Vector Object Server Improvement: Use OpenCV for Polygon Rasterization

[SteveMacenski]

As the description implies, we may be able to gain some performance improvements using OpenCV rather than manually rastering polygons on the costmap for the vector object server. If so and possible, it might be good to pull that source code here to remove the direct dependency on openCV.

Good community improvement issue!

[Rudyy75]

Hi @SteveMacenski, I'd like to work on this.
I can benchmark whether OpenCV's polygon rasterization (cv::fillPoly / scanline) offers performance gains here, and if viable, isolate/adapt the essential algorithm directly into the codebase to keep dependencies lean as you suggested.
Could you point me to the relevant files for the vector object server to get started?
I'll set up a draft PR with some quick benchmarks.

[SteveMacenski]

Great! They are in the map server's package: https://github.com/ros-navigation/navigation2/tree/main/nav2_map_server 😄

[Rudyy75]

@SteveMacenski I see #6286 has been opened covering the scanline approach.
Happy to help review or test if useful, otherwise I'll look for other areas to contribute to Nav2. Thanks for the pointer to nav2_map_server 😄