Upcoming SlideShare
×

# 3d tracking : chapter2-2 kalman filter

3,114 views

Published on

Published in: Technology
• Full Name
Comment goes here.

Are you sure you want to Yes No
• Be the first to comment

• Be the first to like this

### 3d tracking : chapter2-2 kalman filter

1. 1. Monocular Model-Based 3D Tracking of Rigid Objects: A Survey<br />2008. 12. 11.<br />백운혁<br />Chapter 2. Mathematical Tools (Bayesian Tracking)<br />
2. 2. 2.6 Kalman Filtering<br />The kalman filter is the best possible (optimal) estimator for a large class of problems and a very effective and useful estimator for an even larger class<br />
3. 3. 2.6.1. Kalman Filtering<br />Time Update<br />(“Predict”)<br />Measurement Update<br />(“Correct”)<br />
4. 4. Discrete kalman filter time update equations<br />project the state and covariance estimates forward from time step to step .<br />2.6.1. Kalman Filtering<br />Measurements are derived from the internal state<br />New state is modeled as a linear combination of both the previous state and som noise<br />uncertainty<br />state transition<br />actual state<br />estimate state<br />noise<br />posteriori estimate error covariance<br />priori estimate error covariance<br />
5. 5. Discrete kalman filter measurement update equations<br />the next step is to actually measure the process to obtain ,and then to generate an a posteriori state estimate.<br />2.6.1. Kalman Filtering<br />the actual measurement<br />gain or blending factor<br />measurement matrix<br />predicted measurement<br />
6. 6. 2.6.1. Kalman Filtering<br />Time Update (“Predict”)<br />(1) Compute the kalman gain<br />(1) Project the state ahead<br />(2) Update estimate with measurement<br />(2) Project the error covariance ahead<br />(3) Update the error covariance<br />Initialize<br />Measurement Update (“Correct”)<br />Initial estimates for and<br />
7. 7. 2.6.1. Kalman Filtering<br />2D Position-Velocity (PV Model)<br />
8. 8. 2.6.1. Kalman Filtering<br />2D Position-Velocity (PV Model)<br />
9. 9. 2.6.1. Extended Kalman Filtering<br />
10. 10. 2.6 Particle Filters<br />
11. 11. 2.6.2. Particle Filters<br />general representation by a set of weighted hypotheses, or particles<br />do not require the linearization of the relation between the state and the measurements<br />gives increased robustness<br />but few papers on particle based 3D pose estimation<br />
12. 12. 2.6.2. Particle Filters<br />
13. 13. 2.6.2. Particle Filters<br />
14. 14. 2.6.2. Particle Filters<br />
15. 15. Thanks for your attention<br />