Computational on-chip spectrometer based on arrayed waveguide structure and spectral reconstruction method

Abstract

The application discloses a calculation type on-chip spectrometer based on an array waveguide structure and a spectrum reconstruction method, relates to the field of integrated optics, and discloses a calculation type on-chip spectrometer based on an array waveguide structure, which comprises an array waveguide structure, a free diffraction module and an input-output coupler. Incident light enters the closely arranged array waveguide structure through the input coupler, the multiple modes in the array waveguide are coupled and interfered with each other, then the light is remixed and split by the free diffraction module, and the interfered light is output through the output coupler. The array waveguide structure and the free diffraction module jointly form a circulation loop, the effective optical path difference of the light in the waveguide is increased, and the spectral resolution is improved. The calculation type on-chip spectrometer based on the array waveguide structure is calibrated to obtain a measurement matrix, a regularization optimization algorithm is used to reconstruct and recover the unknown incident spectrum, the spectral resolution is further improved, and the working bandwidth is expanded.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of integrated optics, in particular to a computing on-chip spectrometer based on an arrayed waveguide structure and a spectrum reconstruction method. BACKGROUND

[0002] Miniaturized spectrometers have wide applications in material analysis, environmental monitoring, space exploration and other fields. However, due to the fact that traditional optical spectrometers usually need to use free-space dispersion elements such as prisms and diffraction gratings to map different wavelengths of light to different positions in space, and combine line-scan photodetector modules to detect spectral signals, or use mechanical structures to scan the arm length of a Michelson interferometer to realize Fourier transform type spectral detection. Generally speaking, the above two traditional spectral monitoring methods need complex optical splitting structures and photodetection technologies, or rely on mechanical scanning systems, which to some extent increases the complexity of the spectral detection system, and thus makes it difficult to reduce the size and cost of the spectral detection equipment.

[0003] Silicon-based photonic platforms have the advantages of strong light confinement, high integration, and inherent compatibility with complementary metal oxide semiconductor technology, providing a promising way to develop ultra-compact, high-performance and high-resolution integrated spectrometers. Although the on-chip integrated spectrometer technology based on silicon-based photonic platforms has developed rapidly, the on-chip spectrometers demonstrated by the above work need to make a trade-off between spectral resolution and device size, or need to apply thermal tuning to the device, which to some extent increases the cost and size of the equipment, and is not conducive to the high integration and temperature control of on-chip devices. Within a compact chip area, the limited dispersion freedom and optical path length make it still challenging to realize a compact on-chip spectrometer with high resolution and wide optical bandwidth. SUMMARY

[0004] To solve the above problems, the present application provides a computing on-chip spectrometer based on an arrayed waveguide structure and a spectrum reconstruction method, which uses closely arranged arrayed waveguide structures and free diffraction modules to form a loop structure, greatly increasing the effective optical path difference and improving the spectral resolution within a compact chip size, solving the contradiction between spectral resolution and device size.

[0005] To achieve the above purpose, the present application provides a computing on-chip spectrometer based on an arrayed waveguide structure, which includes an input coupler, an output coupler, an arrayed waveguide structure and a free diffraction module arranged between the input coupler and the output coupler.

[0006] The arrayed waveguide structure is composed of straight waveguides or curved waveguides, including a first arrayed waveguide structure, a second arrayed waveguide structure, a third arrayed waveguide structure and a fourth arrayed waveguide structure.

[0007] The free diffraction module is arranged in a star coupler structure, and includes a first free diffraction module and a second free diffraction module.

[0008] One end of the first array waveguide structure is connected with the input coupler, and the other end of the first array waveguide structure is connected with the first free diffraction module; the first free diffraction module, the second array waveguide structure, the second free diffraction module and the third array waveguide structure are sequentially connected to form a circulation loop; and the second free diffraction module is connected with the output coupler through the fourth array waveguide structure.

[0009] Preferably, the input coupler and the output coupler are both arranged as X, the first array waveguide structure, the third array waveguide structure and the fourth array waveguide structure are all composed of X closely arranged waveguides, the second array waveguide structure includes 2X closely arranged waveguides, and X is any integer greater than or equal to 2.

[0010] Preferably, the input coupler and the output coupler are arranged as grating couplers or end face couplers.

