Mini-SLAM

Authors: Juan J. Gómez Rodríguez, Víctor M. Batlle

Minimal implementation of a SLAM system for the practical assignment of the SLAM course.

Install Docker using your preferred method. If you are not sure where to start, below are some recommended options depending on your host operating system:

• Windows or macOS: Docker Desktop
• Linux: Docker CE/EE and Docker Compose

Install Visual Studio Code on your host machine.

Clone this repository:

git clone https://github.com/universidad-zaragoza/Mini-SLAM.git

Launch VS Code and go to File > Open Folder..., then select your Mini-SLAM folder.

Click on Yes, I trust the authors when prompted.

On start-up, VS Code will copy the corresponding devcontainer.[windows|linux|darwin].json to .devcontainer/devcontainer.json. You might need to do that manually if the automatic copy fails.

Click Install to add the recommended VS Code extensions.

If you do not get this prompt, you may have already installed all the recommended extensions.

Click Reopen in Container. The first time you do this, it may take up to an hour to set up the container, depending on your machine.

If you do not get this prompt, you can click the bottom-left corner (Open a Remote Window) and select Reopen in Container.

From now on, every time you open your Mini-SLAM folder in VS Code, you can immediately choose Reopen in Container from the pop-up or via the connect menu in the bottom-left corner:

You can follow the steps provided in the SLAM course handout for each lab and run everything from the terminal. However, this section explains how to make full use of VS Code’s debugging capabilities for your applications (highly recommended).

In the CMake menu, use Build All Projects and Set the Launch/Debug Target (right-click) for each lab. To define new targets in the future, edit CMakeLists.txt inside the Apps/ folder:

set(PROGRAM_NAME mono_tumrgbd)
add_executable(${PROGRAM_NAME} mono_tumrgbd.cc)
target_link_libraries(${PROGRAM_NAME} ${OpenCV_LIBS} mini_slam)

First, download the datasets required for each lab (example fr1_xyz) and place the uncompressed folders inside Datasets/. Then, run the app from VS Code Run > Start Debugging.

If you are unfamiliar with using VS Code for debugging, take a look at the guide Debug code with Visual Studio Code. Skip the video (it is too long) and just read how to start a debug session, set breakpoints, and perform data inspection.

We already provide two example launch configurations in .vscode/launch.json.

You should see two windows similar to the following:

When you are done developing your application, go to the CMake menu and change the Configure setting from Debug to Release. Rebuild and run your target.

• If you encounter an error while loading shared libraries, try running sudo ldconfig before executing your application.

• To execute the SLAM system frame by frame, go to MiniSLAM.cc and uncomment the following lines:

  // Uncomment for step by step execution (pressing the Esc key)
  while ((cv::waitKey(10) & 0xEFFFFF) != 27) {
      mapVisualizer_->update();
  }