Martin Valgur

I'm not sure how `np.fromfile()` works on big-endian machines, but I agree, it would not hurt to use `dtype='

Related: https://github.com/ethz-asl/kitti_to_rosbag/issues/7

@wxt-github The mapping from odometry to raw datasets is provided for the first 11 sequences in the readme of the [odometry devkit](https://s3.eu-central-1.amazonaws.com/avg-kitti/devkit_odometry.zip). | Nr. | Sequence name | Start |...

I guess you can consider it a semi-official plugin now. 😄 I mentioned this PR to Fleep and they were satisfied with the quality, it seems, and even went as...

This looks like a fantastic improvement for both the stereo calibration accuracy and convenience of the calibration workflow. Thank you! The new logic does not seem to share data collected...

Also, if the monocular calibrations are available, then the `self.c.goodenough` / `self.c.compute_goodenough()` could be relaxed considerably. Even a single checkerboard image pair is enough to calculate stereo extrinsics for calibrated...

Indeed, I apparently messed up pretty badly when trying to rebase your repo on top of the original cartographer ones to get a clear overview of the changes made here....

Hi! Sorry, I did not notice your message until now. I agree, the script should first check whether virtualenv already exists and then prefer the apt version if it does...

The coordinate values in the footprint geometries are perfectly correct in this case. The issue is more related to how your specific GIS software handles geometries that cross the dateline....

> So the "fix" would be to save the footprints in a different projection? No, simply changing the overall Project CRS (to some UTM zone perhaps) in QGIS should be...