[TASK] Set Up Image and Pointcloud Fusion Ros Node

[jorgenfj]

Description of task

Set up the ros node that will be responsible for the image and pointcloud fusion. The idea is to transform the pointcloud points onto the image to get a depth estimate. However, this issue is only focused on the setting up the ros node with subscribers and message filters.
Object detection will most likely be done on the GPU with a GPU allocated image. The transformation of the pointcloud on the image will also benefit from the GPU speedup. Therefore we want to do fusion of GPU allocated messages.

Suggested Workflow

• Read up an Nitros and working with Nitros with CUDA nodes
• Reaserch tf2_ros::MessageFilter This will only invoke its registered callback when there is a transform available.
• Message filter documentation
• Look at this example for working with nitros_message_filter for GPU messages

Specifications

• We plan to use image and pointcloud both generated from the zed camera, but we should implement this to work with arbitrary sensors, i.e., only invoke callback when transforms for both sensors are available.
• Should be compatible with NITROS GPU messages.

Contacts

@jorgenfj

Code Quality

• Every function in header files are documented (inputs/returns/exceptions)
• The project has automated tests that cover MOST of the functions and branches in functions (pytest/gtest)
• The code is documented on the wiki (provide link)