luxonis/rae-ros

Image conversion and compression issues

sskorol opened this issue · 2 comments

sskorol commented

Steps:

  • run luxonis/rae-ros-robot:humble docker image
  • run bringup.launch.py w/o LED node
  • open RViz, add Camera and assign different camera topics there:

Expected: camera output is correctly rendered in RViz
Actual: none of the image topics produce the valuable output; the following errors are spammed in the log:

[component_container-1] [ERROR] [1697753825.488095736] [TheoraPublisher]: cv_bridge exception: '[16UC1] is not a color format. but [bgr8] is. The conversion does not make sense'
[component_container-1] [ERROR] [1697753825.500971843] [compressed_depth_image_transport]: Compressed Depth Image Transport - Compression requires single-channel 32bit-floating point or 16bit raw depth images (input format is: bgr8).
danilo-pejovic commented

Hey, I will try to replicate the issue over the weekend/monday. Are raw images displaying correctly in rviz?

sskorol commented

I see the following raw topics in RViz: /rae/left_back/image_raw, /rae/right/image_raw, /rae/stereo_back/image_raw, /rae/stereo_front/image_raw.

And I got the initial stream with a huge delay. But then it immediately stuck forever. Stereo images are just black with some rare white dots.

Note that I could successfully stream video via the RobotHub app w/o any delays, so it's not a bandwidth issue.