AU2021107411B4

AU2021107411B4 - Method of Addressing Alternating Column Noise

Info

Publication number: AU2021107411B4
Application number: AU2021107411A
Authority: AU
Inventors: Mark Bigland; Jamie Flynn; Brenton Gray; Antony Martin; William Palmer
Original assignee: Rapid Phenotyping Pty Ltd
Current assignee: Rapid Phenotyping Pty Ltd
Priority date: 2021-08-25
Filing date: 2021-08-25
Publication date: 2022-10-20
Anticipated expiration: 2029-08-25
Also published as: AU2021107411A4

Abstract

The present invention relates to a method of reducing alternating column noise in a set of data points received from a first detector array and a second detector array, wherein each data point comprises an output at a particular value of a given independent variable, the method comprising the steps of separating the set into a first set comprising first data points received from the first detector array, and a second set comprising second data points received from the second detector array, determining an array of values of the given independent variable, constructing a first adjusted set comprising first adjusted data points by calculating, for each value in the array of values, a corresponding first interpolated output by interpolating the first data points in the first set, constructing a second adjusted set second adjusted data points by calculating, for each value in the array of values, a corresponding second interpolated output by interpolating the second data points in the second set, and constructing an adjusted complete set by calculating, for each value in the array of values, a corresponding averaged output by averaging the first interpolated output and the second interpolated output for the respective value.

Description

METHODOFADDRESSING ALTERNATINGCOLUMN NOISE TECHNICAL FIELD

[0001] The present invention relates generally to the field of spectroscopy, and more particularly to the field of spectroscopy using dual detectors

BACKGROUND

[0002] There is a common noise affect seen in spectroscopic data, referred to as an odd-even effect or alternating column noise (ACN). The source of the noise is due to having two overlapping/alternating pixel detector strips or columns, that are referred to as the odd and even pixels. These odd and even pixels also have their own shift register and/or amplifier. The resulting effect seen between the odd and even pixels is a slight displacement shift that produces a zigzag or underlying sawtooth like structure. Figure 1 is a spectrum of a data set containing raw data from both an odd detector array and an even detector array.

[0003] Most methods to deal with the ACN require an intensive calibration process that looks at the differences in interpolated pixels, using the average or median difference over the entirety of the pixels from both detectors and over a range of intensities with different integration times. This difference is then applied as an offset to all detector pixels from one column (e.g., odd) and in reference to the raw stationary column (e.g., even). Although this averaged offset that is applied to all of one of the columns pixels reduces the effect of ACN, there is still a noticeable amount of ACN that can be seen, especially at higher intensities and with samples that produce a fluorescent background.

[0004] Overall, prior art methods of removing ACN are calibration techniques that are applied at a 'firmware' level, i.e. they are broadly and consistently applied to all throughput, rather than being developed with specific regard to the particular data set being processed, and so cannot account for dataset-specific noise. Otherwise, prior art methods attempt to execute complex interpolation directly between the two data sets, thereby negating the entire purpose of having two separate detector arrays. Figure 2 is a spectrum of a data set that has been processed in such a way to attempt to reduce noise. As can be seen, a significant 'sawtooth' effect is still present, meaning the data set is still substantially noisy.

DESCRIPTION OF FIGURES

[0005] Embodiments of the present invention will now be described in relation to figures, wherein: Figure 1 is a spectrum of raw data received from a first and second detector array; Figure 2 is the spectrum of Figure 1, having been processed to reduce noise through a prior art direct interpolation method; and Figure 3 is the spectrum of Figure 1, having been processed through the method of the present invention.

DETAILED DESCRIPTIONOF PREFERRED EMBODIMENTS

[0006] In a broad first aspect, the method of the present invention comprises separating the signals from the odd and even detector arrays, producing a spectrum created from the odd detector array data and a separate spectrum created from the even detector array data. the two spectra are then individually interpolated to have matching independent variable values, followed by averaging the two spectra at each of the independent variable values.

[0007] It is considered that the method of the present invention may allow both detector columns (odd and even) to have an effect on the spectrum that is produced, while also giving a smooth and more accurate spectrum. The additional benefits of this particular method to remove ACN are that it is specific to the light source being used for excitation, as well as any power settings, integration time and averaging that is being performed. Moreover, this method accounts for the ACN for the specific sample type. This can be contrasted against the prior art 'firmware level' calibration method and the prior art direct inter data-set interpolation methods of attempting to reduce or ameliorate ACN.

[0008] In an embodiment, the present invention provides a method of reducing alternating column noise in a set of data points received from a first detector array and a second detector array, wherein each data point comprises an output at a particular value of a given independent variable. In an embodiment, the method comprises the steps of: (1) separating the set into a first set comprising first data points received from the first detector array, and a second set comprising second data points received from the second detector array; (2) determining an array of values of the given independent variable; (3) constructing a first adjusted set comprising first adjusted data points by calculating, for each value in the array of values, a corresponding first interpolated output by interpolating the first data points in the first set; (4) constructing a second adjusted set comprising second adjusted data points by calculating, for each value in the array of values, a corresponding second interpolated output by interpolating the second data points in the second set; and (5) Constructing a combined set by calculating, for each value in the array of values, a corresponding averaged output by averaging the first interpolated output and the second interpolated output for the respective value.

[0009] In an embodiment, the independent variable may be any appropriate independent variable, variation of which leads to variation in the output. In an embodiment and as an example, the independent variable may be a wavelength that is recorded by a particular pixel in the array. In an embodiment and as an alternate example, the independent variable may be a wavenumber associated with a particular pixel in the array.

