comp.lang.idl-pvwave archive: archive » A better way to find a dip (local minimum with certain conditions)

Home » Public Forums » archive » A better way to find a dip (local minimum with certain conditions)

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: A better way to find a dip (local minimum with certain conditions) [message #69211 is a reply to message #69046]

Thu, 17 December 2009 14:57

James[2]
Messages: 44
Registered: November 2009

Member

what about:

FUNCTION finddip, array, min, range

; find the derivative
deriv = array - SHIFT(array, -1)

; reduce the derivative to negative and positive
pos = deriv gt 0
neg = deriv lt 0

; create convolution kernels for the elements before and after
beforekern = [replicate(1,range),intarr(range-1)]
afterkern = [intarr(range+1),replicate(1,range)]

; find places where the leading derivatives are non-positive
; and the trailing derivatives are non-negative
dip = logical_and(convol(pos, beforekern) eq 0, convol(neg, afterkern)
eq 0)

; find where the value exceeds the threshold
thresh = array gt min

; put it all together
localmin = logical_and(dip, thresh)

return, localmin

end

does what you want, it might be a little heavy on the memory but it
doesn't use any loops.

Report message to a moderator

[Message index]

		A better way to find a dip (local minimum with certain conditions) By: DavidPS on Wed, 16 December 2009 07:47
		Re: A better way to find a dip (local minimum with certain conditions) By: James[2] on Thu, 17 December 2009 14:57

Previous Topic:	Re: Catalyst Object Widget Hierarchy
Next Topic:	smooth function and rounding error

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Sun Oct 12 08:48:58 PDT 2025

Total time taken to generate the page: 0.32061 seconds