CN118882821A - A computational spectrometer, computational spectrum reconstruction method, and readable storage medium - Google Patents

Abstract

The present invention provides a computational spectrometer, including: a laser for providing a calibration light source; a single-mode optical fiber for limiting the mode distribution of incident light and transmitting stable incident light to a broadband filter part; a polarization controller installed on the single-mode optical fiber so that the incident light has a unique polarization; a broadband filter part placed in the tail end direction of the single-mode optical fiber, the broadband filter part including a first broadband filter and a second broadband filter, the first broadband filter and the second broadband filter are used alternately to split the signal light to be measured; a transparent scattering medium placed after the broadband filter part so that the incident light is scattered to obtain a speckle pattern; and a planar array photodetector. The beneficial effects of the present invention are: using a single-mode optical fiber to guide the incident light, being able to reduce the difficulty of reconstruction calculation and increase the bandwidth of the broadband filter part, and using a transparent scattering medium as the core component of the spectrometer, with great advantages of stable structure, low cost, and simple system.

Description

A computational spectrometer, computational spectrum reconstruction method, and readable storage medium

Technical Field

The present invention relates to a computational spectrometer, and in particular to a computational spectrometer, a computational spectrum reconstruction method, and a readable storage medium.

Background Art

Spectrometers are widely used in scientific research and industrial production. Traditional spectrometers usually use gratings or prisms for light splitting, but gratings and splitting prisms have defects such as large size, high cost and limited resolution. In recent years, with the introduction of computational spectrometers, the resolution can be improved and the size of the equipment can be greatly reduced from the perspective of algorithm and device design, realizing a miniaturized spectrometer suitable for production applications. Unlike the simple one-to-one mapping of spectrum to space in traditional spectrometers, computational spectrometers are usually complex one-to-many mappings.

At present, there are many solutions for implementing complex mapping of computational spectrometers, among which computational spectrometers based on speckle reconstruction have the advantages of ultra-compactness and low cost. The core of speckle reconstruction is to map the spectral information into a complex speckle pattern after passing through a scattering medium. The first method to form a speckle pattern was to use multimode optical fiber, single-mode optical fiber, etc., but the optical fiber structure is extremely unstable and easily affected by the environment; disordered photonic crystals, multimode spiral waveguides and other structures can also be used, but such structures require complex designs and have high preparation costs.

Therefore, it is still challenging to develop a miniaturized spectrometer with stable structure, low cost, simple system and the resolution of commercial spectrometers.

Summary of the invention

In order to solve the problems in the prior art, the present invention provides a computational spectrometer, a computational spectrum reconstruction method, and a readable storage medium.

The present invention provides a computational spectrometer, comprising:

Laser, used to provide calibration light source;

Single-mode optical fiber, which limits the mode distribution of the incident light and is used to transmit stable incident light to the broadband filtering part;

A polarization controller, mounted on the single-mode optical fiber, so that the incident light has a unique polarization;

A broadband filter part is placed in the tail end direction of the single-mode optical fiber, and the broadband filter part includes a first broadband filter and a second broadband filter, and the first broadband filter and the second broadband filter are used alternately to split the signal light to be measured;

A transparent scattering medium is placed after the broadband filter part to scatter the incident light to obtain a speckle pattern;

The area array photoelectric detector is arranged in the direction of the speckle pattern and is used to receive and record the scattering pattern.

As a further improvement of the present invention, the laser is a broadband tunable laser or a laser covering a near-infrared band (eg, 900-1600 nm).

As a further improvement of the present invention, the computational spectrometer further includes a computer storing an executable computational program for acquiring pre-calibration information of the planar array photoelectric detector and performing spectrum reconstruction calculation to obtain an input spectrum.

As a further improvement of the present invention, the first broadband filter and the second broadband filter are used to divide the incident light into bands. The first broadband filter and the second broadband filter are located at the same position of the computational spectrometer and are used alternately through a wheel.

As a further improvement of the present invention, the light transmission range of the first broadband filter is 400-600nm, and the light transmission range of the second broadband filter is 600-800nm.

As a further improvement of the present invention, the transparent scattering medium is an object with a random scattering mechanism.

As a further improvement of the present invention, the planar array photoelectric detector adopts a silicon-based detector in the visible light band.

As a further improvement of the present invention, the planar array photoelectric detector adopts an InGaAs material detector in the infrared band.

As a further improvement of the present invention, the output of the calibration light source is single wavelength light.

The present invention also provides a computational spectrum reconstruction method, which is based on the computational spectrometer and performs the following steps:

S1. Record the calibration speckle pattern.

