CN106885631B - A kind of spectral method of detection based on digital projection - Google Patents

Abstract

The invention discloses a spectrum detection method based on digital projection, comprising: according to the pixel size of a photoelectric detector or a single-point detector of a spectrometer, designing an entrance part whose width changes in a gradient, so that the incident light spot is finally projected in a gradient On the photoelectric detector or single-point detector of the spectrometer; according to the number of horizontal projection pixels from few to many, a set of spectrograms with resolutions from high to low is extracted; spectrograms with different resolutions are input into the digital projection algorithm for iterative calculation , until the result converges; the converged result can be restored to the spectrum with the best resolution, that is, the spectrum when all the projections are concentrated on one pixel on the photodetector or single-point detector of the spectrometer. The invention overcomes the problem that the spectrum obtained by the traditional spectrometer cannot have high resolution and high signal-to-noise ratio at the same time by adding an entrance with a gradient width before the spectrum detection system and supporting data processing means in the later stage.

Description

A Spectral Detection Method Based on Digital Projection

technical field

The invention relates to the technical field of spectrum detection, in particular to a digital projection-based spectrum detection method.

Background technique

The spectrum is a pattern in which the dispersed monochromatic light is arranged sequentially according to the wavelength (or frequency) after the polychromatic light is split by the dispersion system. The light waves are generated by the electrons moving inside the atoms, and the electrons in the atoms of various substances move differently, so the light waves they emit are also different. Therefore, it is of great theoretical and practical significance to study the luminescence and light absorption of different substances.

At present, the spectral detection technology is relatively simple, and its structure mainly uses rectangular slits, and the resolution and signal-to-noise ratio of the spectrum are controlled by adjusting the width of the rectangular slits. The shape of the slit can be finally projected on the photodetector (CCD) of the spectrometer. When the projections are all concentrated on one pixel on the CCD, the spectral resolution is the best. The more the number of projected pixels, the worse the spectral resolution. Therefore, when the rectangular slit is narrow, the spectrum can have a higher resolution, but because the luminous flux becomes smaller, it will directly lead to a decrease in the spectral signal-to-noise ratio and affect the spectral quality; correspondingly, when the rectangular slit becomes wider, Although the increased luminous flux can relatively improve the signal-to-noise ratio, but the resolution will be reduced due to too many spectral projection pixels on the CCD. That is to say, traditional spectral detection technology cannot have high resolution and signal-to-noise ratio at the same time.

Contents of the invention

The invention provides a spectrum detection method based on digital projection. The invention overcomes the fact that the spectra acquired by traditional spectrometers cannot have high resolution at the same time by adding an entrance with a gradient width before the spectrum detection system and supporting data processing means in the later stage. and high signal-to-noise ratio, see the description below:

A kind of spectral detection method based on digital projection, described spectral detection method comprises the following steps:

According to the pixel size of the spectrometer photodetector or single-point detector, design an entrance part with a gradient width, so that the incident light spot is finally projected on the spectrometer photoelectric detector or single-point detector in a gradient;

Wherein, the entrance part whose width changes in a gradient is specifically: an optical fiber type and a slit type, and the slit type is divided into a fixed type and an adjustable type;

According to the number of horizontal projection pixels from less to more, a set of spectral images with resolution from high to low is extracted;

Input the spectrograms with different resolutions into the digital projection algorithm, and perform iterative calculations until the results converge;

The converged result is reverted to the best-resolution spectrum, that is, the spectrum when the projections are all centered on a single pixel on the spectrometer's photodetector or single-point detector.

Further, the fixed type is a diamond-shaped slit.

Wherein, the step of making the incident light spot finally gradiently projected on the spectrometer photodetector or single-point detector is specifically:

Design an optical path structure, the optical path structure includes: entrance part, collimating mirror, grating, converging mirror and spectrometer photodetector or single point detector,

The light is incident from the entrance part with a gradient width, diverges into parallel light through the collimator and hits the grating. After the grating is split, the dispersed monochromatic light is arranged in sequence according to the wavelength or frequency;

After converging through the objective lens, it is projected on the photoelectric detector or single-point detector of the spectrometer. When the projection is all concentrated on one pixel of the photoelectric detector or single-point detector of the spectrometer, the spectral resolution is the best. The more projection pixels, the better the spectral resolution. Difference.

The specific steps of inputting the spectrograms with different resolutions into the digital projection algorithm and performing iterative calculation until the results converge are as follows:

According to the digital projection algorithm, the input spectrum group is equal to the product of the projection matrix and the output spectrum;

Pre-estimate according to the input spectrum group, obtain an estimated spectrum, and perform projection transformation on the estimated spectrum; compare the projection transformation result with the input spectrum group, and use the ratio to correct the estimated spectrum;

Iterate repeatedly until the ratio is less than the threshold, at this time, the estimated spectrum meets the expected requirements, and the estimated spectrum at this time is output as the result.

Wherein, the step of comparing the projection transformation result with the input spectrum group and using the ratio to correct the estimated spectrum is specifically:

Among them, f _div (n) is the spectral deviation array; a _in is the column vector of the projection matrix; m is the number of rows of the projection matrix; g _div (n) is the projection deviation array; f(n) is the estimated spectral array; f'(n) is the last estimated spectral array; h _sum (n) is the projection specific gravity array.

The beneficial effect of the technical solution provided by the present invention is: the present invention acquires a group of spectrograms with different resolutions by adding an entrance with a gradient width before the spectral detection system, and the spectrum with the limit resolution of the spectrometer can be obtained through digital projection. This spectral detection method can not only simply and quickly overcome the problem that the spectra obtained by traditional spectrometers cannot have high resolution and high signal-to-noise ratio at the same time, but also can reduce the average laser intensity, realize the measurement of the spectrum of heat-absorbing materials, and prevent laser energy from Excessive damage to the sample; at the same time, due to the change in the shape of the entrance, it is allowed to pass through a large-radius spot, which can reduce the average intensity of the laser and realize the measurement of the spectrum of the endothermic material.

Description of drawings

Fig. 1 is the flowchart of a kind of spectral detection method based on digital projection provided by the present invention;

Fig. 2 is a schematic diagram of the optical path structure of the instrument provided by the present invention;

Among them, 1 is the entrance part; 2 is the collimating mirror; 3 is the grating; 4 is the converging mirror;

Fig. 3 is three kinds of possible entrances provided by the present invention;

Among them, (a) is a fixed slit (rhombic slit); (b) is an adjustable slit; (c) is an optical fiber arranged in a gradient.

Fig. 4 is a difference diagram between a common rectangular slit and a gradient inlet provided by the present invention, and the projection diagram of the light spot on the CCD after passing through four kinds of inlets;

Figure 5 is an effect diagram provided by the present invention;

Among them, (a) is the spectrum produced when the spot projection covers 1 pixel on the CCD; (b) is the spectrum produced when the spot projection covers 95 pixels on the CCD; (c) is the iterative result of the digital projection algorithm ; (d) is the difference value before (c) and (a).

Fig. 6 is a schematic diagram of a digital projection algorithm.

Among them, (a) and (c) are two kinds of original signals; (b) and (d) are signals after dislocation superimposition.

Detailed ways

In order to make the purpose, technical solution and advantages of the present invention clearer, the implementation manners of the present invention will be further described in detail below.

The digital projection algorithm is a data recovery algorithm. The resolution of the signal is reduced due to dislocation and superposition, and it can be restored by the digital projection algorithm. The original signal can be approximated by multiple iterations.

In order to make the spectrum have high resolution and high signal-to-noise ratio at the same time, and overcome the problems existing in the background technology, the embodiment of the present invention proposes a new idea of combining spectrum detection technology and digital projection algorithm. By changing the shape of the light entrance, the luminous flux is relatively small In large cases, post-processing can still be used to obtain high-resolution spectra.

Example 1

A kind of spectral detection method based on digital projection technology, referring to Fig. 1, this spectral detection method mainly comprises the following steps:

101: According to the pixel size of the spectrometer photodetector (CCD) or single-point detector (hereinafter, CCD is taken as an example), design an entrance with a gradient width, so that the incident light spot is finally projected on the CCD in a gradient;

102: According to the number of horizontal projection pixels from less to more, a set of spectral images with resolution from high to low can be extracted from it;

Among them, the spectral resolution is the best when the projections are all concentrated on one pixel on the CCD, and the spectral resolution is poorer as the number of projected pixels increases.

103: Input the spectrograms with different resolutions into the digital projection algorithm, and perform iterative calculations until the results converge;

Wherein, step 103 is specifically:

According to the digital projection algorithm, the input spectrum group is equal to the product of the projection matrix and the output spectrum;

Pre-estimate according to the input spectrum group, obtain an estimated spectrum, and perform projection transformation on the estimated spectrum; compare the projection transformation result with the input spectrum group, and use the ratio to correct the estimated spectrum;

Iterate repeatedly until the ratio is less than the threshold, at this time, the estimated spectrum meets the expected requirements, and the estimated spectrum at this time is output as the result.

104: The convergence result can be restored to the spectrum with the best resolution, that is, the spectrum when all the projections are concentrated on one pixel on the CCD.

According to the digital projection principle mentioned above, the measured spectrum is produced by projection of the spectrum with the best resolution. The digital projection algorithm simulates this projection process, takes the measured spectrum as input, and performs inverse calculation through multiple iterations to restore the projection process. The result is the best-resolution spectrum.

In summary, the embodiment of the present invention realizes the requirements of high resolution and high signal-to-noise ratio of the spectrum through the above steps 101 to 104; at the same time, since the change of the entrance shape allows the light spot with a large radius to pass, the laser average can be reduced. Intensity, to achieve the measurement of the spectrum of endothermic materials.

Example 2

Below in conjunction with Fig. 2 to Fig. 6, and specific examples, the scheme in embodiment 1 is further introduced, see the following description for details:

The embodiment of the present invention proposes a spectral detection method based on digital projection, and designs an entrance with a gradient width to collect spectral images with high information content, and cooperates with later data processing methods to restore spectral images with high resolution and Spectra with high signal-to-noise ratio. The spectral detection method can be divided into two parts: entrance and detection. The entrance part 1 is the focus of the design of the embodiment of the present invention.

There are two options for the entrance part 1 , which are optical fiber type and slit type, and the slit type can be divided into fixed type and adjustable type. The detection part can match the detection part of other existing spectrum detection instruments. Generally, the grating is used to split the spectrum, and the photoelectric detector (CCD) or digital micromirror device (DMD) is used to cooperate with the single-point detector to collect the spectrum. As for the detection part, the embodiments of the present invention take the example of collecting light with CCD5 after grating 2 splitting light.

FIG. 2 is a schematic diagram of the optical path of the instrument provided by the embodiment of the present invention. The entrance part 1 is an adjustable slit, which can be replaced by the entrance part shown in FIG. 3 . Fig. 3 is a schematic diagram of three kinds of entrance parts, including but not limited to the three kinds shown in Fig. 3, as long as the width of the entrance part changes gradually.

As shown in Figure 2, the light is incident from the slit, diverges into parallel light through the collimating mirror 2 and hits the grating 3, after the grating 3 splits the light, the dispersed monochromatic light is in order according to the wavelength (or frequency) Arranged, converged by the objective lens 4, and projected on the photodetector 5 (the embodiment of the present invention uses a CCD as an example for illustration). The spectral resolution is the best when the projections are all concentrated on one pixel on the CCD5, and the spectral resolution is worse when the number of projected pixels is more.

The embodiment of the present invention proposes an inlet portion 1 whose width changes gradually, such as a diamond-shaped slit (hereinafter, the rhombus-shaped slit is used as an example), as shown in FIG. The ratio is projected on the CCD5, so if the traditional rectangular slit wants to obtain the ultimate resolution (that is, the projection on the CCD is one pixel), the width of the slit must be smaller than the width of one pixel of the CCD5, which will cause the light flux to be too small and the signal-to-noise ratio too low. The diamond-shaped slit is different from the commonly used rectangular slit, and has a larger luminous flux and a better signal-to-noise ratio. The light spot is incident from a rhombus-shaped slit, and the projection on the CCD5 is relatively rhombus-shaped, from which a set of spectral images with resolutions ranging from high to low can be extracted.

The principle of spectral resolution reduction is shown in Figure 5. When the light spot is projected on multiple pixels on the CCD5, there will be overlapping of different wavelength spectra, which will lead to a reduction in resolution. Therefore, the resolution of (a) is much lower than that of ( b). As shown in Figure 5, select the 75th to 95th group of data (that is, the spectrum generated when the spot projection covers 75 pixels to 95 pixels on the CCD5) as input, where (a) is the spot projection covering the CCD5 The spectrum generated by 1 pixel has the best resolution, (b) is the spectrum generated when the spot projection covers 95 pixels on the CCD5, (c) is the result obtained after iteration, which greatly improves Spectral resolution, (d) is the difference between (a) and (c). Therefore, the spectrum can have high resolution and high signal-to-noise ratio at the same time; at the same time, because the change of the shape of the slit allows a large-radius spot to pass, the average intensity of the laser can be reduced, thereby realizing the measurement of the spectrum of the endothermic material.

The digital projection algorithm is a data recovery algorithm. The resolution of the signal is reduced due to dislocation and superposition. As shown in Figure 6, it can be restored by the digital projection algorithm, and the original signal can be approximated by multiple iterations. The limit resolution spectrum is obtained through inversion iteration, that is, the spectrum map when the projections are all concentrated on one pixel on the CCD5.

Wherein, the operation of step 103 in embodiment 1 is specifically:

Let the input spectrum group be g ₀ ; the projection matrix is H=H _m*n (a _ij ) _m*n ; H is the projection matrix; a _ij is the element of the projection matrix; m is the number of rows of the projection matrix; n is the projection matrix The number of columns; the estimated spectrum is f; the output spectrum is f ₀ .

