Frequency multiplexing spectrum detection device and method

Abstract

The invention discloses a frequency multiplexing spectrum detection device and method, and belongs to the technical field of computational spectroscopy and micro optical sensing. The device comprises a frequency modulation module which is used for carrying out intensity modulation on incident light with a group of different modulation frequencies to generate a frequency-coded modulation light beam; the spectrum modulation module is composed of a plurality of filtering units with different band-pass characteristics and realizes selective response to different wavelength components; the detection module is used for receiving the modulated and filtered optical signal and converting the modulated and filtered optical signal into a time sequence electric signal; and the signal processing and resolving module is used for carrying out frequency domain analysis on the electric signal and solving an inverse problem in combination with a pre-calibrated system response matrix so as to reconstruct a complete spectrum of the to-be-measured light. Multispectral channel parallel detection is realized through a frequency multiplexing mechanism, the luminous flux and the signal-to-noise ratio of the system are remarkably improved, background noise and channel crosstalk are effectively suppressed, and the method can be widely applied to the fields of environment monitoring, portable analysis equipment and the like.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of computational spectroscopy and miniature optical sensing technology, and particularly relates to a single-detector spectral detection device and method based on frequency coding multiplexing, which is suitable for on-site spectral analysis scenarios requiring high sensitivity, anti-crosstalk and miniaturization, such as online monitoring of environmental pollutants, portable drug component detection, non-invasive spectral diagnosis of biological tissues, etc. BACKGROUND

[0002] Spectral detection technology has a wide range of applications in material composition analysis, environmental monitoring, biomedicine, military reconnaissance, etc. Traditional spectrometers usually use grating or Fourier transform for light splitting, which has high spectral resolution but has problems such as large system size, complex structure, high cost, sensitivity to environmental vibration, etc., making it difficult to meet the needs of on-site, real-time and portable detection. To solve the above problems, spectral detection technology based on filter array has emerged. This technology integrates a series of micro-nano filters with different center wavelengths directly on the pixels of a face array detector, with each pixel responsible for detecting the light intensity of a specific waveband. The spectral information of the target can be obtained through a single exposure, realizing the miniaturization of the system. However, the existing technology has the following limitations: (1) low light flux: each filter corresponds to a detector pixel, and the photon collection efficiency of each spectral channel is limited, which restricts the detection sensitivity. (2) The need for a detector array significantly increases the size and price of the system. Computational spectral reconstruction technology is a new emerging spectral technology that can reduce the manufacturing precision requirements of optical devices, realize a small-sized and low-cost spectral detection system, and take advantage of the potential of the computing system to compensate for the performance of the hardware with high software computing power. SUMMARY

[0003] The present application aims to overcome the problems of low light flux, the need for a detector array and severe spectral crosstalk in the prior art, and provides a frequency multiplexing spectral detection device and method. The device combines frequency coding modulation in the time domain with multi-channel filtering in the spatial / spectral domain, allowing a single detector to receive all the modulated signals and reconstructing the spectrum through frequency domain demodulation and system response matrix inversion, thereby achieving: M times improvement in light flux (M is the number of filter units), as a single detector receives all channel photons; Improvement in signal-to-noise ratio, as frequency coding shifts the signal to the high frequency band, effectively suppressing 1 / f noise; Reduction in channel crosstalk rate, as different filter unit signals are coded at different frequencies, achieving orthogonal separation at the electronics level.

