Raw rosbags missing for 3 of the 11 VAL flights

[Smoolak]

The released rosbags (TD_01_processed.bag through TD_08_processed.bag) cover only 8 of the 11 VAL trajectories. By timestamp they map to:

bag	VAL flight
TD_01	long_trajtr/2023-03-16-18-04-01 (LT.1)
TD_02	long_trajtr/2023-03-18-12-18-25 (LT.2)
TD_03	long_trajtr/2023-03-18-14-38-32 (LT.4)
TD_04	long_trajtr/2023-03-18-15-01-14 (LT.5)
TD_05	long_trajtr/2023-03-18-15-40-18 (LT.6)
TD_06	short_trajtr/2023-03-18-16-30-27 (ST.3)
TD_07	short_trajtr/2023-03-18-16-43-16 (ST.4)
TD_08	short_trajtr/2023-03-18-16-55-37 (ST.5)

The other 3 VAL flights have no corresponding bag:

VAL flight	why it looks uncovered
short_trajtr/2023-03-11-11-48-35 (ST.1)	no bag from 2023-03-11 in the release
short_trajtr/2023-03-16-16-58-43 (ST.2)	starts about 1 h before the earliest 2023-03-16 bag (TD_01)
long_trajtr/2023-03-18-12-47-05 (LT.3)	sits in the gap between TD_02 (ends 04:32 UTC) and TD_03 (starts 06:44 UTC)

The GeoVINS paper reports 11 collected flight datasets (TD_1 to TD_11, over 67 km), so raw data for more than these 8 sequences presumably exists. Two questions:

1. Do rosbags exist for ST.1 / ST.2 / LT.3 (timestamps above)? If so, can they be released or shared on request?
2. If they weren't captured with the rosbag rig, is any IMU / attitude / altimeter data available for them?

[hmf21]

Hi @Smoolak
Thanks for these details, Flight Trajectories in GeoVINS paper and AerialVL paper are two different naming conventions, but the data will overlap partially,
For the rosbag files ( contain IMU / attitude / altimeter data topics and 11 rosbags in total) in AerialVL paper, we can release them if you need them.
I will inform you as soon as possible and it won't be later than this month.

[Smoolak]

Thank you for offering to release the rosbags. That would be very useful.

Looking more closely at the data, I am confused, because it seems I do not have a complete set for either the AerialVL benchmark or the GeoVINS paper, and I am not sure how the two relate. I would be grateful if you could help clear this up.

Here is what confuses me. The rosbag files use a TD numbering, and I can match TD_01 to TD_08 to eight of the eleven AerialVL VAL sequences by timestamp. However, one of the rosbags I have, TD_10, is a flight from September 2024, which cannot be an AerialVL VAL sequence, since those were all recorded in March 2023. This is consistent with the GeoVINS paper, which reports 11 flights collected from 2023 to 2025. So the AerialVL 11 and the GeoVINS 11 do not appear to be the same set, and I cannot tell whether my data is complete for either one.

Could you help me understand the mapping between the TD rosbag numbering and the AerialVL VAL sequences, so I can see which rosbags correspond to which flights, and in particular whether rosbags exist for the three March 2023 sequences I am still missing? If those three were not recorded with the rosbag rig, is any IMU, attitude, or laser altimeter data available for them?

Thanks again.

[hmf21]

Hi @Smoolak
Thanks for these details. These rosbags in GeoVINS paper are preprocessed and I am sure that I can find the raw recorded data. But it still needs some time to find the right corresponding between AerialVL, GeoVINS and those raw recorded rosbags. I will try my best to get the right match.
Please inform me if these is any questions.

[Smoolak]

Hi @hmf21, following up on this issue. Were you able to locate the raw rosbags and determine their mapping to the AerialVL and GeoVINS flights? Thanks.