Visualize Vision Restriction Region

[anthonybrown0528]

Problem and Benefit

The current implementation of the Jetleg vision pipeline relies on a step where only a portion of points within a hard-coded region is retains to process into a heightmap. Although the resulting heightmap is easy to visualize, it is difficult to understand the boundary of the region well. This may be necessary if we need to iteratively adjust the region in the future. There are at least 2 approaches that may work to visualize the "restriction region."

In the first approach, after culling the points outside the region, the resulting point cloud can be published in a ROS 2 topic to be visualized in RVIZ2. This is relatively simple to implement, but it requires publishing potentially many points through a topic, which is inefficient. This also does not necessarily visualize the contours of the region if the points are not dense near the edges.

In the second approach, the boundary of the region is computed and visualized in RVIZ2 using the Marker message type. This requires publishing less data, so it is more memory-efficient. However, computing and ordering the points to form a wire frame of the region is possibly more complicated.

Suggested Action Items

• Modify the appropriate script in the jetleg_vision package to publish multiple ordered Marker's to make enclosing lines.
• Compute the points at the corners of the restriction region
• Determine the pairs of points which form the wire frame of the restriction region
• Publish the pairs of points to visualize the wire frame of the restriction region in RVIZ
• Test the implementation with differently shaped and sized regions
• Push the changes to the dev branch