SLAM Integrated Mobile Platform

University project to integrate a laserscanner, GNSS antenna and remote controled vehicle in order to perform SLAM and georeference the results.

Overview

This repository includes the processing code for data collected with the SIMP vehicle as well as some example data to run the program. The project is structured as follows:

 your_project_folder/
	│
	├── Code/  
        │   ├── functions/  
	│   ├── Georeferencing.m 
	└── Data/  
	    ├── GNSS Trajectories/  
	    ├── GoPro Images/  
            │   └── GoProSchloss/  
 	    ├── Scan Trajectories/  
	    ├── Scans/  
	    └── output/  

How to use

The Georeferencing function includes the main code that will be used to call all functions related with this project. To process the example data simply run Georeferencing.m and select the corresponding data when prompted.

The data can be downloaded from Google-Drive via this link (~ 2GB):
https://drive.google.com/drive/folders/1lvRP46lA1Gb7gnMfUR8oFm5_vJhGkE-_?usp=sharing

Simply download "SIMP Data" and integrate the folders into your_project_folder/Data/ to match the abovementioned project structure.

The following MATLAB Add-Ons need to be installed:

• Aerospace Toolbox (by MathWorks)
• Mapping Toolbox (by MathWorks)
• Signal Processing Toolbox (by MathWorks)
• Symbolic Math Toolbox (by MathWorks)
• Lidar Toolbox (by MathWorks)
• Image Processing Toolbox (by MathWorks)
• lasdata (by Teemu Kumpumäki)

After successfully running the code, the output pointcloud and trajectory can be found in your_project_folder/Data/output/

For a Cesium-based webvisualizer of our results, visit https://ifpwww.ifp.uni-stuttgart.de/philipp/SIMP-viewer/SimpViewer.html or go to https://github.com/tjkayser/SIMP-Viewer to access the viewer code files.

Processing code summary

The main steps of the program are:

• Loading all necessary data (see Data folder)
• Calculating Time Offset (to connect Scan and GNSS data)
• Coarse trajectory match (as preparation for the ICP-based accurate matching)
• Accurate trajectory match (using temporally closest points of both trajectories)
• Estimate point matching accuracy (statistical values for relative accuracy of matching algorithm)
• Transform point cloud (apply combined transformation)
• Colorize point cloud (uses GoPro Data to determine colors)
• Remove moving objects (to clean up the point cloud)
• Ground classification (to seperate the points by ground and get some semantic information)
• Save final cloud (in Earth-Centered-Earth-Fixed (ECEF) coordinates)

The presentation slides of the project can be found here:
https://docs.google.com/presentation/d/1rTlY-6AHuGG7iMSqez656pDZ-7JLuCv1/edit?usp=sharing&ouid=113886872871499490206&rtpof=true&sd=true