Flour moisture detection sensor based on scattering medium gas absorption spectrum

Abstract

The invention provides a flour moisture detection sensor based on a scattering medium gas absorption spectrum, which comprises a tunable semiconductor laser light source module, a signal processing module, an equivalent average optical path (Leq) calculation module, a portable integrated structure and an optical detection module, belongs to the technical field of optical detection, and aims to solve the problems of large size and low signal-to-noise ratio of the existing GASMAS system. The moisture content in the flour is measured by adopting the GASMAS technology, compared with a traditional measuring instrument, the device has the advantages of being high in precision, good in stability, high in response speed, non-loss in measurement, low in power consumption, easier to operate and the like, and compared with other near-infrared spectrum moisture analyzer products, the used measurement principle is different, and the device has the advantages of being lower in price, more portable, convenient to use and the like. And the product can meet the detection requirements of the food detection industry, the granary storage industry and the small and medium-sized food processing industry, so that the problems in the background technology are solved.

Description

Technical Field

[0001] The present invention relates to the technical field, and particularly to a flour moisture detection sensor based on the gas absorption spectrum of a scattering medium. Background Art

[0002] Existing flour moisture detection methods include the drying method, infrared drying method, resistance capacitance method, etc. Among them, the constant weight method is cumbersome in operation, time-consuming, energy-consuming, and the instrument is generally large and inconvenient to operate. The resistance method has low measurement accuracy and poor reliability. In addition, there are many varieties of flour and its processed products, and factors such as electrolyte content, filling density, thickness, and temperature also have a certain impact on the measurement results. In addition, there is also a near-infrared spectrum flour moisture analyzer, which is expensive, and the online measurement device belongs to surface measurement, cannot reflect the sample, and is also affected by the moisture in the air during the measurement process.

[0003] The Gas Absorption Spectroscopy in Scattering Media (GASMAS) technology realizes non-destructive detection by measuring the equivalent average optical path of light in a scattering medium and combining the gas absorption characteristics. However, the existing GASMAS systems on the market are large in volume and low in signal-to-noise ratio, and it is difficult to meet the moisture detection of flour.

[0004] The present invention proposes a flour moisture detection sensor based on the gas absorption spectrum of a scattering medium to solve this problem. Summary of the Invention

[0005] The present invention uses GASMAS technology to measure the moisture content in flour, and has the advantages of high accuracy, good stability, fast response speed, non-destructive measurement, low power consumption, and easier operation compared with traditional measuring instruments. And compared with other near-infrared spectrum moisture analyzer products, it uses a different measurement principle, has the advantages of lower price, more portable, and faster measurement. This product can meet the detection needs of the food detection industry, granary storage, and small and medium-sized food processing industries, thus overcoming the problems in the above background art.

[0006] A flour moisture detection sensor based on the gas absorption spectrum of a scattering medium includes a tunable semiconductor laser light source module, a signal processing module, an equivalent average optical path (Leq) calculation module, a portable integrated structure, and an optical detection module.

[0007] For further limitation of the above technical solution, the tunable semiconductor laser light source module is used to emit near-infrared laser in the 1392nm band. The laser light source is a distributed feedback semiconductor laser (DFB-LD), which is configured with a temperature control module and a current drive circuit with a sawtooth wave superposed sine modulation. The output power is 10mW, and the modulation frequency is 2.3kHz.

[0008] For further limitation of the above technical solution, the tunable semiconductor laser light source module further includes a driving and modulation circuit and an optical component used in conjunction with it. The driving and modulation circuit includes a low-noise constant current source that provides a stable injection current (accuracy ±0.1 mA) to avoid wavelength drift caused by current fluctuations, slow start protection to prevent current mutations from damaging the laser, a high-frequency scanning signal, a sawtooth wave or sine wave signal (frequency 1 Hz~10 kHz) to drive the laser wavelength to periodically scan the absorption peak, and a lock-in amplifier interface that synchronizes with the detection module to extract weak absorption signals (suppress noise through frequency locking); The optical component includes an optical isolator to prevent reflected light from returning to the laser cavity, avoiding wavelength instability or mode hopping caused by optical feedback, with parameters of isolation degree >30 dB and operating wavelength matching the laser, a collimating lens, types including aspherical lens or gradient refractive index (GRIN) lens, collimating the laser beam into a low-divergence beam (divergence angle <1°), and a coating, anti-reflection (AR) coating, reducing interface loss (transmittance >99%).

