Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Error in the data and some of the files are missing in the dataset #1

Open
ozgurkara99 opened this issue Jun 2, 2024 · 0 comments
Open

Comments

@ozgurkara99
Copy link

Dear authors,

Thank you for your valuable work and the release of the dataset. I am a 2nd-year CS PhD student at the Georgia Institute of Technology, working under the supervision of Prof. James Rehg.

I am currently working on a project that requires a dataset composed of the following components:

  • Images of the scene
  • 3D point clouds
  • Projections of the 3D points onto these images

I believe your data would be very useful for our task and attempted to produce the results using the provided codebase. However, I encountered some issues. Could you please clarify my questions and assist me with these problems? Thank you very much.

Here are my questions:

  • Some of the Velodyne files are missing (I have attached the error files for reference).
  • I noticed that the 'raw' data for the 'canyon' map does not include information for all Velodyne sensors, but only includes the velodyne_360 folder.
  • The main problem I have is that I used a script to project the 3D point clouds onto images. However, in most of the cases, the projected points are shifted from their ground truth positions, even though I am using the same script that works for other files. I observed that when the car is on a slope, the projections tend to drift, which ideally should not happen. Could there be any errors in the data? If so, could you please help me resolve this issue? (the example image attached is from 'park' map, and 45th frame. I used the 45th velodyne files of the same timestamp and projected the points to the frame in the same timestamp)
  • Another major problem is that I would like to project all Velodyne data to obtain the points in the world, and then project these points into a specific frame. However, when I do this, the points drift from their original locations. I have double-checked my code and couldn't find any problems. For example, I noticed a difference between the location.txt file and the poses (in the last columns). It would be great if you could help me resolve these issues, perhaps through a chat.

I also include the python script that I used to do projection.

Screenshot 2024-06-02 at 12 38 47 [error.txt](https://github.com/user-attachments/files/15526650/error.txt) [script.txt](https://github.com/user-attachments/files/15526651/script.txt)

Thanks a lot
Ozgur

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant