AUST-Hansen / rgbd-slam-tutor2

The second version of RGBD SLAM tutorial. We will discuss more about pnp, tracking, multi-thread and loop closure methods. Please refer to my blog for details: www.cnblogs.com/gaoxiang12 .

This is the second version of RGBD SLAM tutor. Please visit my blogs for details:

---

Installization

1. Dependencies: OpenCV 2.4.x , PCL 1.7, Eigen. Use commands below if you are using Ubuntu:

   sudo apt-get install libopencv-dev libeigen3-dev

2. Compile third-party libs, including DBoW2 (for loop closure), a modified version of g2o (for solving pnp), and the OrbExtractor from [orb-slam2] (https://github.com/raulmur/ORB_SLAM2).

They are all cmake projects, so just go into the directory and type:

    mkdir build
    cmake ..
    make -j2

For g2o you need to type make install to install it into /usr/local/ otherwise the FindG2O.cmake will not work.

3. Compile this project:

    mkdir build
    cmake ..
    make -j2

You will find some experiment binaries in bin/ like bin/exp_mapping, they are experiments described in the blog. Please download the dataset and edit the parameter file before doing experiments.

Many thanks to the excellent works of g2o and orb-slam!