CN109883549B

CN109883549B - Digital micromirror-based bending spectral line correction method

Info

Publication number: CN109883549B
Application number: CN201910162585.XA
Authority: CN
Inventors: 田地; 王宏霞; 李春生
Original assignee: Jilin University
Current assignee: Jilin University
Priority date: 2019-03-05
Filing date: 2019-03-05
Publication date: 2021-03-12
Anticipated expiration: 2039-03-05
Also published as: CN109883549A

Abstract

The invention relates to a method for correcting a bending spectral line based on a digital micromirror, which comprises the following steps: obtaining three-dimensional full spectrum data; determining the range of each bending spectral line in the spectrogram; acquiring the spectral center position of each line of each curved spectral line and the corresponding signal intensity by adopting a light intensity value weighting method; determining a fittable range; step five, fitting a relational expression of positions and wavelengths of each row of pixels in a fittable range to obtain wavelength values corresponding to all the pixels; calculating the corresponding wavelength value of the pixel of the part which cannot be fitted; and the bending and overturning detection of the digital micromirror is realized. The invention does not need to consider error terms caused by installation and adjustment errors and environmental errors, and can improve the signal-to-noise ratio of the measurement spectral line.

Description

Digital micromirror-based bending spectral line correction method

Technical Field

The invention belongs to the technical field of spectral analysis, and particularly relates to a method for correcting a bending spectral line based on a digital micromirror.

Background

In a spectral analysis instrument, it is a very common method to realize the spatial separation of polychromatic light by using a diffraction grating, wherein an incident slit of a spectrometer is perpendicular to a main cross section of the grating, a principal ray at the center point of the slit is superposed with the main cross section of the grating, and principal rays at other points of the slit form included angles with the main cross section of the grating, thereby causing the bending of spectral lines. Spectral resolution, luminous flux and signal-to-noise ratio are affected by the presence of line bending phenomena. The existing method for correcting the spectral line bending starts from the aspect of optical design, the bending spectral line correction method based on data processing is based on the basic optical formula to carry out complex formula derivation calculation, the calculation process is complicated, and the theoretical derivation is easy to be influenced by the environment to cause deviation. In the method for actually controlling the digital micromirror, the Hadamard template and the array (whole column) overturning mode are mainly used, and the Hadamard template mainly aims at a continuous light source; column inversion, although the simplest control method, affects the accuracy and signal-to-noise ratio of the light intensity signal with respect to the line bending phenomenon. At present, no method for detecting the turning of the bending track of the digital micromirror is available, and the detection method can be used for accurately and rapidly correcting the bent spectral line.

The invention discloses an imaging spectrometer system capable of correcting spectral line bending and a correction method thereof (application publication No. CN 104034419), a spectral element of a grating-prism-grating is designed, the spectral line bending of the center wavelength is compensated and corrected by utilizing the characteristic that the dispersion of the grating and the prism is opposite to the spectral line bending direction, and the residual spectral line bending is compensated and corrected by utilizing the distortion and the image plane inclination generated by a collimating objective and an imaging objective, so that the calibration of the spectral line bending of the full working spectral line is realized.

The invention discloses a spectral line bending long-wave infrared plane grating imaging spectral system (application publication number: CN 103048045A), which utilizes an off-axis lens to eliminate spectral line bending generated by a grating.

The invention discloses a method and a device for correcting spectral line bending of a dispersion type spectral imager (application publication No. CN 104316183A), which carry out spectral line bending correction by utilizing a related wavelength Euclidean distance calculation result and an inverse distance weight method.

The articles "spectral aliasing analysis and correction of hadamard transform spectral imager" (zhou brocade, luqun wave, phase in. "spectral aliasing analysis and correction of hadamard transform spectral imager [ J ] photonics, 2005(10): 1518-.

The paper "analysis and correction of spectral line bending for digital micromirror Hadamard spectrometer" ("analysis and correction of spectral line bending for digital micromirror Hadamard spectrometer," [ J ]. Spectroscopy and Spectroscopy, ", 2016,36(02): 555-.

The above prior art has the following disadvantages: 1. the calculation formula derived from the theoretical formula of the optical system cannot be directly calculated by the formula to obtain a final result due to unavoidable installation and adjustment errors and the like; 2. parameters such as a plurality of angles involved in an optical formula are inconvenient to actually measure, so that the practical application is restricted, and therefore, the existing verification methods are all simulation verification. 3. What is detected in the array (column) flip control mode is local spectral information, resulting in a decrease in spectral signal-to-noise ratio.

Disclosure of Invention

The invention aims to provide a method for correcting a bending spectral line based on a digital micromirror, which does not need to consider error items caused by installation and adjustment errors and environmental errors and has high signal-to-noise ratio of a measured spectral line.

In order to solve the above technical problem, the method for correcting the bending spectral line based on the digital micromirror of the present invention comprises:

step one, obtaining three-dimensional full-spectrum data; controlling the digital micro-mirror array to scan the known spectral line wavelength lambda point by point or block by block₁，λ₂，···λ_n···，λ_NThe excitation light source of a certain element carries out full spectrum scanning to obtain full spectrum data;

step two, determining the range of each bending spectral line in a spectrogram; taking a multiple of a wavelength-free position Noise value Noise in a spectrogram as a signal intensity Limit value Limit, wherein the value of a is selected by taking the aim of extracting all spectrograms as far as possible and not extracting Noise as a target; defining the position of the signal intensity more than or equal to the Limit value in the spectrogram as a spectral line position, and further extracting N curved spectral lines;

thirdly, acquiring the spectral center position of each line of the N curved spectral lines and the corresponding signal intensity by adopting a light intensity value weighting method; for the mth row of the nth curved spectral line, the abscissa of the pixel at the center position of the transverse spectrum

Wherein x_nm-iIs the abscissa of the ith pixel element of the mth line of the nth curved spectral line, I_nm-iThe signal intensity corresponding to the ith pixel of the mth line of the nth curved spectral line; i is_nmThe number of pixel columns contained in the mth row of the nth curved spectral line is set;

step four, sequencing the longitudinal upper limits of the N bending spectral lines from high to low, and sequencing the longitudinal lower limits of the N bending spectral lines from low to high; determining all action fittable ranges at least comprising two curved spectral line row pixels in one row;

step five, according to the theoretically known spectral line wavelength lambda₁，λ₂，···λ_n···，λ_NWith respect to the position of the pixel columnThe method comprises the steps of solving the corresponding relation between the center position of each line spectrum in each curved spectral line and the wavelength in a fitting range, fitting a relation between the position of each line pixel and the wavelength in the fitting range, and obtaining the wavelength value corresponding to all pixels in the fitting range according to the relation between the position of each line pixel and the wavelength;

sixthly, calculating the wavelength value corresponding to the part of pixels which cannot be fitted; finding curved lines n containing non-fittable parts_bCalculating the curved spectral line n by using a light intensity value weighting method_bOrdinate of central pixel of longitudinal spectrum

Wherein y is_nb-kRepresenting curved spectral lines n_bThe longitudinal coordinate of the central pixel of the transverse spectrum of the nth b-K line, and nb-K is a bending spectral line n_bTotal number of lines of (1)_nb-kIs the signal strength of the pixel; obtaining a bending spectral line n by utilizing the relation between all line pixels and the wavelength in the fitting range fitted in the step five_bThe wavelength values corresponding to all the pixels of the row where the center position of the longitudinal spectrum is located; using the central position of transverse spectrum as the alignment reference and using the curved spectral line n_bThe wavelength value corresponding to the row pixel where the longitudinal spectrum center position is located completely fills the non-fittable row pixels;

step seven, realizing the bending and overturning detection of the digital micromirror; when the element detection is carried out on the sample, all pixels corresponding to the wavelength value to be detected are searched by using the wavelength value corresponding to each pixel obtained in the fifth step and the sixth step, and the digital micromirrors in the pixels are controlled to simultaneously turn over, so that the measurement of the bending spectral line of the element to be detected of the sample is realized.

The invention has the advantages that: 1. correcting a bending spectral line caused by aberration of an optical element in an optical path of the analysis instrument by controlling bending and overturning of the digital micromirror; 2. compared with the method for correcting the bending spectral line through light path design, the method has the advantages that the bending control is carried out according to real-time detection results without considering error items caused by installation and adjustment errors and environmental errors; 3. by bending the scanning lines, the signal-to-noise ratio of the measured lines can be improved compared to scanning in an array.

Drawings

The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.

FIG. 1 is a schematic diagram of three-dimensional full-spectrum X-Y view angle curved spectral line detection.

FIG. 2 is a partial enlarged view of 306.80nm wavelength bend line A of the Bi element.

FIG. 3 is a schematic diagram showing wavelength values corresponding to pixels at the longitudinal scanning center position of an 306.80nm wavelength bending spectral line of a Bi element.

FIG. 4 is a diagram showing the correspondence between the unadapted part I at the upper end of the 306.80nm wavelength bending spectral line of the Bi element and the line wavelength value at the longitudinal scan center position.

FIG. 5 is a diagram showing the correspondence between the line wavelength value at the line position of the lower unfixed portion II of the 306.80nm wavelength bending spectral line of the Bi element and the longitudinal scan center position.

Fig. 6 is a flow chart of the present invention.

Detailed Description

As shown in fig. 6, the method for correcting the bending spectral line based on the digital micromirror of the present invention is as follows:

step one, obtaining three-dimensional full spectrum data: as shown in fig. 1, taking a digital micromirror with a specification of 1024 × 768 as an example, controlling a digital micromirror array to perform full-spectrum scanning on an excitation light source of a known element in a point-by-point or block-by-block scanning manner to obtain full-spectrum data; the full spectrum data includes the signal strength corresponding to each pixel (1 or more digital micromirrors may be included in a pixel); wherein the element corresponds to a spectral line wavelength lambda₁，λ₂，···λ_n···，λ_N(ii) a The signal intensity is represented by light color, and the darker the color, the stronger the signal intensity;

step two, determining the range of each bending spectral line in the three-dimensional spectrogram obtained in the step one (namely the abscissa range and the ordinate of a pixel corresponding to each bending spectral line): taking a multiple of a wavelength-free position Noise value Noise in a spectrogram as a signal intensity Limit value Limit, namely, the Limit is a multiplied by Noise, and the value of a is selected by taking the aim of extracting all spectrograms as far as possible and not extracting Noise as a target; the position of the signal intensity which is more than or equal to the Limit value in the spectrogram is specified as the spectral line position, and then the spectral line position is obtainedTaking out N bending spectral lines (corresponding wavelength lambda of N bending spectral lines)₁，λ₂，···，λ_NKnown) corresponding pixel coordinate ranges;

step three, acquiring the spectral center position and the corresponding signal intensity of each line of the N curved spectral lines: because the spectral imaging has nonuniformity, a light intensity value weighting method is needed for calculating the transverse spectral center position of each line of the curved spectral line; for the mth row of the nth curved spectral line, the abscissa of the pixel at the center position of the transverse spectrum

determining the range of the relational expression between the pixel positions and the wavelengths of each row in the fittable spectrogram; because the heights of the curved spectral lines are different in the longitudinal direction, some lines on a spectrogram only have one curved spectral line pixel position, and fitting cannot be performed; therefore, the longitudinal upper limits of the N bending spectral lines need to be sequenced from high to low, and similarly, the longitudinal lower limits of the N bending spectral lines need to be sequenced from low to high; determining all action fittable ranges including at least two bending spectral line pixel positions in a row; as shown in fig. 1, the ordinate is Y_T～Y_BAll rows in the range are fitting ranges;

step five, according to the theoretically known spectral line wavelength lambda₁，λ₂，···λ_n···，λ_NThe wavelength lambda of the spectral line is measured in a fitting range according to the position of the pixel array₁，λ₂，···λ_n···，λ_NAnd respectively assigning pixels to the transverse spectrum centers of each row of the corresponding curved spectral line, and fitting a relation between the positions and the wavelengths of each row of pixels in a fittable range according to the pixel positions: if the fittable range includes J rows of pixels, the relationship between the pixels in the first row to the J-th row and the wavelength is as follows: lambda [ alpha ]¹＝f₁(X)，λ²＝f₂(X)···λ^j＝f_j(X)…λ^J＝f_J(X), so that the corresponding wavelength values of all the pixels in the fitting range can be obtained; as shown in fig. 2, the pixel at which the wavelength of 306.8nm is located is the central pixel of the transverse spectrum of the 306.80nm wavelength bending spectral line of the Bi element;

sixthly, calculating the wavelength value corresponding to the part of pixels which cannot be fitted; finding curved lines n containing non-fittable parts_b(ii) a The light intensity value weighting method is also used for calculating the curved spectral line n_bOrdinate of central pixel of longitudinal spectrum