The broadband tunable laser scans and outputs, and the array photodetector records the calibration speckle pattern. During the calibration process, the broadband filter part is not used, and it is only necessary to divide the scanning wavelength into two groups according to different bands;

S2. Find the transmission matrix,

Processing the calibration speckle pattern on a computer, obtaining a calibration matrix through mean filtering and equal-interval sampling, and calculating a transmission matrix;

S3, solve the input spectrum,

Input the signal light to be measured, use the first broadband filter and the second broadband filter alternately, obtain the signal vector to be measured, and solve to obtain the input spectrum.

As a further improvement of the present invention, step S1 includes: using the broadband tunable laser to scan and output a series of calibration light sources in a set spectral range at a set step length, and using the array photodetector to record calibration speckle patterns corresponding to a series of calibration light sources, that is, there is at least one calibration speckle pattern corresponding to each wavelength.

As a further improvement of the present invention, step S2 comprises: processing the calibration speckle pattern on a computer, dividing the calibration speckle pattern into two groups, one group under the wavelength range corresponding to the first broadband filter, and another group under the wavelength range corresponding to the second broadband filter, obtaining two groups of calibration matrices by an equal-interval sampling method for the calibration speckle pattern after mean filtering preprocessing, and solving two groups of transmission matrices in combination with the spectrum of the calibration light source.

As a further improvement of the present invention, step S3 includes: inputting the signal light to be measured from a single-mode optical fiber, alternately using the first broadband filter and the second broadband filter, collecting two speckle patterns corresponding to the signal light to be measured, and using the equally spaced sampling method in step S2 to obtain two groups of signal vectors to be measured, respectively applying corresponding transmission matrices, combining the compressed sensing principle and the convex optimization solution method to obtain two groups of spectral vectors, and merging the results to obtain the input spectrum of the signal light to be measured.

The present invention also provides a readable storage medium, in which execution instructions are stored. When the execution instructions are executed by a processor, they are used to implement the computational spectrum reconstruction method.

The beneficial effects of the present invention are as follows: a computational spectrometer and a computational spectrum reconstruction method, and a readable storage medium are provided, which use single-mode optical fiber to guide incident light, a broadband filter part that can reduce the difficulty of reconstruction calculation and increase the bandwidth, and a transparent scattering medium as the core component of the spectrometer, and have great advantages of stable structure, low cost, and simple system; in addition, the computational spectrometer uses two broadband filters to separate the incident light beam in space in advance, so that the corresponding speckle patterns have more obvious differences, which improves the contrast of the speckle image, thereby reducing the calculation error; at the same time, the introduction of the two broadband filters divides the speckle images into two groups for reconstruction calculation, reduces the dimension of the calibration matrix, can greatly reduce the calculation amount of spectral reconstruction, and more importantly, can break through the bandwidth limitation of the reconstruction single peak of the computational spectrometer, so that its performance is closer to that of a commercial spectrometer.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other solutions can be obtained based on these drawings without paying creative work.

FIG1 is a schematic diagram of the structure of a computational spectrometer based on a transparent disordered medium and a filter structure provided by the present invention.

FIG. 2 is a flow chart of a spectrum reconstruction method in an example of the present invention.

FIG. 3 is a diagram of narrowband spectrum reconstruction according to an example of the present invention.

FIG. 4 is a diagram of broadband spectrum reconstruction of an example of the present invention.

In the figure: 1. Broadband tunable laser; 2. Single-mode optical fiber; 3. Polarization controller; 4. Incident light beam; 5. Broadband filtering part; 6. Transparent scattering medium; 7. Array photodetector; 8. Computer.

DETAILED DESCRIPTION

It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the scope of protection of the present invention. In addition, the terms "first", "second", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood by specific circumstances.

The present invention will be further described below in conjunction with the accompanying drawings and specific implementation methods.

As shown in FIG1 , the present invention provides a computational spectrometer, which is a computational spectrometer based on a transparent disordered medium (also called a transparent scattering medium) and a filter structure, and mainly includes:

A broadband tunable laser 1, used to provide a calibration light source;

The single-mode optical fiber 2 limits the mode distribution of the incident light 4 and is used to transmit the stable incident light 4 to the broadband filtering part;

A polarization controller 3, mounted on the single-mode optical fiber 2, so that the incident light has a unique polarization;

A broadband filter part 5 is placed in the tail end direction of the single-mode optical fiber 2. The broadband filter part 5 includes two broadband filters, namely a first broadband filter and a second broadband filter. The first broadband filter and the second broadband filter are used alternately to split the signal light to be measured.

A transparent scattering medium 6 is placed after the broadband filter part 5 to scatter the incident light to obtain a speckle pattern;

The area array photoelectric detector 7 is arranged in the direction of the speckle pattern and is used to receive and record the scattering pattern.

