Hi, Thank you for sharing your work. Can you also share other KI

KITTI Ros bags about limo HOT 8 CLOSED

johannes-graeter commented on July 30, 2024

KITTI Ros bags

from limo.

Comments (8)

johannes-graeter commented on July 30, 2024

Hi,

thanks for the interest in my work!
Unfortunately the bags are very big and I do not possess more webspace :) However I am on the brink of finishing integration with a python interface that could read the dataset via pykitti. Help would be appreciated, are you interested?
Best,
Johannes

from limo.

hlezki commented on July 30, 2024

Thanks for quick reply.
I have one more question. When I run your code with KITTI sequence-04, in the output poses file (/tmp/poses_dump.txt), i can not see poses for all images. In the groundtruth poses, there are 271 poses information for 271 images. Limo outputs only 112 poses for 112 images. Why is that? And how can I compare limo outputs with KITTI groudtruth ?? Is it possible to obtain same number pose information with the KITTI odometry groundtruth data?

from limo.

johannes-graeter commented on July 30, 2024

The ros version is built as an only system, so when your machine can not process the poses when they come in, they will be dropped. You can try running it slower with rosbag play -r 0.3 04.bag (i hope ingot the command right from the top if my head ;) ) By the way this is exactly why I wanted to add the pykitti interface going frame by frame :) Best, Johannes Von meinem iPhone gesendet

…

Am 28.06.2020 um 15:54 schrieb hlezki ***@***.***>: I have one more question. When I run your code with KITTI sequence-04, in the output poses file (/tmp/poses_dump.txt), i can not see poses for all images. In the groundtruth poses, there are 271 poses information for 271 images. Limo outputs only 112 poses for 112 images. Why is that? And how can I compare limo outputs with KITTI groudtruth ?? Is it possible to obtain same number pose information with the KITTI odometry groundtruth data? — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

from limo.

hlezki commented on July 30, 2024

So did you get your results that given in the paper by slowing down the rosbags for KITTI benchmark ?? And I wonder if i run rviz in the docker container?? I could not achieve running rviz in the docker. It seems it is not installed. How can I visualize the map with the RVIZ using your docker image?

Thank you

from limo.

johannes-graeter commented on July 30, 2024

Hi hlezki, No the evaluation was in real time (we also used it on our autonomous vehicle in the field). This is described in the paper shortly and more kn depth in my phd thesis (the latter in german) but depending on your machine it might be necessary to slow it down. Especially if your RAM is small, the memory allocation in swap slows it down vastly. If speed is an issue further downscaling of the images could help (in my experience the results are a bit worse but not too much). The docker image was only added very recently, i did not intend to use rviz with it but rather serve the python interface. Probably you would need to setup a remote ros master and connect that with the docker, but i habe no experience with this. you can still build it with the bash script for rviz compatibility. Best, Johannes Von meinem iPhone gesendet

…

Am 28.06.2020 um 17:38 schrieb hlezki ***@***.***>: So did you get your results that given in the paper by slowing down the rosbags for KITTI benchmark ?? And I wonder if i run rviz in the docker container?? I could not achieve running rviz in the docker. It seems it is not installed. How can I visualize the map with the RVIZ using your docker image? Thank you — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

from limo.

hlezki commented on July 30, 2024

Hi,
Thank you for your comments. I am looking forward to your new improvements about adding the pykitti interface.

Have a nice day.

from limo.

johannes-graeter commented on July 30, 2024

:) thx. Since I am working on this as a hobby, improvement is slower than I would like it to be. you are very welcome to contribute! Open source lives and dies with the contributions of the community!

Best,
Johannes

from limo.

Diggle221 commented on July 30, 2024

Thanks for quick reply.
I have one more question. When I run your code with KITTI sequence-04, in the output poses file (/tmp/poses_dump.txt), i can not see poses for all images. In the groundtruth poses, there are 271 poses information for 271 images. Limo outputs only 112 poses for 112 images. Why is that? And how can I compare limo outputs with KITTI groudtruth ?? Is it possible to obtain same number pose information with the KITTI odometry groundtruth data?

Excuse me,when I run code with KITTI sequence-04 and 01, i can not find the output poses file .Can you tell me where it is？Thank you very much.

from limo.

KITTI Ros bags about limo HOT 8 CLOSED

Comments (8)

Related Issues (20)

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent