The problem
The problem we have is that we wish to use some measure of the edge angles at meetpoints to assist in categorizing them.
Specifically, we would like to be able to use the edge angle information to partition a set of meetpoints of the same degree
into at least two subsets. In order to do this we require a measure that has a mean and variance that is distinguishable for
distinct meetpoints.
We have used various comparison methods, and found that for each method the means and variances can be very similar for
greatly distinct meetpoints. This means that the edge angle data from a single sample cannot be accurately placed into a subset.
Examples from the Wean Hall 6th floor data are shown below.
As you can see, there is significant overlap of the samples. There is enough distinctness in the distributions to quantify the
likelihood of a single sample to belong to one or another meetpoint, but neither likelihood is zero, and so we cannot use these
data distributions as subset definitions.
The temporary decision
Given meetpoint minima distance and edge feature map data to localize, the additional information that could be gained from edge
angles is not vital to the process. For the time being, we are omitting this aspect of the meetpoints for node comparison and
categorization. This certainly does not mean that edge angle information will not be used in the future.
The image below shows edge angle symmetry data at a meetpoint of degree three with equidistance of approximately 51 inches.
The image below shows edge angle symmetry data at a meetpoint of degree three with equidistance of approximately 85 inches.
