Minor tweaks to intensity calculation (code clarity, not function)

[lauralindzey]

Tested on Porcine with live data.

I'm not sure why CI is failing -- there's nothing in the logs at all.

[amarburg]

See Zulip

@all I need to do some system maintenance this weekend. The in-house CI system will be offline until early next week.

[lauralindzey]

Ah, OK! Hadn't realized that applied to CI run from Github.

[amarburg]

OK, I will ask myself a dumb question. Why does sonar_postprocess exist? Right now it is a ProjectSonarImage ->ProjectedSonarImage converter, potentially with some logspace 32bit to linear 8bit conversion. But sonar_image_proc can do that natively now (with the min_db / max_db functionality).

If so, maybe it's vestigial?

[lauralindzey]

The sonar postprocessing tools are still a bit messy, in that I use the rqt plugin most of the time, but sometimes I want to see the data in range-theta space rather than euclidean space, so for that I go back to the published drawn_sonar.

I only encountered the code because @micat001 had used it as the starting point for coloring points in sonar_pointcloud.py, and when I asked for an explanation of the calculations he was doing he pointed me here =) And, given that it had confused both of us,

[amarburg]

Hm, my memory of the history is fuzzy but I am wondering if this was a bandaid to experiment with 32-bit logspace rendering before fully integrating the functionality into the core sonar_image_proc library. If it isn't doing anyone any good, maybe change this PR to "Remove postprocess"?

[lauralindzey]

I just checked, and my launch files for visualizing sonar data still seem to depend on these nodelets. So, removing them will be a bigger cleanup task.