[0004] The technical solution of the present application is as follows: A frequency multiplexing spectral detection device, comprising: A frequency modulation module is configured to apply a set of mutually different preset modulation frequencies to the incident light to be measured to generate a composite modulated light beam carrying frequency-encoded information. A spectral modulation module is arranged in the propagation path of the composite modulated light beam and includes a plurality of filter units arranged side by side, each filter unit being composed of microstructure units having a specific bandpass spectral response characteristic, and the bandpass spectral portions of the filter units overlapping with each other; the composite modulated light beam simultaneously irradiates all the filter units; A detection module is a single photodetector arranged behind the spectral modulation module and configured to receive a mixed light signal after frequency and spectral double modulation and output a corresponding time sequence electrical signal. A signal processing and solving module is electrically connected to the detection module and configured to perform frequency domain analysis on the time sequence electrical signal, extract signal amplitudes corresponding to each preset modulation frequency to form a measurement vector V, and perform linear inverse operation based on a pre-calibrated system response matrix H to reconstruct the complete spectrum of the light to be measured. The system response matrix H is an m*n matrix, the i-th row of which corresponds to the sampling values of the spectral response function of the i-th filter unit at n discrete wavelength points, m is the number of filter units, and n is the number of spectral channels.

[0005] The frequency modulation module is any one of a micro-electro-mechanical system (MEMS) mirror, an acousto-optic modulator (AOM), a liquid crystal spatial light modulator (LCoS), or a digital micromirror device (DMD), and has a modulation frequency range of 1 kHz to 10 MHz and a modulation depth of ≥80%.

[0006] Each filter unit is composed of microstructure units, which are Fabry-Perot cavities, superstructures, or photonic crystal structures, and are integrated on the same substrate through micro-nano processing technology.

[0007] A frequency multiplexing spectral detection method based on any one of the above devices includes the following steps: S1. A set of mutually different preset modulation frequencies (f1, f2,..., fn) are applied to the wide-spectrum light to be measured by the frequency modulation module to generate a composite modulated light beam carrying frequency-encoded information. m S2. The composite modulated light beam is simultaneously irradiated to all the filter units of the spectral modulation module, each filter unit selectively transmits the incident light based on its spectral response function, and outputs a time sequence mixed electrical signal. S3. The time sequence mixed electrical signal is collected by the detection module. S4. Frequency domain analysis is performed on the time sequence electrical signal, and signal amplitudes A corresponding to each preset modulation frequency f are extracted. i i ​​, forming a measurement vector V = [A1, A2,..., A m ] T ; S5. Based on the pre-calibration system response matrix , the linear equation set V = H·S is solved by least square method or Tikhonov regularization algorithm to obtain the spectral intensity vector S of the measured light .

[0008] The frequency domain analysis adopts fast Fourier transform (FFT) or digital phase-locked amplification technology; when the modulation frequency interval is greater than the system bandwidth, FFT is adopted; when the signal-to-noise ratio is lower than 10 dB, digital phase-locked amplification is adopted.

[0009] The system response matrix H is calibrated by the following method: using a standard light source with a known spectrum to irradiate the device, measuring the output amplitude of each filter unit at each preset modulation frequency, and taking the amplitude as the element of the matrix H, wherein the i-th row corresponds to the sampling value of the spectral response function of the i-th filter unit at n discrete wavelength channels.

[0010] A program stored on a computer readable medium, when executed on a processor, for implementing steps S4 and S5 in the method, comprising: Performing fast Fourier transform or digital phase-locked amplification on the time sequence electrical signal of each filter unit to extract the amplitude A i of each modulation frequency; Reading the pre-stored system response matrix H, solving by least square or Tikhonov regularization algorithm, and outputting the spectral intensity vector S.

[0011] The beneficial effects of the present application are: The present application realizes the following technical effects through frequency coding multiplexing mechanism: (1) Light flux is increased by ≥5 times: in traditional FPA, only one pixel is used for each wavelength channel, and the photon utilization rate is ≤10%; in the present application, all channel information is received by a single detector, and the photon utilization rate is ≥50%.

[0012] (2) Signal-to-noise ratio is increased by ≥10 dB: the modulation frequency (>1 kHz) is much higher than the environmental light and the 1 / f noise frequency band (<100 Hz) of the detector, and background interference can be suppressed by FFT or phase-locked amplification, and the experimental SNR reaches 35 dB (the traditional FPA is 22 dB).

[0013] (3) Channel crosstalk rate <3%: the crosstalk of traditional FPA is >15% due to spectral overlap, and the present application realizes channel isolation at the electronic level even if the filter center wavelength interval is only 10 nm and the bandwidth overlap is 30%.