[0011] The spectral reconstruction method of the calculation type on-chip spectrometer based on the array waveguide structure includes the following steps:

[0012] S1: cross-measure the response spectrum of the channel in the on-chip spectrometer, obtain the response spectrum of M physical channels, and construct the relationship matrix between the physical channels and the response spectrum as the measurement matrix Φ of the on-chip spectrometer, wherein M is equal to the product of the number X of input channels and the number X of output channels;

[0013] S2: input the to-be-measured spectrum S into the physical channel of the on-chip spectrometer, measure the output light intensity of different channels, and form an output light intensity distribution O; the relationship function among the to-be-measured spectrum S, the output light intensity distribution O and the calibrated measurement matrix Φ is O=Φ×S;

[0014] S3: according to the relationship function among the to-be-measured spectrum S, the output light intensity distribution O and the calibrated measurement matrix Φ obtained in step S2, the recovery and reconstruction of the to-be-measured spectrum S are realized by using a regularization optimization algorithm, and a reconstructed spectrum is obtained.

[0015] Preferably, in step S2, the to-be-measured spectrum S is a one-dimensional vector of N×1; the output light intensity distribution O is a one-dimensional vector of M×1; and the measurement matrix Φ is a two-dimensional matrix of M×N. Wherein, N is the sampling number of the to-be-measured spectrum.

[0016] Preferably, in step S3, the regularization optimization algorithm includes gradient descent method, conjugate gradient method, two-step iterative shrinkage threshold method, split Bregman method, alternating direction multiplier method and Lagrange multiplier method.

[0017] Preferably, the regularization term is one or a combination of L1 regularization, L2 regularization, elastic net regularization, and total variation regularization.

[0018] Therefore, the application has the following beneficial effects:

[0019] (1) Compared with the existing micro spectrometer, the on-chip spectrometer device of the application is completely compatible with the existing CMOS process, which is conducive to reducing the processing cost and realizing mass production.

[0020] (2) Compared with the existing on-chip spectrometer, the application does not need to know the line type of the measured spectrum in advance, and can directly recover and reconstruct sparse spectrum, continuous spectrum and mixed spectrum, thereby enhancing the convenience of spectrum detection.

[0021] (3) Compared with the existing on-chip spectrometer, the application can solve the underdetermined system problem with the number of spectrum samples N greater than the number of physical channels M of the spectrum device, and can effectively alleviate the inherent contradiction between the resolution and the working bandwidth of the spectrometer under the limited chip size.

[0022] (4) Compared with the existing dispersion type on-chip spectrometer, the application uses closely arranged array waveguide structure and free diffraction module to form a loop structure, which greatly increases the effective optical path difference and improves the spectral resolution under the compact chip size, thereby solving the contradiction between the spectral resolution and the device size. Therefore, the application can reduce the spatial size of the on-chip integrated spectrum system based on waveguide dispersion, and facilitate high-density integration.

[0023] (5) Compared with the existing on-chip spectrometer based on heating electrode tuning waveguide, the application does not need to design and process additional electrode materials for thermal tuning of the waveguide, thereby greatly suppressing the noise interference and reducing the energy consumption caused by temperature change, and the design structure is simple, the process tolerance is large, and it is easy to process and integrate other on-chip devices.

[0024] The technical solutions of the application will be further described in detail below with reference to the drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 is a schematic diagram of the principle of the on-chip spectrometer based on the array waveguide structure of the application;

[0026] Figure 2 is a flowchart of the on-chip spectrometer based on the array waveguide structure and the spectrum reconstruction method of the application;

[0027] Figure 3 is a schematic diagram of the on-chip spectrometer based on the array waveguide structure in Example 1 of the application;

[0028] Figure 4 The spectral correlation function calculation result diagram of the mid-infrared calculation type on-chip spectrometer based on the array waveguide structure in the embodiment one of the present application;

[0029] Figure 5 The simulated methane gas absorption spectrum reconstruction result diagram of the mid-infrared calculation type on-chip spectrometer based on the array waveguide structure in the embodiment one of the present application;

[0030] Figure 6 The structural schematic diagram of the 1550nm waveband calculation type on-chip spectrometer based on the array waveguide structure in the embodiment two of the present application;

[0031] Figure 7 The spectral correlation function calculation result diagram of the 1550nm waveband calculation type on-chip spectrometer based on the array waveguide structure in the embodiment two of the present application;

[0032] Figure 8 The simulated to-be-measured spectrum reconstruction result diagram of the 1550nm waveband calculation type on-chip spectrometer based on the array waveguide structure in the embodiment two of the present application;

[0033] Reference numerals

[0034] 1, input coupler; 2, first array waveguide structure; 3, first free diffraction module; 4, second array waveguide structure; 5, third array waveguide structure; 6, second free diffraction module; 7, fourth array waveguide structure; 8, output coupler. DETAILED DESCRIPTION

[0035] The technical solutions of the present application are further described below through the drawings and embodiments.

