Data Collection Procedure

For data collection, a Velodyne VLP16 and an Applanix POS LVX is needed. These 2 sensors needed to be connected to each other with PPS cable. In that way, the needed time information is provided into the VLP16.

Basically, the data can be collected with a Mobile Mapping System (MMS). A sensor setup is needed for this procedure. With the system, fixed position sensors are provided. A computer is able to collect PCAP for point cloud data comes from Velodyne VLP16 and GNSS/INS positions comes from Applanix POS LVX via ethernet. Additionally, from the web user interface of the Applanix POS LVX GNSS should be collected for post-processing the GNSS/INS positions to get better accuracy. The example data collection is made with the car in below images.

mapora_setup1 mapora_setup2

System Requirements

  • PC which runs Ubuntu 22.04 (Ubuntu version must be at least 20.04)

    • Mapora is tested on Ubuntu 22.04 with ROS2 Humble.

  • Velodyne VLP16 LiDAR

    • One of the needed data are PCAPs which contain Velodyne VLP16 ethernet packets. These packets consist of data packets and position packages. Position packages is needed to transmit to the LiDAR with PPS (Pulse Per Second) and GPRMC messages. In that way, the time information that is needed for program provided in PCAPs.

  • Applanix POS LVX GNSS/INS

    • The other needed data is GNSS position .txt data. These can be got in any way. Any software or library that can extract GNSS/INS positions in the read format of Mapora can be used.
    • One of the preferable way to get te GNSS/INS position .txt data is written below.
      • Collection the GNSS/INS data with RTK (Real Time Kinematics).
      • Post-Processing of the collected RTK GNSS/INS data for getting a better position accuracy.
      • Exporting the .txt files of GNSS/INS positions as Mapora can use. (An example of the format is provided in examples directory.)

Dependencies

Installation & Build

Install some dependencies with apt-get:

sudo apt-get install libgeotiff-dev libboost-all-dev libeigen3-dev libpcap-dev

Install ROS2 dependencies:

sudo apt install ros-humble-point-cloud-msg-wrapper ros-humble-tf2-ros ros-humble-tf2

Create the ROS2 workspace:

mkdir -p ~/projects/mapora_ws/src
cd ~/projects/mapora_ws/src

Clone the repository:

git clone https://github.com/leo-drive/mapora_dev.git
cd mapora_dev/

Run build.sh script to install thirdparty libraries:

sudo ./scripts/build.sh

Build the ROS2 environment:

colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=RelWithDebInfo -DCMAKE_EXPORT_COMPILE_COMMANDS=1

Input Files

Input Description
PCAP Files PCAP files are needed for reading the point clouds.
Pose .txt Pose txt file is needed for reading the positions for knowing the point cloud origin positions.

Output Files

Output Description
Point Clouds Georeferenced .las formatted point cloud files.

Run the Mapora

To running the Mapora, PCAP files collected with Velodyne VLP16 and GNSS poses taken from Applanix PosLVX are needed. PCAP files can be seperated or can be only one piece. If it is one piece, then program may fail depending on your RAM size. So, it is highly recommended to divide the PCAP files. It can be made with below command via Wireshark API.

editcap -c 1000 <input.pcap> <output.pcap>

It takes input PCAP and exports the seperated PCAP with file indexes like output_{index}_{timestamp}.pcap.

After filling the params/mapora_params.yaml, the program can be executed with:

ros2 launch mapora mapora_launch.py

Example Data

To see how Mapora works, an example data directory is provided. This link contains the example data of the Mapora. The example data needs to be in the example_data/ directory in the Mapora package. In params/mapora_params.yaml file, the params are set to run the example data. You can observe the file paths in this file. After you put all data files into right directory which is example_data/, .las exports of the Mapora would be in the example_data/output_las/ directory. You can change the path files in the params file as you wish.

ros2 launch mapora mapora_launch.py

Params

Param Description
project_namespace What is the name of this project for the point clouds that will be exported.
pcap_dir_path Input directory of PCAPs.
pose_txt_path Input file of pose txt.
las_export_directory Export directory of the point clouds.
max_point_distance_from_lidar What is the maximum distance of LiDAR points for mapping?
min_point_distance_from_lidar What is the minimum distance of LiDAR points for mapping?
r_x LiDAR-IMU calibration roll angle (angle around X-axis).
r_y LiDAR-IMU calibration pitch angle (angle around y-axis).
r_z LiDAR-IMU calibration yaw angle (angle around z-axis).