Open pablofdezalc opened 4 years ago
Hi Pablo,
thanks, that's a good suggestion. Priority-wise, we might not address this immediately (we're all busy PhDs / Postdocs), but I'll write it down for the next version of our dataset at the very least :wink: . I'm marking this as enhancement, please let me know if this blocks you from participating in the competition. If you do this conversion, feel free to share the files with us, would save us some time.
If anyone else feels this issue is important, please like the issue, so we can track our priorities.
Best, Titus
Hi Titus,
Thanks for the info. I will look into this conversion eventually. I want to investigate if the calibration results are good enough and if I can reproduce those results across different calibrators. The evaluation results with initial calibration parameters are quite suboptimal which suggests that online calibration is needed and/or that initial parameters can also be improved.
I will share the files with you when I have done the conversion.
Regards, Pablo
On Mon, 21 Sep 2020 at 14:12, Titus Cieslewski notifications@github.com wrote:
Hi Pablo,
thanks, that's a good suggestion. Priority-wise, we might not address this immediately (we're all busy PhDs / Postdocs), but I'll write it down for the next version of our dataset at the very least 😉 . I'm marking this as enhancement, please let me know if this blocks you from participating in the competition. If you do this conversion, feel free to share the files with us, would save us some time.
If anyone else feels this issue is important, please like the issue, so we can track our priorities.
Best, Titus
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/uzh-rpg/IROS2020-FPV-VIO-Competition/issues/4#issuecomment-696104804, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJY6CD5732ZL55QNKDU4JDSG5GMNANCNFSM4ROSRPDQ .
Interesting, thanks for letting us know. I'm transferring this issue to the "main" dataset repo https://github.com/uzh-rpg/uzh_fpv_open
Hi there,
Is it possible that you can provide the calibration sequences also in ZIP format and not only in the Rosbag format like the rest of the sequences? I know that you can always convert from Rosbag to a different format but it is just to avoid doing that step if you have already the calibration sequences in the ZIP format.
Regards, Pablo