thanks for your great work

[sume-cn]

It's the first LOAM project I found that can works well on KITTI dataset.
But why my run shows best result than yours in loam_velodyne/issues/117 ? Especially in the last giant arc. I've changed nothing.

Another question:
Why the result trajectory in sky direction is always within range [-1.5, +1.5], it's wired.

A typo:
README.md
scritps -> scripts

[claydergc]

One drawback of the LOAM algorithm is its computational complexity of the LM optimization. That part can make the dropping of some ROS messages and consequently the feature matching could fail. Maybe your result is better than mine because your computer is faster than my computer and the messages are not being dropped.

What do you mean with "sky direction"?

Regards!.

[sume-cn]

sorry, I mean up direction

[sume-cn]

BTW, how to compensate the rotating effect of the raw LiDAR data?
It's strange, I can't find any description on this topic in KITTI website or papers.
I also run the raw LiDAR data too, drift in yaw can be observed.

The raw bin files, timestamp files and IMU files seems not enough to do the compensation by yourself, because no individual timestamp of each point was provided as in VeloView captured pcap files.

[claydergc]

According to what I read in laboshinl's repo, the compensation you mention for the raw data is done in the line 103 of the file laserOdometry. That's why I commented that line in this repo.

[LongruiDong]

It's the first LOAM project I found that can works well on KITTI dataset.
But why my run shows best result than yours in loam_velodyne/issues/117 ? Especially in the last giant arc. I've changed nothing.

Another question:
Why the result trajectory in sky direction is always within range [-1.5, +1.5], it's wired.

A typo:
README.md
scritps -> scripts

range [-1.5, +1.5], it's wired.

Hi~
have you tested all 11 training sequence with kitti devkit?
What is the relative translation error and rotationg error of your results?

My result is strangely bad(Maybe I have someting wrong), I found that the “x direction” of the resulting trajectory is opposite to the groundtruth x axes....

[xuwuzhou]

It's the first LOAM project I found that can works well on KITTI dataset.
But why my run shows best result than yours in loam_velodyne/issues/117 ? Especially in the last giant arc. I've changed nothing.
Another question:
Why the result trajectory in sky direction is always within range [-1.5, +1.5], it's wired.
A typo:
README.md
scritps -> scripts

range [-1.5, +1.5], it's wired.

Hi~
have you tested all 11 training sequence with kitti devkit?
What is the relative translation error and rotationg error of your results?

My result is strangely bad(Maybe I have someting wrong), I found that the “x direction” of the resulting trajectory is opposite to the groundtruth x axes....

My result is strangely bad too,have you found what happened?