This code implements a semi-direct monocular visual odometry pipeline which has been integrated with a EKF to infused IMU and CMD as well. The feature extraction and image alignment part are implemented in OpenCL to take advantage of GPU.

OpenCL documentation

Please be sure that you have the OpenCL driver installed

You can edit the algorithm parameters in the follwoing file, please look at the config file to underestand the meaning of the parameters (GPU_version/vio/include/vio/config.h)

GPU_version/vio/param/vo_fast.yaml

1. OpenCV 4
2. Eigen
3. G2o # build g2o with -DG2O_HAVE_OPENGL=ON -DBUILD_WITH_MARCH_NATIVE=ON # G2o version => https://github.com/RainerKuemmerle/g2o/tree/memory_management
4. Boost
5. OpenCL

1. The estimated oriantetion is not accurate 
2. There is a scale map issue in the algorithm 
3. The number of matched points are increasing a lot we need to limit them in a way that will not cause some error in the estimated odometry

 log files will be written in the project folder, you can change the path in the cmake files as well as activating debug mode or not
 https://github.com/Pilot-Labs-Dev/vio/blob/111141365d86e3260cb75a9235afa47f0a1397fe/GPU_version/vio/CMakeLists.txt#L89