Hi cdub,
We have a few follow up questions for you on how you would like to define your peaks. This is where the difficulty in creating a universal peak search comes in as everyone's definition of a peak can be different.
First, is there a certain y value, that above which would constitute a peak? Are you only wanting the single high peaks from your data or the high peak, and the medium peak to the right of this peak.
Secondly, if there is a maximum, how would you like to define what your width would be? For example, we can start at the beginning of your data, and find the first peak, and have a period of 10 data points where no maximum can occur, and repeat this process. Secondly, we can find the highest maximum in your entire data set, and break our data up into 10 width chunks based on this. As you can see, there are several different options in this binning process (especially when combined with a possibly y level).
This can also lead to a situation where we have for example three peaks. If we choose the highest peak as our starting point, and block of x points to each side of it, we can block a second peak from being registered. If there is then a third peak outside of this width that we pick up, it may be shorter than the secon peak we did not register, and thus is not the highest peak within the X point width from it. I have attached an screenshot depicting this behavior.
What we need to further advise you is a very clear definition of what you define as a peak (in terms of width, how you want to deal with these corner cases, etc). Please let me know if you have any questions.
Cheers,
Kelly R.
Applications Engineer
National Instruments