[0009] For further limitation of the above technical solution, the signal processing module includes a modulation and demodulation circuit, an analog-to-digital converter (ADC), and a microcontroller. The modulation and demodulation circuit uses lock-in amplification technology to extract the second harmonic (2f) signal and performs noise reduction processing on the signal through the FIR filtering algorithm.

[0010] For further limitation of the above technical solution, the signal processing module further includes a photoelectric conversion circuit, an analog filtering circuit, and a reference signal synchronization channel. The photoelectric conversion circuit includes a detector, a high-speed InGaAs photodiode (response wavelength 900-1700 nm) or an avalanche photodiode (APD), converting the optical signal into a current signal, and a transimpedance amplifier (TIA) converting the current signal into a voltage signal (gain adjustable range 1kΩ~10MΩ) to suppress high-frequency noise; The analog filtering circuit includes a low-pass filter with a cut-off frequency of 1-10 kHz (matching the laser modulation frequency) to suppress high-frequency noise, and a band-pass filter for the absorption peak characteristic frequency (such as 1-100 Hz) to eliminate power frequency interference (50 / 60 Hz).

[0011] For further limitation of the above technical solution, the equivalent average optical path (Leq) calculation module includes measuring the normalized second harmonic peak-to-peak value of the water vapor absorption signal in the scattering medium by the incremental method and obtaining a linear relationship model between the equivalent optical path and the moisture content by combining the least squares fitting.

[0012] For further limitation of the above technical solution, the equivalent average optical path (Leq) calculation module further includes a Monte Carlo simulation unit to pre-calculate the mapping relationship between Leq and scattering parameters through photon transport simulation. Input parameters include flour optical parameters (μs, μa, anisotropy factor g), geometric parameters (sample thickness d, light source-detector spacing), an experimental calibration module, and a standard sample library. Flour samples with known moisture content (0.5% - 15%) and density are prepared, and the absorption signal intensity is measured through a transmission experiment.

[0013] For further limitation of the above technical solution, the portable integrated structure has an overall size of 16.3×14.8×8.5 cm³ and includes a driving circuit board, a modulation and demodulation circuit board, and a power supply module with a stacked design, supporting in-situ detection within 10 seconds. The optical detection module includes a scattering medium cavity for accommodating the flour sample, a G12180-020A type photodetector, and a transimpedance amplifier circuit. The scattering medium cavity uses a 3D-printed polylactic acid (PLA) cuvette container, and the optical path is designed as a transmission-type multiple scattering structure.

[0014] For further limitation of the above technical solution, the operating parameters of the tunable semiconductor laser light source module are: central wavelength 1392.53nm, output power 10mW, modulation frequency 2.3kHz, sawtooth wave scanning frequency 0.5Hz, temperature control accuracy ±0.1℃. The modulation and demodulation circuit includes: a 2.3kHz square wave modulation signal generated by a voltage-controlled oscillator (VCO); a 1f reference signal and a 2f frequency-doubled signal generated by a complex programmable logic device (CPLD); a fifth-order Butterworth low-pass filter for smoothing the demodulated signal; a balanced modulator-demodulator for synchronously demodulating the output signal of the photodetector. FIR algorithm filtering is used to reduce noise in the demodulated signal. The implementation method of the FIR filtering algorithm is: , the FIR filtering algorithm is a 50th-order finite impulse response filter, with a passband cut-off frequency of 500Hz, stopband attenuation ≥40dB, passband ripple ≤0.1dB, and stopband start frequency of 600Hz. The calculation formula for the equivalent average optical path (Leq) is: , where by continuously increasing the air cavity with a known optical path, the linear relationship between the normalized second harmonic signal intensity and the optical path is measured, and the goodness of fit R²≥0.996. The power supply module includes: an input stage using an IRW-60-24V switching power supply to convert 220V AC to 24V DC; a secondary voltage regulation circuit using LT3045 / LT3094 chips to generate ±15V and ±12V low-noise power supplies; power supply ripple ≤1mV, load regulation rate ≤0.01%. The optical path length of the 3D-printed cuvette container is 5mm, the light transmittance of the quartz window is ≥98%, it supports rapid sample replacement, and the air interference suppression rate is ≥90%.

