Ocean Optics Ocean View Install - Manual (Page 62)

4: Wizards
54 000-20000-300-02-201312
2. Use the Select Processed Node panel to choose which data node supplies the information for the
calculation of peak values. Select the desired device or node by clicking on that row in the
table. The selected source will be highlighted in
blue . Click Next to proceed.
3. Use the Configure Baseline panel to subtract off any underlying spectral structures that may
obscure the peaks that you are trying to identify.
- None -- The peak finder will evaluate all peaks at any y-axis value.
--OR--
- Advanced -- Allows you to adjust the spectrum to eliminate any non-pertinent data. Advanced
activates the advanced features and the window's preview graph illustrates how the changes
affect which peaks will be evaluated. As the advanced parameters are adjusted, the preview
spectrum in the graph changes position. Only peaks visible above the preview graph's x-axis
will be evaluated by the peak finder.
Use the Horizontal threshold adjustment when your data has a relatively flat baseline structure with
protruding peaks (like the spectrum of an Ocean Optics HG-1 mercury line light source). The value
entered into the horizontal threshold box is subtracted off the y-axis value of each pixel. As the horizontal
threshold value gets larger, the preview spectrum drops closer to the x-axis and fewer peaks will be
analyzed.
Curve fit regression attempts to eliminate any underlying base structure by subtracting off a best fit
polynomial from the spectrum. This is useful when there is an underlying curve with superimposed
peaks. An Ocean Optics PX-2 pulsed xenon light source is a good example of a blackbody curve with
protruding atomic emission peaks. In order to eliminate the underlying blackbody structure and expose
the emission peaks, a curve fit regression can be used to match and subtract the blackbody shape. A curve
fit regression of order 0 subtracts a constant that best fits the data (the average). A first order polynomial
subtracts off a straight line that best fits the data. A second order regression subtracts off a parabola that
most closely matches the data shape. As the order number increases, the underlying curve is more closely
matched and subtracted away, leaving only the sharper peaks.
4. Click Next after selecting the desired settings.
5. Use the Peak Finder panel to characterize which of the remaining spectral features should be
evaluated. Different methods of peak analysis may be used as well as different filtering options to
reduce unwanted noise effects. The preview graph at the bottom of this panel displays the
adjusted spectrum from the previous panel.
a. There are two peak finding methods available:
Local Maxima -- searches for all of the local maxima that fall within the user-defined
search parameters established in the previous wizard panel.
--OR--
Derivative -- selecting this option requires that you choose a derivative order number. A
first derivative search will locate all points where the first derivative of the spectrum is
zero (stationary points). A second derivative search will locate all points where the second
derivative of the spectrum is zero (inflection points). For either method, you must choose a
minimum peak width. This parameter, measured in nanometers, will exclude any peaks
that are narrower than the user-defined value. This is useful for eliminating small, random
fluctuations due to noise. Note that the number of peaks shown in the blue text in the