[0036] Unless otherwise defined, the technical terms or scientific terms used in the present application shall be understood as the usual meanings understood by those skilled in the art to which the present application belongs.

[0037] The terms "comprising" or "including" as used in this invention mean that the element preceding the term encompasses the element listed after the term, and do not exclude the possibility of encompassing other elements. Terms such as "inner," "outer," "upper," and "lower" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. When the absolute position of the described object changes, the relative positional relationship may also change accordingly. In this invention, unless otherwise explicitly specified and limited, the term "attached" and similar terms should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral part; it can refer to a direct connection or an indirect connection through an intermediate medium; it can refer to the internal communication of two elements or the interaction relationship between two elements. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0038] like Figure 1 As shown, a computational on-chip spectrometer based on an arrayed waveguide structure includes an input coupler 1, an output coupler 8, an arrayed waveguide structure disposed between the input coupler 1 and the output coupler 8, and a free diffraction module.

[0039] The arrayed waveguide structure consists of straight waveguides or curved waveguides, including a first arrayed waveguide structure 2, a second arrayed waveguide structure 4, a third arrayed waveguide structure 5, and a fourth arrayed waveguide structure 7.

[0040] The free diffraction module is configured as a star coupler structure, including a first free diffraction module 3 and a second free diffraction module 6;

[0041] One end of the first array waveguide structure 2 is connected to the input coupler 1, and the other end of the first array waveguide structure 2 is connected to the first free diffraction module 3. The first free diffraction module 3, the second array waveguide structure 4, the second free diffraction module 6 and the third array waveguide structure 5 are connected in sequence to form a loop. The second free diffraction module 6 is connected to the output coupler 8 through the fourth array waveguide structure 7.

[0042] Both input coupler 1 and output coupler 8 are configured to have X units. The first array waveguide structure 2, the third array waveguide structure 5, and the fourth array waveguide structure 7 are each composed of X closely arranged waveguides. The second array waveguide structure 4 includes 2X closely arranged waveguides, where X is any integer of 2 or higher. Input coupler 1 and output coupler 8 are configured as grating couplers or end-face couplers.

[0043] The incident light enters X waveguides in the first array waveguide structure 2 after passing through X input couplers 1, and enters the first free diffraction module 3 after mode mixing in the closely arranged waveguides. The light enters 2X waveguides in the second array waveguide structure 4 after re-mixing and splitting in the first free diffraction module 3, and enters the second free diffraction module 6 after mode coupling in the second array waveguide structure 4. The light enters X waveguides in the third array waveguide structure 5 and X waveguides in the fourth array waveguide structure 7 after re-mixing and splitting in the second free diffraction module 6. The light in the third array waveguide structure 5 re-enters the first free diffraction module 3 after mode coupling, and enters the cycle to increase the effective total optical path. The light in the fourth array waveguide structure 7 is strengthened by mode coupling and interference, and is emitted through the output coupler 8.

[0044] As shown in Figure 2 The spectral reconstruction method of the array waveguide structure-based computational on-chip spectrometer includes the following steps:

[0045] S1: cross-measure the response spectrum of the channel of the on-chip spectrometer, obtain the response spectrum of M physical channels, and construct the relationship matrix between the physical channels and the response spectrum as the measurement matrix Φ of the on-chip spectrometer, wherein M is equal to the product of the number X of input channels and the number X of output channels.

[0046] S2: input the to-be-measured spectrum S into the physical channels of the on-chip spectrometer, measure the output light intensity of different channels, and form an output light intensity distribution O. The relationship function among the to-be-measured spectrum S, the output light intensity distribution O and the calibrated measurement matrix Φ is O = Φ × S. In step S2, the to-be-measured spectrum S is a one-dimensional vector of N × 1; the output light intensity distribution O is a one-dimensional vector of M × 1; and the measurement matrix Φ is a two-dimensional matrix of M × N. Wherein, N is the sampling number of the to-be-measured spectrum.

