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

288 fix uyvy2rgb size #292

Merged
merged 2 commits into from
Nov 13, 2023
Merged

288 fix uyvy2rgb size #292

merged 2 commits into from
Nov 13, 2023

Conversation

flynneva
Copy link
Collaborator

@flynneva flynneva commented Nov 4, 2023

Closes #288

Signed-off-by: Evan Flynn <evanflynn.msu@gmail.com>
Signed-off-by: Evan Flynn <evanflynn.msu@gmail.com>
@zymouse
Copy link

zymouse commented Nov 8, 2023

image Thank you very much, problem solved!

Copy link

@zymouse zymouse left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Tested with sensing camera

@zymouse
Copy link

zymouse commented Nov 10, 2023

Subscribe to Topics: /gmsl/image_raw/compressed
CPU Result:CPU resource usage is relatively high

image

Subscribe to Topics: /gmsl/image_raw
CPU Result:~2.0
image

Cause analysis:

Using rqt_image_view software to subscribe to compressed topics causes cpu to rise

@zymouse
Copy link

zymouse commented Nov 10, 2023

Unusual frequency of topic releases

  • pixel_format: "uyvy2rgb"

  • Use rqt_image_view software to subscribe to compression topics
    image

  • pixel_format: "uyvy"

  • Subscribe to Topics: /gmsl/image_raw

    • Use rqt_image_view software to subscribe to compression topics
      image

Cause analysis[Decrease in frequency]:

Related to rqt_image_view software subscription

Analyze the causes:

Cause 1

Subscription to image_raw/compressed Topic frequency is normal ✅
So:
image_raw should not be subscribed to during camera use. Instead, subscribe to image_raw/compressed to process camera data.

My way:

  • Currently ros2 subscribes to topics with bandwidths of 100MB/S or more, which can make frequency degradation.
    Frequency reduction is fatal for sensor fusion.
  • With compression, there is a problem of reduced resolution, which reduces image quality.
    Image quality degradation - fatal for a long range object recognition [traffic light recognition].

I tested using opencv's resize, which solves the problem of large bandwidth lowering frequency and compression-picture quality degradation.

image

@flynneva
Copy link
Collaborator Author

@zymouse this MR just fixes the uyvy size issue. The other frequency issue I think can be tackled in another issue / PR. Could you approve this PR to close the sizing issue and open a new issue for the frequency issue?

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