According to the digital projection algorithm, g ₀ =H*f ₀ ;

in, a _in is the column vector of the projection matrix; h _sum (n) is the projection proportion array.

Perform pre-estimation according to the input spectrum group g ₀ to obtain an estimated spectrum f=H ^T *g ₀ ; perform projection transformation on the estimated spectrum, g=H*f; T is transpose.

The result g is compared with the input spectrum group g ₀ , g _div =g ₀ /g, and the estimated spectrum is corrected by the ratio g _div .

f _div (n)=(a _in ) _m*1 *g _div (n); g _div (n) is an array of projection deviations; f _div (n) is an array of spectral deviations.

f(n) is the estimated spectrum array for this time; f'(n) is the last estimated spectrum array.

The iterations are repeated until the estimated spectrum meets the expected requirement, that is, g _div is smaller than the threshold D, where the threshold D is set according to the actual application requirements, which is not limited in the embodiment of the present invention.

The final estimated spectrum is output as a result, ie f ₀ =f.

To sum up, the embodiment of the present invention realizes the requirements of high resolution and high signal-to-noise ratio of the spectrum through the above operations; at the same time, because the change of the shape of the entrance allows the light spot with a large radius to pass, the average intensity of the laser can be reduced, and the absorption can be achieved. Measurement of thermal material spectra.

In the embodiments of the present invention, unless otherwise specified, the models of the devices are not limited, as long as they can complete the above functions.

Those skilled in the art can understand that the accompanying drawing is only a schematic diagram of a preferred embodiment, and the serial numbers of the above-mentioned embodiments of the present invention are for description only, and do not represent the advantages and disadvantages of the embodiments.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection of the present invention. within range.

Claims (4)

1. A spectral detection method based on digital projection, the spectral detection method comprising the steps of:

designing an inlet part with gradient width according to the size of the pixel size of a photoelectric detector or a single-point detector of a spectrometer, so that an incident light spot is finally projected on the photoelectric detector or the single-point detector of the spectrometer in a gradient manner;

wherein, the inlet part with the width changing in a gradient way is specifically as follows: the optical fiber type or the slit type is divided into a fixed type and an adjustable type;

extracting a group of spectrograms with resolution from high to low according to the number of transverse projection pixels from small to large;

inputting spectrograms with different resolutions into a digital projection algorithm, and performing iterative computation until the result is converged;

the convergence result can be reduced to a spectrum with the best resolution, namely a spectrum when the projection is totally concentrated on one pixel on a photoelectric detector or a single-point detector of the spectrometer;

inputting the spectrograms with different resolutions into a digital projection algorithm, and performing iterative computation until the result converges, wherein the steps are specifically as follows:

according to the digital projection algorithm, the input spectrum group is equal to the product of the projection matrix and the output spectrum;

pre-estimating according to the input spectrum group to obtain an estimated spectrum, and performing projection transformation on the estimated spectrum; comparing the projection transformation result with the input spectrum group, and correcting the estimated spectrum by using the ratio;

and repeating iteration until the ratio is smaller than the threshold value, the estimated spectrum meets the expected requirement at the moment, and the estimated spectrum at the moment is output as a result.

2. The method according to claim 1, wherein the fixed type is a diamond slit.

3. The spectrum detection method based on digital projection according to claim 1, wherein the step of projecting the incident light spot on the spectrometer photodetector or single-point detector in a final gradient manner is specifically:

designing an optical path structure, the optical path structure comprising: an inlet part, a collimating mirror, a grating, a converging mirror, and a spectrometer photodetector or single point detector,

light enters from an entrance part with the width changing in a gradient manner, is diffused into parallel light through a collimating mirror and then is emitted onto a grating, and monochromatic light dispersed by colors is sequentially arranged according to the wavelength or the frequency after the grating is subjected to light splitting;

and the light is converged by the objective lens and projected on a photoelectric detector of the spectrometer or a single-point detector, when the projection is totally concentrated on one pixel on the photoelectric detector of the spectrometer or the single-point detector, the spectral resolution is optimal, and the spectral resolution is worse when the number of projection pixels is more.

4. The spectrum detection method based on digital projection according to claim 1, wherein the step of comparing the projection transformation result with the input spectrum group and correcting the estimated spectrum by using the ratio specifically comprises:

wherein f is_div(n) is a spectral deviation array; a is_inIs a column vector of the projection matrix; m is the number of rows of the projection matrix; g_div(n) is a projection deviation array; f (n) is the spectrum array estimated this time; f' (n) is the last estimated spectrum array; h is_sumAnd (n) is a projection proportion array.