[0015] A method for detecting the moisture content of flour based on the gas absorption spectrum of a scattering medium, characterized by comprising the following steps: Calibration stage: Dynamically adjust the optical path by the incremental method, measure the equivalent average optical path (Leq) of multiple groups of flour samples with known moisture content, and establish a linear calibration model between Leq and the moisture content; Detection stage: Load the flour sample to be measured into the scattering medium cavity, transmit the 1392nm laser through the sample, and use a photodetector to collect the transmitted light signal after multiple scatterings; Signal processing: Extract the second harmonic signal using wavelength modulation spectroscopy (WMS), and calculate the normalized peak-to-peak value after FIR filtering; Result output: Input the processed signal into the calibration model, display the moisture content in real time. After comparative tests on 8 groups of samples with different moisture contents (3.75% - 8.49%), the maximum relative error is 9.3%, and the average error is 4.6%; The least squares fitting formula in the calibration stage is: Moisture content = , where the slope k = 0.152 ± 0.005% / cm and the intercept b = 6.8 ± 0.3%; In the signal processing, a fifth-order Butterworth band-pass filter is used, with a passband range of 2.25 - 2.35kHz and a stopband attenuation ≥ 30dB / octave.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. Lower error: The detection method used in the present invention has a smaller error rate compared to traditional detection methods, and can reduce the error rate to less than or equal to 9.3% during actual use; 2. Portable structure: The optical components used in the present invention have a higher structural integration degree during actual use, thereby reducing the usage difficulty of the present invention during actual use; 3. High degree of automation: The present invention adopts a signal processing module, which can automatically calculate the data that needs to be integrated after collecting the input information, reducing the usage difficulty of the staff. Description of the Drawings

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is the schematic diagram of the sensor system; Figure 2 It is the design drawing of a 3D printed cuvette container; Figure 3 It is the comparison diagram before and after FIR filtering of the 2f signal; Figure 4 It is the linear fitting curve of the average equivalent optical path and the water content. Specific implementation mode

[0019] The following is a further detailed description of the present invention in conjunction with the attached Figures 1-4 drawings.

[0020] Embodiment 1: This embodiment provides a flour moisture detection sensor based on the absorption spectrum of a scattering medium gas. As Figures 1-4 shown, it includes a tunable semiconductor laser light source module, a signal processing module, an equivalent average optical path (Leq) calculation module, a portable integrated structure, and an optical detection module.

[0021] The tunable semiconductor laser light source module is used to emit near-infrared laser in the 1392 nm band. The laser light source is a distributed feedback semiconductor laser (DFB-LD), configured with a temperature control module and a current drive circuit with a sawtooth wave superimposed sine modulation. The output power is 10 mW, and the modulation frequency is 2.3 kHz.

[0022] The tunable semiconductor laser light source module also includes a drive and modulation circuit and an optical component used in conjunction with it. The drive and modulation circuit includes a low-noise constant current source, which provides a stable injection current (accuracy ±0.1 mA) to avoid wavelength drift caused by current fluctuations, slow start protection to prevent current mutation from damaging the laser, a high-frequency scanning signal, a sawtooth wave or sine wave signal (frequency 1 Hz~10 kHz) to drive the laser wavelength to periodically scan the absorption peak, and a lock-in amplifier interface to synchronize with the detection module and extract weak absorption signals (suppress noise through frequency locking); the optical component includes an optical isolator to prevent reflected light from returning to the laser cavity and avoid wavelength instability or mode hopping caused by optical feedback. Parameters: isolation degree >30 dB, working wavelength matching the laser, a collimating lens, type: aspherical lens or gradient refractive index (GRIN) lens, collimating the laser beam into a low-divergence beam (divergence angle <1°), and coating: anti-reflection (AR) coating to reduce interface loss (transmittance >99%).

[0023] The signal processing module includes a modulation and demodulation circuit, an analog-to-digital converter (ADC), and a microcontroller. The modulation and demodulation circuit uses lock-in amplification technology to extract the second harmonic (2f) signal and performs noise reduction processing on the signal through the FIR filtering algorithm.

