US 10,760,966 B2Granted

Method for the correction of background signals in a spectrum

AssigneeAnalytik Jena AG

InventorEike Thamm

Life of this patent

Estimating the term…

Loading patent…

Abstract

A method for the determination and correction of background signals in a spectrum, consisting of signals of a plurality of spectral points, characterized by the steps of: Calculating at least three statistic or analytic functions of the signal values of the spectrum, attributing probabilities Pi(band) for the presence of bands to each point in each of the calculated functions: Adding the probabilities Pi(band) up to an overall probability ΣPi(band) from all calculated functions for each point; calculating a probability P(background) for the presence of background for each point in the spectrum from said overall probability ΣPi(band) according to P(background)=1−ΣPi(band) wherein negative values are set to zero; and calculating a fit of the signal values at all points of the original spectrum wherein the signal in each point is taken into account in the fit only with the respective probability for the presence of background P(background), and subtraction of the background function determined in such a way from the signal values of the original spectrum in order to generate a background corrected spectrum.

Claims12

Dependent claims

A method for the determination and correction of background signals in a spectrum to generate a background corrected spectrum for use in spectral analysis, comprising:

obtaining a spectrum consisting of a plurality of signals, each signal forming a spectral point in said spectrum and having a signal value, some signal values in said spectrum representing background;
and
removing said signal values representing background to generate a background corrected spectrum, the method for removal of signal values representing background comprising calculating fit values for each of said signals, said fit values being generated by the steps of:
(a) calculating function values according to at least three functions of said signal values, wherein said at least three functions are selected from:
(A) mean;
(B) variance;
(C) skewness;
(D) kurtosis;
(E) moving average with a window width of 3 points;
(F) second derivative;
(G) median;
(H) maximum;
(I) minimum;
(J) difference between maximum and minimum;
(K) quotient of the central signal value and the mean value;
(L) quotient (maximum-mean value)/(maximum-minimum);
(b) normalizing said calculated function values by division by the mean value of the corresponding function values over all of said spectral points;
(c) attributing probabilities Pi(band) for the presence of bands or peaks to each of said spectral points in each of said calculated function values wherein said probabilities Pi(band) are calculated by attributing a probability to each of said spectral points having a normalized value above or below a threshold, in particular above 1.3 and below 0.7, said attributed probability being proportional to the distance of said normalized value from said threshold;
(d) adding said probabilities Pi(band) up to an overall probability ΣPi(band) from all of said calculated function values for each of said spectral points;
(e) calculating a probability P(background) for the presence of background for each of said spectral points from said overall probability ΣPi(band) according to
P(background)=1−ΣPi(band)
wherein negative values of P(background) are set to zero;
and
(f) calculating a fit of said signal values with fit values at all of said spectral points wherein said signal values in each of said spectral points is taken into account in the fit values only with said background probability calculated for said spectral point, and
subtracting said fit values from said signal values of said spectrum thereby generating a background corrected spectrum;
making said background corrected spectrum available for spectral analysis in place of said spectrum.

23 more elements

A method for improving spectral analysis by generating a background corrected spectrum for use in spectral analysis, comprising:

obtaining a spectrum consisting of signals of a plurality of spectral points, said signals represented by signal values of respective points in said spectrum, some signal values in said spectrum representing background;
and
determining the signal values in said spectrum representing background comprising the steps of:
(a) selecting at least three functions to perform on the spectrum, wherein said at least three functions are selected from:
(A) mean;
(B) variance;
(C) skewness;
(D) kurtosis;
(E) moving average with a window width of 3 points;
(F) second derivative;
(G) median;
(H) maximum;
(I) minimum;
(J) difference between maximum and minimum;
(K) quotient of the central signal value and the mean value;
(L) quotient (maximum-mean value)/(maximum-minimum);
(b) selecting a window width and calculating function values using the selected window width for the spectrum according to each of the at least three selected functions;
(c) normalizing said calculated function values by division by the mean value of the corresponding function values over all of said spectral points;
(d) attributing probabilities Pi(band) for the presence of bands or peaks to each of said spectral points in each of said calculated function values wherein said probabilities Pi(band) are calculated by attributing a probability to each of said spectral points having a normalized value above or below a threshold, in particular above 1.3 and below 0.7, said attributed probability being proportional to the distance of said normalized value from said threshold;
(e) adding said probabilities Pi(band) up to an overall probability ΣPi(band) from all of said calculated function values for each of said spectral points;
(f) calculating a probability P(background) for the presence of background for each of said spectral points from said overall probability ΣPi(band) according to
P(background)=1−ΣPi(band)
wherein negative values of P(background) are set to zero;
and
(g) calculating a fit of said signal values with fit values at all of said spectral points wherein said signal values in each of said spectral points is taken into account in the fit values only with said background probability calculated for said spectral point;
using the calculated fit values, subtracting said fit values from said signal values of said spectrum thereby removing said signal values representing background from said spectrum to generate a background corrected spectrum;
and
making said background corrected spectrum available for spectral analysis in place of said spectrum.

25 more elements

Dependencies are inferred from the claim text.

Facts

Application: 15/093,398Filed 7 Apr 2016
Priority: DE 10 2015 105 2397 Apr 2015
Published: US 2016/0299004 A113 Oct 2016
Granted: 1 Sep 2020
Term ends: Not estimatedThe source records don’t support an estimate.

Classification7 codes

G01J 3/0297
Details: constructional arrangements for removing other types of optical noise or for performing calibration
G01J 3/28
Spectrometry; Spectrophotometry; Monochromators; Measuring colours: investigating the spectrum
H01J 49/022
Details: circuit arrangements, e.g. for generating deviation currents or voltages ; Components associated with high voltage supply
G01J 2003/2859
Investigating the spectrum: peak detecting in spectrum