daniilidis-group / ffmpeg_image_transport Goto Github PK

Am transmitting the ffmpeg image to a remote computer, Is there a way to decode the ffmpeg image after transmitting the image over UDP

When i try to this i got the following warning

Hi,
I'm trying to do video streaming via wireless channel using ROS.
Since the wireless channel has low bandwidth, video encoding(ENC) and decoding(DEC) processes are essential, and this code is very useful for what I'm trying to do.

By following the provided instructions, I successfully displayed the video stream after ENC and DEC processes.
The video source used for testing was 720p@20Hz, which is not quite high quality.
In my case, however, the decoded video has a long delay (about 2.5 seconds).
Since the ENC and DEC are performed on the same computer, it is not caused by network delay.

Is there any way to shorten the ENC+DEC time?
Currently I am using libx264 codec. (I did not test yet with nvenc due to NVIDIA driver version problem)
Can nvenc speed up the ENC+DEC process significantly?

Thank you,

Seo

Hello,

to be able to deal with my available data rate I was very happy to stumble across this project! Using noetic your code works perfectly for live feeds. Well done!

But I am not able to store the images/video in a rosbag. Have you experience with that?
My test was:

• record just /camera/image_raw/ffmpeg in a bag
• replay that bag (on a fresh new roscore with nothing running)
• open rqt_image_view and try to subscribe to the image - but here the whole mission fails as I am not able to see the topic in the drop down menu. I can perfectly see the topic beeing published using rostopic hz /camera/image_raw/ffmpeg

Hi,
I am trying to build the workspace , it is throwing me a lot of declaration errors for the variables or functions used. Here is the screenshot

I am using : Ubuntu 18.04, ROS melodic

Hello! Is there a ROS2 version planned? Thanks.

Hello,

I have installed ffmpeg_image_transport and ffmpeg_image_transport_msg, compiled everything from source (also ffmpeg) and works. But when I try to watch the video with rqt_image_view this is the result:

Any ideas about the reason?

Thank you,

Alberto

Hi,
I saw in the description that nvidia hardware could accelerate the processes.
I'm running without GPU and ffmpeg -i ./test.mp4 output.webm works fine, but If I want to use
rosrun image_transport republish raw in:=/kitti/camera_color_right/image_raw_crop ffmpeg out:=/kitti/camera_color_right/image_raw_crop/ffmpeg

I get this error message.

[hevc_nvenc @ 0x56547a989800] Cannot load libcuda.so.1
[ERROR] [1635969757.108212684]: cannot open codec!
[ERROR] [1635969757.108369255]: cannot initialize encoder!

Is it possible to use the script without CUDA kernels?

My lab group and I are very interested in using this for streaming video in our underwater robotics project but noticed that there is no license posted -- would you consider releasing the code under either a permissive license or a license permitting academic and/or non-commercial use?

Thank you very much for creating this!