search

dusty-nv / ros_deep_learning

Deep learning inference nodes for ROS / ROS2 with support for NVIDIA Jetson and TensorRT
879 stars 257 forks source link

image is in rgb8 format, expected bgr8, #24

Open vxgu86 opened 4 years ago

vxgu86 commented 4 years ago

I am getting image_raw from a gscam drived camera, with /camera/image_raw subscribed. the detectNet error with the following:

[cuda] cudaAllocMapped 131072 bytes, CPU 0x101c30000 GPU 0x101c30000 [cuda] cudaAllocMapped 32768 bytes, CPU 0x101b40000 GPU 0x101b40000 [ INFO] [1570526393.448824707]: model hash => 965427319687731864 [ INFO] [1570526393.448906165]: hash string => /usr/local/bin/networks/ped-100/snapshot_iter_70800.caffemodel/usr/local/bin/networks/ped-100/class_labels.txt [ INFO] [1570526393.452781894]: node namespace => /detectnet [ INFO] [1570526393.452866842]: class labels => /detectnet/class_labels_965427319687731864 [ INFO] [1570526393.467713092]: detectnet node initialized, waiting for messages [ INFO] [1570526393.722328769]: converting 1920x1080 rgb8 image [ERROR] [1570526393.722481165]: 1920x1080 image is in rgb8 format, expected bgr8 [ INFO] [1570526393.722541269]: failed to convert 1920x1080 rgb8 image [ INFO] [1570526393.922863144]: converting 1920x1080 rgb8 image [ERROR] [1570526393.923110279]: 1920x1080 image is in rgb8 format, expected bgr8 [ INFO] [1570526393.923377988]: failed to convert 1920x1080 rgb8 image

how can I change the image_encodings???

cocodirasta commented 4 years ago

Hi,

Have you found a way to solve your issue ? I have the same problem.

Thank you

gclarke42 commented 4 years ago

It looks like an in code fix for this is in progress but incase you need a faster solution. I was able to make a sub/pub node using cv_bridge which converted the rgb to bgr and then republish the new image. Not an ideal solution but was pretty simple. The cv_bridge example was basically what I used for the node. http://wiki.ros.org/cv_bridge

qwedaq commented 4 years ago

It looks like an in code fix for this is in progress but incase you need a faster solution. I was able to make a sub/pub node using cv_bridge which converted the rgb to bgr and then republish the new image. Not an ideal solution but was pretty simple. The cv_bridge example was basically what I used for the node. http://wiki.ros.org/cv_bridge

hi i am facing the same error. can you please tell me where in the code can i fix this issue? I am new to this and i am not able to figure it out