/RGBD_SLAM

Implementation of RGBD SLAM pipeline with SIFT feature tracking, incremetal ICP and pose graph optimization

Primary LanguageC++

RGBD_SLAM

Implementation of RGBD-SLAM pipeline with SIFT feature tracking, pairwise ICP and global pose-graph optimization.

Details of the implementation are documented here.

Dependencies

  • OpenCV
  • Point Cloud Library
  • g2o
  • Eigen

Build and Run

mkdir build && cd build
cmake ..
make

To run the code, go download an example rgbd dataset (fr1/room for example) from the TUM Benchmark Dataset for SLAM and specify the path to the data in src/config_params.xml, and then run the code with

cd build
./main

Results

Global reconstruction

img1

img2

img3

Local reconstruction

img4

Loop Closure Comparison:

Before LC:

img4

After LC:

img4