[0014] (4) Miniaturized and low-cost system: No CCD / CMOS array is required, only a single detector and programmable modulator are used, and the system volume is <10cm³. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall structure of the spectral detection device of the present invention, wherein the light to be measured (101) generates a composite modulated beam through the frequency modulation module (102), and the beam simultaneously illuminates all the filtering units of the spectral modulation module (103), and is converted into a timing electrical signal by the detection module (104) and then input into the signal processing and calculation module (105).

[0016] Figure 2 This is a schematic diagram of the spectral modulation module, in which the filter unit (1031) is composed of microstructure units (1032, i.e., Fabry-Perot cavities), and each filter unit is arranged in parallel on the same substrate.

[0017] Figure 3 This is a flowchart illustrating the principle of frequency reuse and signal processing, showing: composite modulated beam → filtering → single detector acquisition → FFT frequency domain analysis → amplitude extraction → matrix inverse solution.

[0018] Figure 4 This is a schematic diagram of the spectral response curves of the filtering unit in the spectral modulation module. The three curves correspond to filtering units A (center wavelength at 4000nm), B (center wavelength at 2600nm), and C (center wavelength at 2150nm), respectively. The overlap of adjacent curves in the 2500–3000nm band is ≥30%. Detailed Implementation

[0019] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0020] See Figure 1 The light to be tested 101 (such as a white LED) is incident on the frequency modulation module 102. In this embodiment, the module is a multi-region driven MEMS galvanometer (such as a Texas Instruments DMD chip), whose reflective surface is divided into three independent control regions, which oscillate periodically at frequencies f1=50 Hz, f2=120 Hz, and f3=800 Hz, respectively, thereby spatially dividing the incident light into three beams (composite modulation beams) with different modulation frequencies, and simultaneously illuminating all the filtering units (non-time-division or spatially isolated) of the spectral modulation module 103.

[0021] See Figure 2 The spectral modulation module 103 consists of three microstructure filter units (1031) integrated on a glass substrate, each unit being a photonic crystal structure (1032), fabricated using electron beam evaporation and reactive ion etching processes.Figure 4 As shown, the center wavelength of filter unit A is 4000nm; filter unit B is 2600nm; and filter unit C is 2150nm. The three have significant overlap in the 2500–3000nm band (average transmittance >30%), ensuring continuous spectral coverage.

[0022] After the modulated beam is illuminated across the entire field of view, the signal received by each filter unit is:

[0023] Where S(λ) is the spectral intensity to be measured, and T i (λ) is the transmittance function of the i-th filter unit.

[0024] The detection module 104 uses an InAsSb photodetector, and its output timing voltage signal is sent to the signal processing and calculation module 105. This module first amplifies the signal (gain = 60dB) and samples it using a 12-bit ADC (sampling rate 10 kHz), then performs Fast Fourier Transform (FFT) analysis, such as... Figure 3 As shown, amplitudes V1, V2, and V3 are extracted at 50Hz, 120Hz, and 800Hz respectively, forming a measurement vector V = [V1, V2, V3]. T .

[0025] The system response matrix H was obtained through calibration: the system was illuminated using a NIST standard tungsten lamp (spectrum known), and the output amplitude at each modulation frequency was recorded. After calibration, H is a 3×10 matrix (corresponding to 10 wavelength points: 2500–3000 nm, step size 50 nm). mn This represents the transmittance response of the m-th filter unit at the n-th wavelength.

[0026] Finally, the solution is obtained using Tikhonov regularized least squares: Where the regularization parameter λ=10 -4 Determined by the L-curve method.

[0027] In an alternative embodiment, the light can be first filtered by a spectral modulation module before frequency modulation. In this case, it is essential to ensure that the modulator is located behind the filter and that the modulation frequency is lower than the system bandwidth (>10 kHz) to avoid frequency aliasing caused by filter response hysteresis. This approach is suitable for applications with stable light sources and low modulation speed requirements.