[0024] The signal processing module further includes a photoelectric conversion circuit, an analog filtering circuit, and a reference signal synchronization channel. The photoelectric conversion circuit includes a detector, a high-speed InGaAs photodiode (response wavelength 900 - 1700 nm) or an avalanche photodiode (APD), which converts an optical signal into a current signal. A transimpedance amplifier (TIA) converts the current signal into a voltage signal (gain adjustable range 1 kΩ - 10 MΩ) to suppress high-frequency noise. The analog filtering circuit includes a low-pass filter with a cut-off frequency of 1 - 10 kHz (matching the laser modulation frequency) to suppress high-frequency noise, and a band-pass filter for the absorption peak characteristic frequency (such as 1 - 100 Hz) to eliminate power frequency interference (50 / 60 Hz).

[0025] The equivalent average optical path (Leq) calculation module includes measuring the normalized second harmonic peak-to-peak value of the water vapor absorption signal in the scattering medium by the incremental method, and obtaining a linear relationship model between the equivalent optical path and the moisture content by least squares fitting.

[0026] The equivalent average optical path (Leq) calculation module further includes a Monte Carlo simulation unit to pre-calculate the mapping relationship between Leq and the scattering parameters through photon transport simulation. Input parameters include the optical parameters of flour (μs, μa, anisotropy factor g), geometric parameters (sample thickness d, light source-detector distance), an experimental calibration module, and a standard sample library. Flour samples with known moisture content (0.5% - 15%) and density are prepared, and the absorption signal intensity is measured through a transmission experiment.

[0027] The portable integrated structure has an overall size of 16.3×14.8×8.5 cm³ and includes a stacked driving circuit board, a modulation and demodulation circuit board, and a power supply module, supporting in-situ detection within 10 seconds. The optical detection module includes a scattering medium cavity for accommodating the flour sample, a G12180-020A photodetector, and a transimpedance amplification circuit. The scattering medium cavity uses a 3D-printed polylactic acid (PLA) cuvette container, and the optical path is designed as a transmission-type multiple scattering structure.

[0028] The operating parameters of the tunable semiconductor laser light source module are: center wavelength 1392.53 nm, output power 10 mW, modulation frequency 2.3 kHz, sawtooth wave scanning frequency 0.5 Hz, and temperature control accuracy ±0.1°C. The modulation and demodulation circuit includes: a 2.3 kHz square wave modulation signal generated by a voltage-controlled oscillator (VCO); a 1f reference signal and a 2f frequency-doubled signal generated by a complex programmable logic device (CPLD); a fifth-order Butterworth low-pass filter for smoothing the demodulated signal; and a balanced modulator-demodulator for synchronously demodulating the output signal of the photodetector. FIR algorithm filtering is used to reduce noise in the demodulated signal. The implementation method of the FIR filtering algorithm is: , the FIR filtering algorithm is a 50 - order finite impulse response filter, with a pass - band cut - off frequency of 500 Hz, a stop - band attenuation ≥ 40 dB, a pass - band ripple ≤ 0.1 dB, and a stop - band start frequency of 600 Hz; the calculation formula for the equivalent average optical path (Leq) is: , where by continuously increasing the air cavity with a known optical path, the linear relationship between the normalized second - harmonic signal intensity and the optical path is measured, and the goodness of fit R² ≥ 0.996; the power supply module includes: the input stage uses an IRW - 60 - 24V switching power supply to convert 220V AC into 24V DC; the secondary voltage - stabilizing circuit uses LT3045 / LT3094 chips to generate ±15V and ±12V low - noise power supplies; the power supply ripple ≤ 1 mV, and the load regulation rate ≤ 0.01%; the optical path length of the 3D - printed cuvette container is 5 mm, the light transmittance of the quartz window ≥ 98%, it supports rapid sample replacement, and the air interference suppression rate ≥ 90%.

[0029] Example 2: This example provides a flour moisture detection sensor based on the absorption spectrum of a scattering medium gas, as Figures 1-4 shown, including the following design of the detection circuit module: It is mainly divided into three parts, namely the circuit power - supply module, the laser driver module, and the modulation - demodulation module. The circuit power - supply module provides the power supply voltage for the entire circuit system. The main function of the laser driver module is to control the temperature and current of the laser. In addition, the high - performance micro - processor on the driver circuit board is responsible for collecting and processing signal data. The role of the modulation - demodulation module is to add a modulation signal to the drive current and demodulate the data at the data - receiving end.

