Iterative X-Ray Background Fitting with Combined Algorithms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining background spectra in X-ray spectroscopy, such as the SNIP algorithm and polynomial fitting, often fail to accurately estimate background radiation, especially in the presence of broad continuum features, leading to inaccurate peak intensity calculations and subsequent quantitative analysis.
Innovation Solution
A method involving multiple iterations of different background algorithms, such as SNIP and iteratively fitted orthogonal polynomials, to determine a combined background spectrum, which is then used in a deconvolution model to extract net intensities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the SNIP algorithm is used to estimate the background spectrum, then the calculation is fast and easy to implement, but the background estimate becomes inaccurate in the presence of broad continuum background features
Solution Approach 1:
The patent combines multiple background estimation algorithms (SNIP, orthogonal polynomials, and other algorithms) to create a composite background estimate. This merging approach allows the system to leverage the speed of SNIP while compensating for its inaccuracies with polynomial fitting, thereby resolving the contradiction between calculation speed and estimation accuracy.
Solution Approach 2:
The patent creates a composite background estimation method by integrating results from different algorithms. Similar to how composite materials combine properties of individual materials, this composite approach combines the advantages of fast calculation (from SNIP) with accurate background modeling (from polynomial fitting), achieving both speed and precision.
2Measurement precision
If iteratively fitted orthogonal polynomials are used to estimate the background radiation, then the background estimation can be improved, but significant fluctuations occur particularly if higher order polynomials are applied
Solution Approach 1:
The patent dynamically adjusts the polynomial order parameter based on the characteristics of the spectrum being analyzed. By changing this parameter adaptively rather than using a fixed high order, the system achieves accurate background estimation without the instability and fluctuations that occur with consistently high-order polynomials.
Solution Approach 2:
The patent applies polynomial fitting selectively and partially - using it to complement rather than replace the SNIP algorithm entirely. This partial application avoids the excessive action of applying high-order polynomials across the entire spectrum, thereby reducing fluctuations while maintaining accuracy where polynomial fitting is most beneficial.
3Device complexity
If a single background algorithm is used, then the processing is simple, but the accuracy of peak intensity extraction is insufficient
Solution Approach 1:
The patent segments the background estimation process into multiple independent algorithmic components (SNIP, orthogonal polynomials, and other algorithms). Each segment handles specific aspects of background estimation, and their results are combined. This segmentation allows the system to achieve high accuracy through multiple specialized approaches rather than one complex monolithic algorithm.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
There is provided a method of analysis of detected X-ray spectra in an instrument, comprising applying a first iteration of a first background algorithm to the detected X-ray spectrum to determine a first partial background spectrum, subtracting the first partial background spectrum from the detected X-ray spectrum to form a first net spectrum, applying a second iteration of a second background algorithm to the first net spectrum to determine a second partial background spectrum wherein the second background algorithm is different from the first background algorithm, forming an estimated background spectrum comprising linear combination of the first partial background spectrum and the second partial background spectrum and using the estimated background spectrum as an input into a deconvolution model.