[0010] In an embodiment, the output may be the dependent variable, or otherwise the property of interest being measured and detected by the first and second detector arrays. In an embodiment and as an example, the output may be a measured intensity value.

[0011] The actual output set of the first and second detector arrays is a combination of the outputs of the first detector array and the second detector array. Although the outputs of each array could be utilised separately, this would effectively halve the resolution of the collected data, and so it is desirable to find a way to combine the outputs into a combined set. It is also advantageous to have a final combined set that comprises data points that are at known values of the given independent variable, such as at equally spaced values. This promotes ease of applying transformations and other algorithms to the data set, which often rely upon the array of outputs that are being fed into the algorithm having associated independent variable values that are, for example, regularly spaced. Therefore, it is beneficial to develop an array of values of the given independent variable for use in constructing the first and second adjusted sets as well as the combined set, as the values within the array of values may be selected to be best suited for future processing.

[0012] The first and second adjusted sets are constructed independently of one another by applying an interpolation algorithm, which involves feeding in the determined array of values and one of the first and second sets. With regard to construction of the first adjusted set, determining a first interpolated output at a particular value within the array of values is achieved by identifying appropriate data points within the first set and interpolating therebetween. This process is repeated until a first adjusted set is constructed, comprising adjusted data points associated with each value within the array of values. The same process is then applied to construct the second adjusted set, by feeding the determined array of values and the second set into the interpolation algorithm.

[0013] The first and second adjusted sets so constructed comprise matching values of the independent variable, but have differing and unique outputs (as calculated through the interpolation algorithm) at each value.

[0014] The skilled person will appreciate that various interpolation methods known in the art may be employed in construction of the interpolation algorithm. For example, one of a linear interpolation method, a polynomial interpolation method, or a spline interpolation method may be employed. Alternatively, other interpolation methods such as Gaussian interpolation, rational interpolation or trigonometric interpolation may be employed. The skilled person will appreciate that the most appropriate interpolation method will depend upon the type of data being processed, and selection of a particular interpolation method over another does not fall outside of the scope of the present invention.

[0015] The combined set is constructed by averaging, for each value in the array of values, the first interpolated output and the second interpolated output associated with that particular value. It is considered that using an averaging algorithm enables both the first set and the second set to affect the combined set, in that constructing the combined set through averaging means that variations between the first and second sets will be blunted, while preserving and potentially clarifying any trends that may be otherwise obscured within the collected data. The skilled person will appreciate that various averaging methods known in the art may be employed in construction of the averaging algorithm. For example, calculating the arithmetic mean may be most appropriate for a set receiving outputs from a first and second detector array. However, if the method is applied to produce a combined set from more than two detector arrays, other averaging methods - such as calculation of a geometric mean, a harmonic mean, a quadratic mean, a weighted mean or any other method known in the art - may be better suited to the data being processed.

[0016] Referring now to Figure 3, shown is the same spectrum as in Figure 1, processed through the above-disclosed method of the present invention comprising separating the data, independently interpolating each set, and developing a combined and averaged set. As depicted, the 'sawtooth' noise characteristic that is visible in Figure 2 (processed through a prior art noise reduction method) is substantially absent, while trends that can be seen in Figure 1 remain visible.

[0017] While the invention has been described with reference to preferred embodiments above, it will be appreciated by those skilled in the art that it is not limited to those embodiments, but may be embodied in many other forms, variations and modifications other than those specifically described. The invention includes all such variation and modifications. The invention also includes all of the steps, features, components and/or devices referred to or indicated in the specification, individually or collectively and any and all combinations or any two or more of the steps or features.

[0018] In this specification, unless the context clearly indicates otherwise, the word "comprising" is not intended to have the exclusive meaning of the word such as "consisting only of", but rather has the non-exclusive meaning, in the sense of "including at least". The same applies, with corresponding grammatical changes, to other forms of the word such as "comprise", etc.

[0019] Other definitions for selected terms used herein may be found within the detailed description of the invention and apply throughout. Unless otherwise defined, all other scientific and technical terms used herein have the same meaning as commonly understood to one of ordinary skill in the art to which the invention belongs.

[0020] Any promises made in the present document should be understood to relate to some embodiments of the invention, and are not intended to be promises made about the invention in all embodiments. Where there are promises that are deemed to apply to all embodiments of the invention, the applicant/patentee reserves the right to later delete them from the description and they do not rely on these promises for the acceptance or subsequent grant of a patent in any country.

Claims

CLAIM

1) A method of reducing alternating column noise in a set of data points received from a first detector array and a second detector array, wherein each data point comprises an output at a particular value of a given independent variable, the method comprising the steps of: (I) separating the set into a first set comprising first data points received from the first detector array, and a second set comprising second data points received from the second detector array; (II) determining an array of values of the given independent variable; (III) constructing a first adjusted set comprising first adjusted data points by calculating, for each value in the array of values, a corresponding first interpolated output by interpolating the first data points in the first set; (IV) constructing a second adjusted set second adjusted data points by calculating, for each value in the array of values, a corresponding second interpolated output by interpolating the second data points in the second set; and (V) constructing an adjusted complete set by calculating, for each value in the array of values, a corresponding averaged output by averaging the first interpolated output and the second interpolated output for the respective value.

2) The method of claim 1 wherein interpolating comprises applying feeding the array of values and one of the first set and second set into an interpolation algorithm; and the interpolation algorithm, for a particular value within the array of values, identifies appropriate data points within the set that was fed thereinto, and interpolating therebetween to determine the interpolated output for that particular value.

3) The method of claim 2 wherein identifying appropriate data points comprises identifying at least one data point having a value lower than the particular value and identifying at least one data point having a value higher than the particular value.