Computing Cloud to Cloud distance , I get
[18:49:44] [ComputeDistances] Mean distance = 0.005331 / std deviation = 0.006731
Why std deviation >Mean. Is it normal . Does it make sense.
Secondly, how can I get sigma and max relative error just like approximate values. Is std deviation is 1 sigma value?
Can I get some sort of indication of degree of completness.
Understanding Cloud to cloud distance
Re: Understanding Cloud to cloud distance
I think you should start here: http://en.wikipedia.org/wiki/Standard_deviation ;)
'Sigma' is just the classical symbol used to represent the standard deviation. And of course you can have a distribution with a mean of 0 and any standard deviation value (think of a series of alternating 1 and -1 values).
In the case of distances, especially in the case of unsigned distances, the distribution is rarely 'Normal' (Gaussian). So the mean and especially the std. dev. are not very meaningful. It's just to get an idea (and also because a lot of people asked for those values ;). The histogram itself is much more informative I think.
And there's no theoretical 'max relative error' with non-approximate distances... because they are not approximate. If there's an error (mainly due to measurements noise, registration errors and missing data or sparse density) we generally have no information about it in CC.
'Sigma' is just the classical symbol used to represent the standard deviation. And of course you can have a distribution with a mean of 0 and any standard deviation value (think of a series of alternating 1 and -1 values).
In the case of distances, especially in the case of unsigned distances, the distribution is rarely 'Normal' (Gaussian). So the mean and especially the std. dev. are not very meaningful. It's just to get an idea (and also because a lot of people asked for those values ;). The histogram itself is much more informative I think.
And there's no theoretical 'max relative error' with non-approximate distances... because they are not approximate. If there's an error (mainly due to measurements noise, registration errors and missing data or sparse density) we generally have no information about it in CC.
Daniel, CloudCompare admin