Format Description
Each line corresponds to a measurement, either odometry or
landmark. For each discrete step there is on
STEP line and 0 to many landmark
lines.
Example of a STEP line
STEP largemap240803e.2347 -0.00595 -0.19776
-0.00204 0.000019729280 -0.000000126442 0.000015893272 0.000029223686
-0.000000977252 0.000296127252
The single fields are separated by whitespace. The meaning of each
field is then:
STEP <image> <dX> <dY> <dPhi> <cXX> <cXY> <cYY> <cXPhi> <cYPhi> <cPhiPhi>
| Field | Explanation |
|---|
| <image> | The name of the image file without suffix |
| <dX> | Movement on the x-axis in respect to last robot position |
| <dY> | Movement on the y-axis in respect to last robot position |
| <dPhi> | Movement in rotation in respect to last robot position |
| <cXX> | Covariance matrix of the odometry |
| <cXY> | Covariance matrix of the odometry |
| <cYY> | Covariance matrix of the odometry |
| <cXPhi> | Covariance matrix of the odometry |
| <cYPhi> | Covariance matrix of the odometry |
| <cPhiPhi> | Covariance matrix of the odometry |
LANDMARK_C and LANDMARK_L lines
The information about circle and line measurement differ. There are two formats:
LANDMARK_C -0.754798 -2.38293 0.961176 0.00174179 0.000243156 0.0026056 6
LANDMARK_C <pX> <pY> <quality> <cXX> <cXY> <cYY> <ID>
| Field | Explanation |
|---|
| <pX> | x-coordinate of the position of the landmark in the robots frame |
| <pY> | x-coordinate of the position of the landmark in the robots frame |
| <quality> | The quality of the landmark (not so useful) |
| <cXX> | Covariance matrix of the odometry |
| <cXY> | Covariance matrix of the odometry |
| <cYY> | Covariance matrix of the odometry |
| <ID> | The ID from ground-truth |
LANDMARK_L -0.300722 -0.586752 -0.220978 0.0441848 0.255339 0.00867793 -0 0.0027596 4 5 6
For the lines the meaning is:
LANDMARK_L <xntop> <yntop> <xnbottom> <ynbottom> <alpha> <beta> <quality> <sigma_alpha> <ID> [<ID> ... ]
| Field | Explanation |
|---|
| <xntop> | Top and bottom pixel position of the lines, only for visualization |
| <yntop> | Top and bottom pixel position of the lines, only for visualization |
| <xnbottom> | Top and bottom pixel position of the lines, only for visualization |
| <xnbottom> | Top and bottom pixel position of the lines, only for visualization |
| <alpha> | Angle from robots coordinate system to landmark (top view) in radian |
| <beta> | Angle of the vertical of the line in respect to gravity frame in radian |
| <quality> | Geometric quality (0..1.0) of the line |
| <sigma_alpha> | Covariance of the line )(each measurement can combine more then one line extracted from the vision processor) |
| <ID> | The ground truth ID (from far away many lines are seen as one, so each line may split up when the robot gets nearer to a feature. Then there are more fields with IDs |
The exact meaning of each field is also documented in the header of corresponding file. The provided example programs
can give an example of how to read the files. Or even better use the example programs to run your algorithm.
| 