[0028] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A frequency-reused spectral detection device, characterized in that, include: The frequency modulation module (102) is used to apply a set of different preset modulation frequencies to the incident light to be tested (101) to generate a composite modulated beam carrying frequency encoding information; A spectral modulation module (103) is disposed on the propagation path of the composite modulation beam and includes multiple filter units (1031) arranged in parallel. Each filter unit is composed of a microstructure unit (1032) with specific bandpass spectral response characteristics, and the bandpass spectra of each filter unit overlap. The composite modulation beam simultaneously illuminates all filter units (1031). The detection module (104) is a single photodetector, located behind the spectral modulation module (103), used to receive the mixed optical signal after frequency and spectral modulation, and output the corresponding timing electrical signal; The signal processing and calculation module (105) is electrically connected to the detection module (104) and is used to perform frequency domain analysis on the time-series electrical signal, extract the signal amplitude corresponding to each preset modulation frequency, form a measurement vector V, and perform linear inverse operation based on the pre-calibrated system response matrix H to reconstruct the complete spectrum of the light to be measured. The system response matrix H is an m×n dimensional matrix, where the i-th row corresponds to the sampled values ​​of the spectral response function of the i-th filter unit at n discrete wavelength points, m is the number of filter units, and n is the number of spectral channels.

2. The apparatus according to claim 1, characterized in that, The frequency modulation module (102) is any one or a combination of the following: a microelectromechanical system (MEMS) galvanometer, an acousto-optic modulator (AOM), a liquid crystal spatial light modulator (LCoS), or a digital micromirror device (DMD), with a modulation frequency range of 1 kHz to 10 MHz and a modulation depth ≥80%.

3. The apparatus according to claim 1, characterized in that, Each filter unit (1031) is composed of microstructure units (1032), which are Fabry-Perot cavities, metasurfaces or photonic crystal structures, and are integrated on the same substrate through micro-nano fabrication processes.

4. A frequency multiplexing spectral detection method based on the device according to any one of claims 1-3, characterized in that, Includes the following steps: S1. Using the frequency modulation module (102), apply a set of different preset modulation frequencies (f1, f2, ..., f) to the broadband light under test. m ), generating a composite modulated beam carrying frequency-coded information; S2. The composite modulation beam is simultaneously irradiated onto all the filtering units (1031) of the spectral modulation module (103). Each filtering unit selectively transmits the incident light based on its spectral response function and outputs a time-series mixed electrical signal. S3. Acquire the timing-mixed electrical signal through the detection module (104); S4. Perform frequency domain analysis on the time-series electrical signal and extract each preset modulation frequency f. i The corresponding signal amplitude A i This forms the measurement vector V = [A1, A2, ..., A m ] T ; S5. Based on a pre-calibrated system response matrix The spectral intensity vector of the light to be measured is obtained by solving the linear equation system V = H·S using the least squares method or the Tikhonov regularization algorithm. .

5. The method according to claim 4, characterized in that, The frequency domain analysis employs either Fast Fourier Transform (FFT) or digital phase-locked amplification (PLA). When the modulation frequency interval is greater than the system bandwidth, FFT is used; when the signal-to-noise ratio is less than 10 dB, digital PLA is used.

6. The method according to claim 4, characterized in that, The system response matrix H is calibrated in the following way: the device is illuminated with a standard light source of known spectrum, the output amplitude of each filter unit is measured at each preset modulation frequency, and the amplitude is used as the element of matrix H, where the i-th row corresponds to the sampled value of the spectral response function of the i-th filter unit on n discrete wavelength channels.

7. A program stored on a computer-readable medium, characterized in that, When executed on a processor, steps S4 and S5 of the method of claim 4 are implemented, including: Perform a Fast Fourier Transform or Digital Locked-Lock Amplification on the timing electrical signal of each filtering unit to extract the amplitude A of each modulation frequency. i ; Read the pre-stored system response matrix H, solve it using the least squares or Tikhonov regularization algorithm, and output the spectral intensity vector S.

Images