carla-simulator / carla

Open-source simulator for autonomous driving research.
http://carla.org
MIT License
11.04k stars 3.55k forks source link

Obtaining Visualizations from Trailer Videos #367

Closed jcyriac2 closed 6 years ago

jcyriac2 commented 6 years ago

Hey! Great work on the new release 0.8.* We are really looking forward to the performance improvements

1) I just wanted to know how I can obtain the visualizations used in the 0.8.0 trailer video on Ray Cast based LIDAR (https://youtu.be/S_XbKHC_hN0?t=25s), depth-based point Cloud (https://youtu.be/S_XbKHC_hN0?t=46s). Is the code available in this repository?

2) Also, I was wondering how to get the overlay of the camera image, the steering wheel and the gas/brake level visualizations as seen in (https://youtu.be/cFtnflNe5fM?t=1m1s). Again, is this code available in the repository or do we have to implement it ourselves?

Thanks again!

felipecode commented 6 years ago

Hello @jcyriac2,

Answering your question 2. We have code for that. Right now I don't have much time to polish it but i can make a gist for you: https://gist.github.com/felipecode/9157947712fbc5923bf4325216bf85c0 The function is the plot_driving_interface The overlay, of course, was made using some video editor.

jcyriac2 commented 6 years ago

Thanks alot @felipecode.

marcgpuig commented 6 years ago

Hello @jcyriac2,

For your question 1. The code is full provided in this repository :) To obtain the LIDAR visualization you must run the manual_control.py with the argument --lidar (or -l). If you are looking for the rendering implementation, please take a look at these lines:

https://github.com/carla-simulator/carla/blob/79c81bf3feacc532235358e2096a4dc015517a2c/PythonClient/manual_control.py#L324-L336

To get the depth-based point Cloud, you only need to run the point_cloud_example.py. It will generate .ply files that you can easily open and edit with MeshLab.

Regards!