[0047] S3: according to the relationship function among the to-be-measured spectrum S, the output light intensity distribution O and the calibrated measurement matrix Φ obtained in step S2, the recovery and reconstruction of the to-be-measured spectrum S is realized by using a regularization optimization algorithm, and a reconstructed spectrum In step S3, the regularization optimization algorithm includes gradient descent method, conjugate gradient method, two-step iterative shrinkage threshold method, split Bregman method, alternating direction multiplier method and Lagrange multiplier method. The regularization term is one or a combination of more than one of L1 regularization, L2 regularization, elastic net regularization and total variation regularization.

[0048] Embodiment one

[0049] As shown in Figure 3As shown, the mid-infrared computing on-chip spectrometer based on the arrayed waveguide structure, the input coupler 1 and the output coupler 8 are set as grating couplers, and the input coupler 1 and the output coupler 8 are both set as 20, and are connected to the input end of the first arrayed waveguide structure 2 and the output end of the fourth arrayed waveguide structure 7 respectively; the materials of the input coupler 1, the output coupler 8, the arrayed waveguide structure and the free diffraction module are all silicon nitride on insulator; the first arrayed waveguide structure 2, the third arrayed waveguide structure 5 and the fourth arrayed waveguide structure 7 are all composed of 20 closely arranged waveguides, the second arrayed waveguide structure 4 is composed of 40 closely arranged waveguides, the waveguide width in the arrayed waveguide structure is 1.2um, the waveguide thickness is 0.8um, the gap width between adjacent waveguides is 0.4um, and the total length is 500um; the first free diffraction module 3 and the second free diffraction module 6 adopt a star coupler structure; the lower end of the first free diffraction module 3 is connected to the first arrayed waveguide structure 2 and the third arrayed waveguide structure 5, and the upper end of the first free diffraction module 3 is connected to the second arrayed waveguide structure 4; the lower end of the second free diffraction module 6 is connected to the third arrayed waveguide structure 5 and the fourth arrayed waveguide structure 7, and the upper end of the second free diffraction module 6 is connected to the second arrayed waveguide structure 4; the arrayed waveguide structure and the free diffraction module together constitute a loop structure.

[0050] After the incident light passes through the 20 input couplers 1, it enters the 20 waveguides in the first arrayed waveguide structure 2 respectively, and after the mode mixing in the closely arranged waveguides, the incident light enters the first free diffraction module 3, and after the light re-mixing and splitting in the first free diffraction module 3, it enters the 40 waveguides in the second arrayed waveguide structure 4, and after the inter-mode coupling in the second arrayed waveguide structure 4, the light enters the second free diffraction module 6, and after the light re-mixing and splitting in the second free diffraction module 6, it enters the 20 waveguides in the third arrayed waveguide structure 5 and the 20 waveguides in the fourth arrayed waveguide structure 7 respectively. Among them, the light entering the third arrayed waveguide structure 5 re-enters the first free diffraction module 3 through inter-mode coupling, enters the loop, and increases the effective total optical path; the light entering the fourth arrayed waveguide structure 7 is strengthened through inter-mode coupling interference and then exits through the output coupler 8.

[0051] The spectral reconstruction method of the mid-infrared computing on-chip spectrometer based on the arrayed waveguide structure specifically includes the following steps:

[0052] (1) The scanning range of the tunable continuous-wave laser was set to 2150nm-2350nm with a step size of 0.05nm. The laser output light was coupled into the on-chip spectrometer via a single-mode fiber. The response spectrum of each channel of the on-chip spectrometer was cross-measured, resulting in the acquisition of response spectra I(λ,M) for M (M=400) physical channels, where λ is the sampling wavelength. Each response spectrum includes 4000 sampling wavelengths. The relationship matrix between the physical channels and the response spectra was constructed, which is the measurement matrix Φ1 (400×4000) of the on-chip spectrometer. The above measurement steps were repeated to obtain the measurement matrix Φ2. To preliminarily characterize the resolution performance of the computational on-chip spectrometer, the spectral correlation function was calculated.

[0053]

[0054] Where Δλ is the wavelength interval between two adjacent spectral lines, <·> λ 、<·> M These are averages over the sampling wavelength and the physical channel, respectively. The spectral correlation function calculation results are as follows: Figure 4 As shown, the wavelength interval corresponding to its half maximum value is 0.41 nm, indicating that when the wavelength interval between adjacent input spectral lines is 0.41 nm, the designed computational on-chip spectrometer can achieve decorrelation of 0.5, that is, it can be clearly distinguished.

