Spatial median filter

[YBachmann]

This PR adds a 1D spatial median filter.
The size of the filter can be set via the parameter window_size.

In this screenshot you can see the raw laserscan in blue and the result of the spatial median filter in red (window_size=31).
The filter smoothes the walls and also removes isolated/scattered noise (see e.g. directly right of the robot).

[jonbinney]

This looks quite useful! I'll take a closer look tomorrow.

[jonbinney]

Added some questions. Also could you add an example yaml and python launch file in the examples folder?

[jonbinney]

It looks like you've made some changes to address my comments. When you have a chance, could you reply to each review comment to note that you've addressed it, and let me know when this PR is ready for another look?

[YBachmann]

It looks like you've made some changes to address my comments. When you have a chance, could you reply to each review comment to note that you've addressed it, and let me know when this PR is ready for another look?

Hey @jonbinney ,
I replied to each comment and marked it as resolved, please re-open a comment if you think it still needs something changed.
Edit: I Just noticed, that an example yaml and Python launchfile are still missing. I will add those on monday.

[jonbinney]

Thanks! One more small request: could you add a comment in the code about why the window size needs to be incremented if it is even? Your explanation was good, just copy it into the code there: "If the window_size is even, the current range, which will be modified cannot be centered in the window." Once you do that and add the example files, this should be ready for me to merge.

[YBachmann]

Hello @jonbinney ,
sorry for the late response, I had some urgent things to do this week.
I added an example xml and python launchfile and a parameter yaml file that's used by the launchfiles. I also added a comment regarding the odd window_size_ in the code.
Should this be the last update, I would like to thank you for taking the time to reviewing this PR!

[jonbinney]

Merging - thanks for contributing this @YBachmann !