Estimation of water interference for spectral correction
Inventors
Judge, Kevin • Andersson, Greger • Zou, Peng
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-10101269-B2
Publication Date
2018-10-16
Expiration Date
Abstract
A method includes decomposing a training set to obtain a principal component matrix having a plurality of principal component vectors. The method also includes variably rejecting portions of a sample spectrum vector that do not correspond to a selected one of the plurality of principal component vectors by incrementally providing a coefficient indicative of the weighting of the selected principal component vector for selected sub-regions. A corrected spectrum vector can be obtained by excluding certain sub-regions of the sample spectrum vector and corresponding principal component vector, multiplying the sample spectrum vector with the principal component matrix for non-excluded sub-regions, providing a predicted interference vector, and subtracting the predicted interference vector from the sample spectrum vector.
Core Innovation
The invention corrects infrared spectrometry vapor-phase spectral data for water vapor interference by decomposing a training set corresponding to spectral data obtained for an interfering substance using an infrared spectrometry system to obtain a principal component matrix having a plurality of principal component vectors. A sample spectrum vector for a sample mixture is generated using the infrared spectrometry system, and the method estimates an interference contribution that can mask a presence or an absence of one or more chemicals of interest in the sample mixture.
The method variably rejects portions of the sample spectrum vector that do not correspond to a selected one of the plurality of principal component vectors by incrementally selecting a sub-region of the sample spectrum vector and a corresponding sub-region of the selected principal component vector, and multiplying the selected sub-region of the sample spectrum vector with the corresponding sub-region of the selected principal component vector to provide a coefficient indicative of the weighting of the selected principal component vector for the selected sub-regions. Sub-regions are excluded based on the incrementally provided coefficients, and a predicted interference vector is produced by projecting the non-excluded sub-regions through the principal component matrix.
After subtracting the predicted interference vector from the sample spectrum vector to provide a corrected spectrum vector, the corrected spectrum vector is compared to known spectrum data for one or more chemicals of interest to determine if the corrected spectrum vector is representative of a chemical of interest. In addition, one or more visual or audible indicia can be activated when the corrected spectrum is representative of a chemical of interest, using PCA-based modeling including incremental sliding sub-regions, coefficient-based outlier handling, and recomputed weighting over non-excluded sub-regions.
Claims Coverage
The independent claims are clm-00001, clm-00008, and clm-00015, together covering three inventive features: PCA-based correction and comparison, optional visual or audible indicia activation, and a computer-readable storage medium implementing the same subject matter.
Principal component matrix from interfering substance training set
Decomposing a training set corresponding to spectral data obtained for an interfering substance using an infrared spectrometry system to obtain a principal component matrix having a plurality of principal component vectors.
Incremental sub-region coefficient-based variably rejecting
Variably rejecting portions of the sample spectrum vector that do not correspond to a selected one of the plurality of principal component vectors by incrementally selecting a sub-region of the sample spectrum vector and a corresponding sub-region of the selected principal component vector; and multiplying the selected sub-region of the sample spectrum vector with the corresponding sub-region of the selected principal component vector to provide a coefficient indicative of the weighting of the selected principal component vector for the selected sub-regions.
Predicted interference vector via non-excluded principal component matrix projection
Excluding sub-regions of the sample spectrum vector and corresponding principal component vector based on the incrementally provided coefficients; multiplying the sample spectrum vector with the principal component matrix for the non-excluded sub-regions to provide a weighting vector indicative of the contribution of the principal component matrix; multiplying the weighting vector by the principal component matrix to provide a predicted interference vector, the predicted interference vector corresponding to the interfering substance in the sample mixture.
Corrected spectrum vector generation and comparison to known chemical spectra
Subtracting the predicted interference vector from the sample spectrum vector to provide a corrected spectrum vector; and comparing the corrected spectrum vector to known spectrum data for one or more chemicals of interest to determine if the corrected spectrum vector is representative of a chemical of interest.
Visual or audible indicia activation upon chemical representativeness
Activating one or more visual or audible indicia when the corrected spectrum is representative of a chemical of interest.
Computer-readable storage medium implementing PCA correction and comparison
A computer-readable storage medium comprising code, the code comprising decomposing a training set to obtain a principal component matrix, receiving a sample spectrum vector, variably rejecting portions of the sample spectrum vector via incrementally computed sub-region coefficients, computing a predicted interference vector, subtracting it to provide a corrected spectrum vector, and comparing the corrected spectrum vector to known spectrum data for one or more chemicals of interest.
Across clm-00001, clm-00008, and clm-00015, the claims are directed to PCA-based correction of a sample spectrum by incrementally excluding sub-regions using coefficient-weighting relative to selected principal component vectors, computing a predicted interference vector corresponding to an interfering substance, subtracting it to form a corrected spectrum vector, and comparing the corrected spectrum vector to known spectrum data for one or more chemicals of interest, with optional visual or audible indicia activation.
Stated Advantages
Corrects infrared spectrometry vapor-phase spectral data for interference so that one or more chemicals of interest can be determined from the corrected spectrum vector.
Provides predicted interference subtraction based on principal component modeling derived from an interfering substance training set.
Can activate one or more visual or audible indicia when the corrected spectrum is representative of a chemical of interest.
Documented Applications
Infrared spectrometry vapor-phase spectroscopy to determine whether a corrected spectrum vector is representative of one or more chemicals of interest despite masking by an interfering substance.
Alarm/user interface use case where one or more visual or audible indicia are activated when a corrected spectrum indicates a chemical of interest.
Interested in licensing this patent?