[0055] (2) The absorption spectrum of methane gas in the 2260nm-2300nm band was measured using a commercial spectrometer to simulate the spectrum to be measured, and the corresponding bands were extracted from the measurement matrices Φ1 and Φ2 respectively to obtain T1 and T2. The output light intensity distribution O was calculated using the extracted measurement matrix T2, i.e. O=T2×S, to simulate the noise interference in the experimental measurement.

[0056] (3) Based on the relationship function between the simulated spectrum S, the simulated output light intensity distribution O, and the intercepted measurement matrix T, this embodiment uses the split Bregman method to solve the optimization problem that introduces total variational regularization.

[0057]

[0058] in, To reconstruct the spectrum, ||·||² is the L2 regularization, TV(·) is the total variational regularization, and λ is the penalty parameter for the regularization term, which is obtained in this example using the K-fold cross-validation method. Furthermore, this embodiment utilizes a relative error function to evaluate the accuracy of the spectrum reconstruction:

[0059]

[0060] The methane gas absorption spectrum reconstruction result in this embodiment is as follows: Figure 5As shown, through the proof of concept, the on-chip spectrometer can accurately reconstruct the methane gas absorption spectrum, and the relative error of the reconstructed spectrum is only 0.221.

[0061] Embodiment two

[0062] As Figure 6 The 1550nm band on-chip spectrometer based on the arrayed waveguide structure shown: the input coupler 1 and the output coupler 8 are both set as end face couplers, each having 10, respectively connected to the input end of the first arrayed waveguide structure 2 and the output end of the fourth arrayed waveguide structure 7; the materials of the input coupler 1, the output coupler 8, the arrayed waveguide structure and the free diffraction module are all silicon on insulator; the first arrayed waveguide structure 2, the third arrayed waveguide structure 5 and the fourth arrayed waveguide structure 7 are all composed of 10 closely arranged waveguides, the second arrayed waveguide structure 4 is composed of 20 closely arranged waveguides, the waveguide width in the arrayed waveguide structure is 0.45um, the waveguide thickness is 0.22um, the gap width between adjacent waveguides is 0.15um, and the total length is 100um; the first free diffraction module 3 and the second free diffraction module 6 are both star coupler structures; the lower end of the first free diffraction module 3 is connected to the first arrayed waveguide structure 2 and the third arrayed waveguide structure 5, the upper end of the first free diffraction module 3 is connected to the second arrayed waveguide structure 4, the lower end of the second free diffraction module 6 is connected to the third arrayed waveguide structure 5 and the fourth arrayed waveguide structure 7, and the upper end of the second free diffraction module 6 is connected to the second arrayed waveguide structure 4; the arrayed waveguide structure and the free diffraction module together constitute a loop structure.

[0063] The incident light enters the 10 input couplers 1 respectively, and then enters the 10 waveguides in the first arrayed waveguide structure 2, the incident light is mixed in the closely arranged waveguides, and then enters the first free diffraction module 3, the light reenters the second arrayed waveguide structure 4 after being mixed and split by the first free diffraction module 3, the light is coupled between the modes in the second arrayed waveguide structure 4, and then enters the second free diffraction module 6, the light reenters the 10 waveguides in the third arrayed waveguide structure 5 and the 10 waveguides in the fourth arrayed waveguide structure 7 after being mixed and split by the second free diffraction module 6. Among them, the light entering the third arrayed waveguide structure 5 reenters the first free diffraction module 3 after being coupled between the modes, enters the loop, and increases the effective total optical path; the light entering the fourth arrayed waveguide structure 7 is strengthened by the coupling between the modes and then exits through the output coupler.

[0064] To evaluate the device resolution of the 1550nm band computational on-chip spectrometer based on the arrayed waveguide structure, the embodiment adopts numerical simulation, sets the wavelength scanning range of the incident light as 1530nm-1570nm, sets the step as 0.1nm, measures the output light intensity distribution of different channels, and obtains a measurement matrix Φ (100x400). The spectral correlation function calculation result is shown in Figure 7 The wavelength interval corresponding to the half maximum value is 0.76nm, which indicates that the resolution of the designed 1550nm band computational on-chip spectrometer can reach 0.76nm. In the 1530nm-1570nm band range, a simulated to-be-measured spectrum S with a resolution of 0.8nm is constructed. In order to simulate the experimental environment noise, random noise is added to the measurement matrix Φ to obtain Φ e with a signal-to-noise ratio of 20dB, and the simulated output light intensity distribution O is calculated, that is, O = Φ e x S. In order to accurately reconstruct the simulated to-be-measured spectrum The alternating direction multiplier method is used to minimize the function

[0065]

[0066] So that Where ||·||1, ||·||2 are L1, L2 regularization, D1, D2 are first-order difference operation and second-order difference operation respectively, e is the estimated value of noise, λ1 and λ2 are penalty parameters of the regularization term, and the K-fold cross-validation method is used to obtain the parameters in the embodiment. The simulated to-be-measured spectrum reconstruction result of the embodiment is shown in Figure 8 The relative error of the reconstructed spectrum is only 0.064.

