• Direct Sparse Odometry, J. Engel, V. Koltun, D. Cremers, In arXiv:1607.02565, 2016

• Stereo DSO:Large-Scale Direct Sparse Visual Odometry with Stereo Cameras, Rui Wang, Martin Schwörer, Daniel Cremers, 2017 IEEE International Conference on Computer Vision

• Direct Sparse Visual-Inertial Odometry using Dynamic Marginalization, Lukas von Stumberg, Vladyslav Usenko, Daniel Cremers, 2018 IEEE International Conference on Robotics and Automation (ICRA)

Please follow https://github.com/JakobEngel/dso.

1. Modify the bash file

2. For stereo:

   bash ./run.bash
   

3. For mono:

   bash ./run_mono.bash
   

• T_stereo file:

    transform matrix from right camera to left camera (4x4).

• imu_info file:

    transform matrix from left camera to imu (4x4).
    gyroscope noise density [ rad / s / sqrt(Hz) ] ( gyro "white noise" )
    accelerometer noise density [ m / s^2 / sqrt(Hz) ]   ( accel "white noise" )
    gyroscope random walk [ rad / s^2 / sqrt(Hz) ] ( gyro bias diffusion )
    accelerometer random walk [ m / s^3 / sqrt(Hz) ].  ( accel bias diffusion ) 

• 20190424

SE(3) Umeyama alignment：

No alignment（Initialization may have an impact on the results）:

• 20190416

Apply First Estimates Jacobians to scale to maintain consistency.

EuRoC MH01:

• EuRoC V1_03:

• EuRoC V2_03:

• 20190409:

green line: groundtruth, redline: VI-Stereo-DSO

This project aims at verifying the location performance of VI-Stero-DSO. The computational efficiency is to be improved. Actually in current code, many preintergration operations are redundant and can be simplified.