Revision history [back]

I personally do the assembly of my Point Clouds in a separate node. I do this so that you abstract the assembly from your code. This increases re-usability of code and makes the code more generic for future uses. It is also good programming practice to create sub-systems within code that each perform a function without affecting other sections of code. This makes the code cleaner, simpler, and easier to maintain.

Additionally, if you generate the assembled cloud in a different node, you can write your point cloud processing node as a filter plugin to the laser filter chain provided by ROS (which also works on Point Clouds).