The computational spectrometer further comprises a computer 8 on which an executable computational program is stored for acquiring the pre-calibration information of the planar array photoelectric detector 7 and performing spectrum reconstruction calculation to obtain an input spectrum.

The present invention provides a computational spectrometer, which has the characteristics of low computational complexity, low cost and high resolution.

The present invention adds two broadband filters to perform wavelength decomposition processing on the wavelength to be reconstructed in time, which not only improves the reconstruction accuracy of the computational spectrometer, but also greatly reduces the calculation complexity, and is expected to realize a more reliable and low-cost spectrometer.

In one embodiment of the present invention, the output of the calibration light source is a single wavelength light, and its half-peak width and scanning interval can be selected according to the application.

In one embodiment of the present invention: the first broadband filter and the second broadband filter are used to divide the incident light into bands, the first broadband filter and the second broadband filter are located at the same position of the computational spectrometer, and are used alternately through a wheel.

In one embodiment of the present invention: the first broadband filter and the second broadband filter are used alternately when reconstructing the spectrum. When applied in the visible light band, the light transmission range of the first broadband filter is 400-600nm, and the light transmission range of the second broadband filter is 600-800nm. And so on in the near infrared. The light transmission ranges of the two broadband filters can be appropriately adjusted according to the application.

In one embodiment of the present invention, the transparent scattering medium 6 is an object with a random scattering mechanism, which may be frosted glass, a diffraction hole or other objects with a random scattering mechanism.

In one embodiment of the present invention, the planar array photoelectric detector 7 adopts a silicon-based detector in the visible light band, and the detector material can be appropriately selected according to different applications.

In one embodiment of the present invention, the planar array photoelectric detector 7 uses an InGaAs material detector in the infrared band, and the detector material can be appropriately selected according to different applications.

Referring to FIG1 , the present invention is a computational spectrometer based on a transparent disordered medium and a filter structure. For near-infrared applications, the broadband tunable laser 1 can be replaced with a laser covering a near-infrared band such as 900-1600 nm, and the array photodetector 7 can be replaced with a detector made of InGaAs materials.

Referring to FIG. 2 , FIG. 2 is a flow chart of a spectrum reconstruction method in an example of the present invention. The present invention also provides a computational spectrum reconstruction method, which performs the following steps based on the computational spectrometer:

S1. Record the calibration speckle pattern.

The broadband tunable laser 1 scans and outputs, and the area array photodetector 7 records the calibrated speckle pattern;

S2. Find the transmission matrix,

Processing the calibration speckle pattern on a computer 8, obtaining a calibration matrix through mean filtering and equally spaced sampling, and obtaining a transmission matrix;

S3, solve the input spectrum,

Input the signal light to be measured, use the first broadband filter and the second broadband filter alternately, obtain the signal vector to be measured, and solve to obtain the input spectrum.

In one of the embodiments of the present invention, step S1 comprises: using the broadband tunable laser 1 to scan and output a series of calibration light sources in a set spectral range at a set step length, and using the array photodetector 7 to record calibration speckle patterns corresponding to a series of calibration light sources, that is, at least one calibration speckle pattern corresponds to each wavelength.

In one of the embodiments of the present invention, step S2 comprises: processing the calibration speckle pattern on a computer, dividing the calibration speckle pattern into two groups, one group under the wavelength range corresponding to the first broadband filter, and another group under the wavelength range corresponding to the second broadband filter, obtaining two groups of calibration matrices by an equal-interval sampling method of the calibration speckle pattern after mean filtering preprocessing, and solving two groups of transmission matrices in combination with the spectrum of the calibration light source.

In one of the examples of the present invention: step S3 includes: inputting the signal light to be measured from a single-mode optical fiber, alternately using the first broadband filter and the second broadband filter, collecting two speckle patterns corresponding to the signal light to be measured, and using the equally spaced sampling method in step S2 to obtain two groups of signal vectors to be measured, respectively applying corresponding transmission matrices, combining the compressed sensing principle and the convex optimization solution method to obtain two groups of spectral vectors, and merging the results to obtain the input spectrum of the signal light to be measured.

In one embodiment of the present invention, step S2 includes the following sub-steps:

Step S21: cropping and adjusting image processing;

Step S22: removing background noise;

Step S23: Gaussian filtering to remove noise;

Step S24: taking a number of pixels as a sampling point and sampling at equal intervals to obtain a calibration matrix composed of column vectors;

For example, 9 pixels are taken as a sampling point and sampled at equal intervals to obtain a calibration matrix composed of column vectors, but this is not limited to this;

Step S25: combining the input light source intensity and the calibration matrix, calculating the corresponding inverse matrix to obtain the transmission matrix.

