thanks for the great dataset!
However, looking at the ground truth provided with the loop1 sequence, I noticed that there are no loop closures in the trajectory. I plotted the poses using the evo library, and the results look like this:
To reproduce this plot, you can use
evo_traj tum gt_odom.txt -p --plot_mode xyz
in the directory where gt_odom.txt is stored. The figure of this sequence in the paper looks different, with correct loops detected.
In loop2 and loop3, the loop closures look correct, and the ground truth path aligns with the figure in the paper.
Hi,
thanks for the great dataset! However, looking at the ground truth provided with the loop1 sequence, I noticed that there are no loop closures in the trajectory. I plotted the poses using the evo library, and the results look like this:
To reproduce this plot, you can use
in the directory where gt_odom.txt is stored. The figure of this sequence in the paper looks different, with correct loops detected. In loop2 and loop3, the loop closures look correct, and the ground truth path aligns with the figure in the paper.
Looking forward to your reply!