Dr. G wrote:Perspective Corrections to Smearograms
Thank you Dr. G, this is an excellent contribution. More than ever, I feel the need to construct a list of links to posts like these. Thankfully, the board has been set up to allow long term editing. A table of contents, of sorts, could be placed into the first post to try to keep great references like this from getting buried.
OWE, I guess (in theory) you could set up a variable pixel-to-vertical height correction factor that could automatically be applied to each smearogram as it was being generated. Alternatively, the necessary correction factors could be applied to the data as required when drop vs. time plots were being created.
Absolutely. No problem to throw that into the mix*. For my own uses, I've sometimes included a value that represents the best conversion available at the time. Only recently has this occasionally gone into posts. We've discussed this a bit on physorg, you may recall.
*ah, but we probably need to discuss the mix. Buzzwords like 'process pipeline' and 'supervised learning' are no longer abstract. The required data reduction isn't cooking of the sort practiced by social scientists, so relax if that's a concern. There is mapping and correction to be applied, some best done while in the pixel domain.
----
My read_me post will have to wait. It's even better now just to get the data, so here's the summary:
1) be very, very careful; there are many pitfalls, many more than the thread has touched
2) there is very little in a smear that directly translates to physical rectilinear motion, except in "approximation"
3) the closer you look, the bigger the warts