[0067] Therefore, the computational on-chip spectrometer based on the arrayed waveguide structure and the spectrum reconstruction method can use the negative dispersion ytterbium-doped fiber as the gain medium of the high-repetition-frequency fiber laser to simplify the dispersion management, reduce the threshold, optimize the nonlinear effect control, and improve the repetition frequency performance, and finally obtain a low-threshold mode-locked pulse output with a repetition frequency of hundreds of GHz.

[0068] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application but not to limit them, although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can still be modified or replaced by equivalents, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present application.

Claims

1. A computational on-chip spectrometer based on an arrayed waveguide structure, characterized in that: It includes an input coupler, an output coupler, an array waveguide structure disposed between the input coupler and the output coupler, and a free diffraction module; The arrayed waveguide structure consists of straight waveguides or curved waveguides, including a first arrayed waveguide structure, a second arrayed waveguide structure, a third arrayed waveguide structure, and a fourth arrayed waveguide structure; The free diffraction module is configured as a star coupler structure, including a first free diffraction module and a second free diffraction module; One end of the first array waveguide structure is connected to the input coupler, and the other end of the first array waveguide structure is connected to the first free diffraction module. The first free diffraction module, the second array waveguide structure, the second free diffraction module and the third array waveguide structure are connected in sequence to form a loop. The second free diffraction module is connected to the output coupler through the fourth array waveguide structure.

2. The computational on-chip spectrometer based on an arrayed waveguide structure as described in claim 1, characterized in that: Both the input coupler and the output coupler are set to X. The first array waveguide structure, the third array waveguide structure and the fourth array waveguide structure are each composed of X closely arranged waveguides. The second array waveguide structure includes 2X closely arranged waveguides, where X is any integer of 2 or higher.

3. The computational on-chip spectrometer based on an arrayed waveguide structure as described in claim 2, characterized in that: The input and output couplers are configured as grating couplers or end-face couplers.

4. The spectral reconstruction method for a computational on-chip spectrometer based on an arrayed waveguide structure according to any one of claims 1-3, characterized in that: Includes the following steps: S1: Cross-measure the response spectra of the channels in the on-chip spectrometer to obtain the response spectra of M physical channels, construct the relationship matrix between the physical channels and the response spectra, and use it as the measurement matrix Φ of the on-chip spectrometer, where M is equal to the number of input channels X multiplied by the number of output channels X; S2: Input the spectrum to be measured S into the physical channels of the on-chip spectrometer, measure the output light intensity of different channels, and form the output light intensity distribution O. The relationship function between the spectrum to be measured S, the output light intensity distribution O and the calibrated measurement matrix Φ is O=Φ×S; S3: Based on the relationship function between the measured spectrum S, the output light intensity distribution O, and the calibrated measurement matrix Φ obtained in step S2, a regularization optimization algorithm is used to recover and reconstruct the measured spectrum S, resulting in the reconstructed spectrum.

5. The spectral reconstruction method for a computational on-chip spectrometer based on an arrayed waveguide structure as described in claim 4, characterized in that: In step S2, the spectrum to be measured S is an N×1 one-dimensional vector; the output light intensity distribution O is an M×1 one-dimensional vector; and the measurement matrix Φ is an M×N two-dimensional matrix, where N is the number of samples of the spectrum to be measured.

6. The spectral reconstruction method for a computational on-chip spectrometer based on an arrayed waveguide structure as described in claim 4, characterized in that: In step S3, the regularization optimization algorithms include gradient descent, conjugate gradient, two-step iterative shrinkage threshold, split Bregman method, alternating direction multiplier method, and Lagrange multiplier method.

7. The spectral reconstruction method for a computational on-chip spectrometer based on an arrayed waveguide structure as described in claim 6, characterized in that: The regularization term is one or more combinations of L1 regularization, L2 regularization, elastic net regularization, and total variational regularization.

Images