In one embodiment of the present invention, step S3 includes:

Combined with the principle of compressed sensing, the spectral channels in the calibration stage are N, and the sampling channels corresponding to each wavelength are M, where M is less than N. At the same time, convex optimization solution methods include L2 norm regularization solution, orthogonal matching pursuit, etc.

A readable storage medium stores execution instructions, which are used to implement the computational spectrum reconstruction method when executed by a processor.

As shown in Figures 3 and 4, both figures show the restoration of the preset spectral signal by the spectrometer. The detected output speckle is used as a special fingerprint, and the input spectral signal is reconstructed and solved using a convex optimization algorithm. From the results, it can be seen that both discrete spectra and continuous spectra can be reconstructed and restored, and spectral information of different intensities can be restored more accurately.

The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.

Claims (10)

1. A computational spectrometer, characterized in that: comprising the following steps:

The laser is used for providing a calibration light source;

A single mode optical fiber limiting a mode distribution of incident light for transmitting stable incident light to the broadband filtering part;

a polarization controller mounted on the single mode optical fiber to make the incident light have unique polarization;

The broadband filtering part is arranged in the tail end direction of the single-mode optical fiber, and comprises a first broadband optical filter and a second broadband optical filter, and the first broadband optical filter and the second broadband optical filter are used for alternately splitting signal light to be detected;

a transparent scattering medium arranged behind the broadband filtering part for scattering incident light to obtain speckle patterns;

And the area array photoelectric detector is arranged in the direction of the speckle pattern and is used for receiving and recording the scattering pattern.

2. The computing spectrometer of claim 1, wherein: the calculation spectrometer further comprises a computer, wherein an executable calculation program is stored on the computer and is used for acquiring preset target information of the area array photoelectric detector and performing spectrum reconstruction calculation to obtain an input spectrum.

3. The computing spectrometer of claim 1, wherein: the first broadband optical filter and the second broadband optical filter are adopted to sub-band incident light, and are positioned at the same position of the calculation spectrometer and are used alternately through a wheel disc.

4. The computing spectrometer of claim 1, wherein: the light transmission range of the first broadband optical filter is 400-600nm, and the light transmission range of the second broadband optical filter is 600-800nm.

5. The computing spectrometer of claim 1, wherein: the transparent scattering medium employs an object with a random scattering mechanism.

6. The computing spectrometer of claim 1, wherein: the area array photoelectric detector adopts a silicon-based detector of a visible light wave band or an InGaAs material detector of an infrared wave band.

7. The computing spectrometer of claim 1, wherein: the laser adopts a broadband tunable laser or a laser covering a near infrared band.

8. A method of computing spectral reconstruction, characterized by: the following steps are performed based on the computational spectrometer of any of claims 1 to 7:

S1, recording a calibration speckle pattern,

The laser scans and outputs, the area array photoelectric detector records a calibration speckle pattern, and a broadband filtering part is not used in the calibration process;

s2, obtaining a transmission matrix,

Processing the scaled speckle pattern on a computer, obtaining a scaling matrix through mean filtering and equidistant sampling, and solving a transmission matrix;

s3, solving the input spectrum,

Inputting signal light to be detected, alternately using the first broadband filter and the second broadband filter to obtain a signal vector to be detected, and solving to obtain an input spectrum.

9. The method of computing spectral reconstruction of claim 8, wherein: the step S1 comprises the following steps: scanning and outputting a series of calibration light sources in a set spectrum range according to a set step length by using the laser, and recording calibration speckle patterns corresponding to the series of calibration light sources by using the area array photoelectric detector, namely at least one calibration speckle pattern is corresponding to each wavelength; the step S2 comprises the following steps: processing the calibration speckle pattern on a computer, dividing the calibration speckle pattern into two groups, dividing the first broadband filter into one group under a wavelength range corresponding to the first broadband filter, dividing the second broadband filter into the other group under a wavelength range corresponding to the second broadband filter, preprocessing the calibration speckle pattern through mean filtering, and then obtaining two groups of calibration matrixes by an equidistant sampling method, and solving the two groups of transmission matrixes by combining the spectrum of the calibration light source; the step S3 comprises the following steps: inputting signal light to be measured from a single mode fiber, alternately using the first broadband filter and the second broadband filter, collecting two speckle patterns corresponding to the signal light to be measured, adopting an equidistant sampling method in the step S2 to obtain two groups of signal vectors to be measured, respectively applying corresponding transmission matrixes, combining a compressed sensing principle and a convex optimization solving method to obtain two groups of spectrum vectors, and combining the results to obtain the input spectrum of the signal light to be measured.

10. A readable storage medium, characterized by: the readable storage medium has stored therein execution instructions which, when executed by a processor, are for implementing the method of computing spectral reconstruction according to any one of claims 8 to 9.