Wherein y is_nb-kRepresenting curved spectral lines n_bOrdinate, I, of center pixel of transverse spectrum of the nth b-k line_nb-kIs the signal strength of the pixel; obtaining a bending spectral line n by utilizing the relation between all line pixels and the wavelength in the fitting range fitted in the step five_bThe wavelength values corresponding to all the pixels of the row where the center position of the longitudinal spectrum is located; using the central position of transverse spectrum as the alignment reference and using the curved spectral line n_bThe wavelength value corresponding to the row pixel where the longitudinal spectrum center position is located completely fills the non-fittable row pixels; taking FIG. 1 as an example, the curved spectral line n_bIs the curved line N in fig. 1; as shown in fig. 3, the curved line n_bLongitudinal spectrum center pixel line pixel (X)_O-3,Y_nb-O),(X_O-2,Y_nb-O),(X_O-1,Y_nb-O),(X_O,Y_nb-O),(X_O+1,Y_nb-O),(X_O+2,Y_nb-O),(X_O+3,Y_nb-O) The corresponding wavelength values are 306.2nm, 306.4nm,306.6nm, 306.8nm, 307.0nm,307.2nm and 307.4nm respectively; as shown in FIG. 4, the picture element (X)_O,Y_nb-O) A wavelength value of 306.8nm, is copied and filled into the curved spectral line n_bPixel (X) of the upper unfit part I_nbT1,Y_nbT1) Pixel (X)_O,Y_nb-O),(X_O+1,Y_nb-O) The wavelength values of 306.8nm and 307.0nm are respectively copied and filled in the bending spectral line n_bPixel (X) of the upper unfit part I_nbT1,Y_nbT1)，(X_nbT2,Y_nbT2),(X_nbT2+1,Y_nbT2) (ii) a As shown in FIG. 5, the picture element (X)_O-1,Y_nb-O),(X_O,Y_nb-O),(X_O+1,Y_nb-O) Respectively 306.6nm, 306.8nm and 307.0, are copied and filled in the bending spectral line n_bThe lower extreme does not fit the picture element (X) of part II_nbB4-1,Y_nbB4),(X_nbB4,Y_nbB4),(X_nbB4+1,Y_nbB4) Let pixel (X)_O,Y_nb-O),(X_O+1,Y_nb-O) 306.8nm, 307.0nm copied to fill the bend line n_bThe lower extreme does not fit the picture element (X) of part II_nbB3,Y_nbB3),(X_nbB3+1,Y_nbB3) (ii) a Pixel (X)_O,Y_nb-O) The value of 306.8nmm of the wavelength is copied and filled into the curved spectral line n_bThe lower extreme does not fit the picture element (X) of part II_nbB2,Y_nbB2) (ii) a Finally, the pixel (X)_O,Y_nb-O) The value of 306.8nmm of the wavelength is copied and filled into the curved spectral line n_bThe lower extreme does not fit the picture element (X) of part II_nbB1,Y_nbB1)；

Claims

1. A method for correcting bending spectral lines based on a digital micromirror is characterized by comprising the following steps:

step one, obtaining three-dimensional full-spectrum data; controlling the digital micro-mirror array to scan the known spectral line wavelength lambda point by point or block by block₁，λ₂，…λ_n…，λ_NThe excitation light source of a certain element carries out full spectrum scanning to obtain full spectrum data;

step five, according to the theoretically known spectral line wavelength lambda₁，λ₂，…λ_n…，λ_NThe corresponding relation between the center position of each line spectrum in each bending spectral line and the wavelength in the fitting range is worked out according to the relation between the center position of each line spectrum and the position of the wavelength in each bending spectral line in the fitting range, the relation between the position of each line pixel in the fitting range and the wavelength is fitted, and the wavelength value corresponding to all the pixels in the fitting range can be obtained according to the relation between the position of each line pixel and the wavelength;

Wherein y is_nb-kRepresenting curved spectral lines n_bThe longitudinal coordinate of the central pixel of the transverse spectrum of the nth b-K line, and nb-K is a bending spectral line n_bTotal number of lines of (1)_nb-kIs the signal strength of the pixel; by using stepsObtaining a bending spectral line n by the relation between all line pixels and the wavelength in the fitting range fitted in the fifth step_bThe wavelength values corresponding to all the pixels of the row where the center position of the longitudinal spectrum is located; using the central position of transverse spectrum as the alignment reference and using the curved spectral line n_bThe wavelength value corresponding to the row pixel where the longitudinal spectrum center position is located completely fills the non-fittable row pixels;