PoseTool download page

PoseTool is a Qt based software for offline processing of 6D pose sequences. Features include Kalman filtering (including data fusion of different sequences) intuitive manual Kalman filter tuning automatic Kalman filter tuning (requires GSL for optimization) pose inversion (with correct update of covariance) combining different sequences, similar to multiplying homogeneous matrices (with covariance computation) time shifting covariance scaling

Requirements

• Qt
• GNU scientific library (GSL) -- optional for automatic kalman filter tuning

Note that the software was so far only tested on SuSE Linux.

Short tutorial

This short introduction shows how to use the PoseTool for Kalman filtering. Other operations can be performed in a similar manner.

1. Download and unpack one of the sample sequences from below.
2. Open PoseTool and load the sample file.
3. Select the newly loaded sequence in the main window and click on "Kalman Filter".
4. A new sequence will show up in the main window. Select it together with the original sequence and click on "Display".
5. Enable "Position" in the Kalman filter window for the Kalman filter to do something.
6. Now you can change the parameters of the Kalman filter and you will immediately (depending on the speed of your computer, of course) see the effects in the display window.

From this example, you can already see the basic principle of the software: Each operation creates a new sequence on which further operations can be carried out. If one of the input sequences changes, the result is also updated. This way, a long chain of operations can be constructed, where modifications on every stage are propagated to the results. Note, however, that recomputation only occurs as long as the window controlling each operation is still open...

Data format

Todo

Download

• source code: PoseTool-1.0.tar.gz
• sample pose sequence (ART tracker): PoseLogART.gz
• sample pose sequence (AR toolkit): PoseLogArToolkit.gz

Further reading

• DaPustka -- my diploma thesis, contains a lot of information about the mathematics
• The following ISMAR paper contains some information about how uncertainties are represented and propagated

J. Newman, M. Wagner, M. Bauer, A. MacWilliams, T. Pintaric, D. Beyer, D. Pustka, F. Strasser, D. Schmalstieg, G. Klinker Ubiquitous Tracking for Augmented Reality International Symposium on Mixed and Augmented Reality (ISMAR) 2004, Arlington, VA, USA, Nov 2-4, 2004, pp. 192-201. (bib)

-- DanielPustka - 03 Jun 2005