How to put all agents to the same map? about coperception

[simbaforrest]

Do you mean how to visualize/draw/plot all the agents and all the detections in a single view?

@dekunma I think we have a visualization module for this, right?

[dekunma]

@simbaforrest Yes, we have visualization located in coperception/coperception/tools/visualization, but this module is used to visualize the original V2X-Sim dataset.
@fandulu If you are trying to put the detection boxes of all the agents on a single global map, unfortunately it is not supported by our SDK currently.
However, we implemented the functionality to fuse the detection boxes of every other agents to the ego map of one agent:

coperception/tools/det/test_codet.py

Lines 312 to 315 in 59e8fa3

	if apply_late_fusion == 1 and len(result[k]) != 0:
	box_colors = late_fusion(
	k, num_agent, result, trans_matrices, box_color_map
	)

this process is called late fusion by us.
Hopefully this would be helpful to you.

[fandulu]

@dekunma and @simbaforrest, thanks very much for your kind replies!
To @dekunma , may I understand that the variable trans_matrices are used to describe the relationship between each ego view? And we can use them to transfer the detections across different views?

[dekunma]

@fandulu You are right.
For example trans_matrices[0, j, i] is the transformation matrix to transfer the detections from agent i to agent j.
This way you can transfer detections across different views.

[fandulu]

Got it. Thanks again!