L3: VSLAM 1 Flashcards
What is WIndowed BA?
Iteratively refine over the last m number of frames to obtain a more accurate estimate of the local trajectory.
❗️❗️❗️What defines VO?
Process of incrementally estimating the pose of the camera by
examining the changes that motion induces on the images.
What is VSLAM?
❗️❗️❗️What is the difference between VO and VSLAM?
VO:
1. aims at LOCAL consistency of the trajectory
2. building block of SLAM
3. VO is SLAM before closing the loop!
4. There is drift (“solved” with windowed BA)
VSLAM:
1. aims at GLOBAL consistency of the trajectory
2. uses LC
What is the difference between VO and SFM?
VO:
What different moion estimation exist?
- 2D-2D
- 3D-3D
- 3D-2D
What is Loop Closure (LC)?
What is Structured for Motion (SFM)?
What properties are important when performing VO?
- Sufficient illumination in the environment
- Dominance of static scene over moving objects (Stationary object preferred over e.g. moving cars)
- Enough texture to allow apparent motion to be extracted
- Sufficient scene overlap between consecutive frames
Advantages of VO
VO > wheel odometry → Not affected by wheel slip
Flow chart of VO
Image sequence
↓
Feature detection
↓
Feature matching (tracking)
↓
Motion estimation
(2D-2D, 3D-3D, 3D-2D)
↓
Local optimization
Why would we use BA and not just VO?
Computes camera path incrementally (pose after pose), the errors introduced by each new frame-to-frame motion accumulate over time. Generates drift of the estimated trajectory from the real path.
To keep small as possible BA is needed as it determined the projection error that minimizes it.
WIndowed BA as a solution to VO locally
Which two approaches can be used to estimate the relative motion (T_k) between frames?
Under input sequence process
- Appearance-based → Intensity information of all pixels in both images. Slow, computationally heavy, worse at estimating and dense.
- Feature-based → Repeatable features extracted across the images. Faster and more accurate, and sparse.
What is 2D-2D?
- Both features are defined in 2D. - Mostly used in monocular VO
- Minimal-case solution involves 5-point correspondences (Nister).
- Or you can use 8-point correspondences (Longurt). Uses SVD at the end.
What is 3D-3D?
- Both features are specified in 3D
- Triangulate 3D points
- Minimal-case solution involves 3 non-collinear correspondences
- Solution is found by aligning transformation that minimizes 3D-3D distance