• This project visualizes the results of robot and vehicle position and pose estimation and lidar observations using Google Earth. This is very useful for evaluating and demonstrating the results of localization and SLAM.

Examples

Comparison of two types of position estimations

example_compare_two_positions.m

https://github.com/user-attachments/assets/44f42e9e-c169-410e-93d6-784791bb7219

Visualization of vehicle driving and lidar data

example_vehicle_lidar_animation.m

https://github.com/user-attachments/assets/7cf90a3f-3ca4-4d13-b62e-40bf3530ce5c

https://github.com/user-attachments/assets/cecb54b5-2ec8-4ae7-809b-ce55b559873b

Drone flight and visualization of lidar data

example_drone_lidar_animation.m

https://github.com/user-attachments/assets/cd34e387-e564-4b97-98db-b750061653ff

Requirements

• MATLAB (>R2022a)
  • Image Processing Toolbox and Computer Vision Toolbox are required for example_vehicle_lidar.m and example_vehicle_lidar_animation.m
• MatRTKLIB
• Google Earth Pro

How to use

Setup

• Install Google Earth Pro
• Clone or download MatRTKLIB and add to path in MATLAB
• Clone or download ge-drive-visualizer

1. KML generation

Execute the MATLAB script. A [m-filename].kmz file will be generated afterwards.

• example_compare_two_positions.m: Comparison of two types of position estimations
• example_drone_lidar_animation.m: Animation of a drone flight and Lidar data
• example_vehicle_animation.m: Animation of a vehicle driving
• example_vehicle_lidar.m: Display of vehicle model and Lidar data
• example_vehicle_lidar_animation.m: Animation of vehicle driving and Lidar data

2. Open [m-filename].kmz in Google Earth Pro

• Check Photorealistic and Terrain in Layers

3. Tools->Movie Maker

• Select DriveTour
• Enter the name of the file to save to
• Set the video recording settings as you want

4. Click Create Movie

• Usually this process takes quite a while!
  • In my experience, running Google Earth in a Linux (Ubuntu) environment is much faster than running it in Windows
  • On Windows, it is faster to use OpenGL instead of DirectX as the graphics mode for Google Earth
  • The type of graphics card you have can also affect performance

Note

• In the example of Lidar data animation (xxx_lidar_animation.m), there are a large number of objects to be visualized, so there are limits to how much can be displayed in Google Earth
• Long-term lidar data visualization is difficult
  • For animation, the limit is around 4000 points x 150 frames
• It is necessary to increase the memory allocated to Google Earth
  • Preferences->Cache->Memory Cache Size (MB): 1024
  • Preferences->Cache->Disk Cache Size (MB): 2048
• For large data, it is difficult to change the viewpoint with the mouse, but Movie Maker rendering and video generation work