Closed lotterf closed 6 months ago
@lotterf pls take a look at the following: https://github.com/nutonomy/nuscenes-devkit/blob/4df2701feb3436ae49edaf70128488865a3f6ff9/python-sdk/nuscenes/utils/data_classes.py#L329-L330
if you do not want to use vx_comp
and vy_comp
, you can try working with vx
and vy
Yes, I know that the given vx and vy can be used to correct the data and reduce the error, but I find from the paper that vx and vy can not get the actual speed, I want to know whether there will be any error after correction. If so, I wonder if it will be big?
I don't recall the exact number, but the error shouldn't be too significant, I think
Ok, I get it. Thank you for your patience.
@lotterf pls take a look at the following:
if you do not want to use
vx_comp
andvy_comp
, you can try working withvx
andvy
Hello, I would like to ask how vx and vy are obtained. What radar can measure is radial velocity and azimuth. Are vx and vy obtained by decomposing radial velocity in radar coordinate system? Is the source code available and where can I find instructions?
@HYY-ying I believe this is internal to the radar
Hi, the data set published by your team is very helpful to me, thank you and your team for your free effort. However, I still have some questions about the multi-frame radar accumulation in the data set. I found that the self-driving motion had been calibrated and aligned by the self-driving sensors. However, how should I calibrate the speed of various dynamic objects in the scene? It might be possible to use the acquired radial velocity to correct, but I think this would lead to a certain bias.