|
|
Spot Diagrams:The following illustrations are approximate because scaling to web pages causes fluctuations in actual percent of black on white backgrounds, especially on small sized areas. These graphs are not intended to be used as standards! However, they are intended to illustrate how coverage at the same percentage can look quite different when the range of spot sizes varies. Address questions and comments to info@advisorycouncil.org.
Editorial Comments: Part I I am often asked about the "degree of coverage" as if this is some simple idea that somehow quantifies how many patches, pieces, splotches, splashes, or otherwise defined detritus resides on a substrate that may or may not (probably not) be uniform in texture, color, or some other foggily defined character. The whole notion of comparing two "pictures" or otherwise graphical representations and deciding whether they are in some way similar is a sophisticated concept. In fact, whole volumes of mathematics are devoted to this idea of similarity, so I doubt if I can actually pick up a picture of spots on surfaces and compare analytically the "coverage" to another picture of "coverage" on another surface. Nor can I devise a practical machine that could do more than increase the accuracy of my "bias" for how I decided to make the comparison in the first place. The problem is more than technologically intractable; it is, in fact, impossible. Yes, I know some are going to bring in the argument about "practical limitations" and "within reason" and so forth...but these are valid only if one recognizes what one means by "practicality". The list of parameters that one would have to consider if one were to attempt to build a "practical" (othr than the human eye) machine are formidable: lighting, color, contrast, saturation, reflectivity, angle of incidence, diffusivity, spectral absorption at different spectral ranges, just to mention a few that are born out of my own considerable experience in analytical spectroscopy. This assumes one would want to make a measurement of the "visual appearance" of the surface. What about the "compositional" appearance of the surface? One could measure the "amount" of something on the surface without regard to its distribution. One thing comes to mind, here. Salt contamination is often measured on surfaces as if the salt is uniformly distributed. This assumption may be bad if you reduce the sampled area too severely. Charles Frenzel Editorial Comments: Part II
|
|