luxonis/depthai-ros

Support for Image Transport

berickson opened this issue · 3 comments

I started playing with the examples, and it looks like the ros endpoints for camera images only support the raw format. I was previously publishing images with python and could see a huge performance improvement for remote subscribers using compressed images.

It looks like the recommended way to do this is to use image_transport (see http://wiki.ros.org/image_transport). That adds support for multiple publish endpoints, compressed, Theora, etc.

Do you support / have any plans to support this? It looks like it might require BridgePublisher specializations for sensor_msgs::Image. I might try to code it up in a fork if you are interested.

The reason why I didn't use it is because it needed specialization.
And I left NV12 to be published as it is to provide backward compatibility to depthai data structures.
Wouldn't it be easier to use a nodelet to republish using image_transport ?

A nodelet would probably be a great way to do it.

Image Transport Support added. Closing the issue.