[0030] A. Design of the system power - supply module: It includes a positive - power - supply circuit and a negative - power - supply circuit. In the laboratory, 220V industrial frequency alternating current is used. When powering the device, stable and low - noise power supply needs to be considered. However, the commercial power will generate problems such as surges, current harmonics, electrical interference, and noise, which will seriously interfere with the measurement accuracy of the entire system. Therefore, the measurement system requires a stable power - supply module and has the characteristic of miniaturization. During the voltage - conversion process, the noise impact needs to be considered particularly, and the power - supply filtering should be optimized as much as possible, which can greatly improve the measurement accuracy of the entire system. To ensure safer and more stable output, the positive - power - supply circuit is equipped with a 10 µF ceramic output capacitor and various protection circuits, including power - supply reverse - connection protection, current limiting, and thermal limiting, where the current limiting has a fold - back function. In addition, the current - limiting degree and power of this regulator can be adjusted according to requirements. The negative - power - supply circuit also has the functions of current - limiting and power adjustment, and also has bipolar disable pins and fast - start capabilities.

[0031] B. Design of Laser Driver Module: It mainly includes temperature control, current control and FIR filtering. In the driver module, the key component for temperature control of the Peltier semiconductor cooler is the MAX1978 chip. It is not only the most reliable and accurate temperature control solution, but also built-in with field-effect transistors and temperature control circuits, greatly reducing the need for external components without affecting efficiency. After the temperature is set, the MAX1978 measures the operating temperature of the laser by detecting the voltage of the negative temperature coefficient thermistor inside the laser. This resistor changes with the change of the laser temperature, and the voltage also changes accordingly. The MAX1978 uses this voltage value to compare with the set target voltage value, and the generated error value is used as the input value of the PID controller inside the chip. After processing, it outputs control for temperature compensation to achieve precise control of the laser temperature. The current control circuit is mainly a constant current source composed of an operational amplifier. The operational amplifier uses the OP213 dual operational amplifier chip, which has the characteristics of low noise and low temperature drift, with built-in calibration, and the power supply range is a single power supply from 4 V to 36 V. Many operational amplifiers can calibrate offset and gain, but cannot correct temperature drift and noise. However, the OP213 combines the advantages of excellent analog performance and digital correction, enabling it to optimize temperature drift and noise. Due to scattering, the received optical power is greatly reduced, and the waveform signal-to-noise ratio of the second harmonic also becomes very poor. Therefore, we added the FIR filtering algorithm in the STM32F405 signal processing to filter the 2f signal before calculating the peak-to-peak value.

[0032] C. Modulation and Demodulation Module: It includes laser sine modulation and second harmonic signal demodulation. In the laser spectrum, the wavelength of the light source can be modulated. Wavelength modulation can be achieved by passing the modulation signal through a modulator or by changing the wavelength of the light source itself. This modulation can be in the form of sine waves, square waves, etc. Wavelength modulation spectroscopy can improve the signal-to-noise ratio, sensitivity, and stability, and suppress baseline drift. We use a Butterworth low-pass filter to convert square waves into sine waves. Wavelength modulation spectroscopy uses a high-frequency modulation signal to lift the spectral signal from low frequency to high frequency for more efficient transmission. Finally, after the signal is received, it is still high frequency and mixed with noise of other frequencies. To analyze this signal, it is necessary to filter out the noise, extract the signal of the target information frequency, and remove the initially modulated signal to restore the original signal, which is equivalent to the inverse process of modulation. This is demodulation. The lock-in amplifier was originally mainly used for interference suppression in radar and communication systems and was later used in fields such as optics to improve the sensitivity of experimental measurements. Its core principle is to perform phase locking through a reference signal, thereby selectively amplifying the components with the same frequency as the reference signal. Today, the technology of lock-in amplifiers has developed rapidly, combining advanced digital signal processing and microprocessing technologies, enabling higher performance and faster data acquisition speeds, thus achieving excellent demodulation effects. The laser sine modulation module and the second harmonic signal demodulation module complement each other, thus constituting the modulation and demodulation module we need.

[0033] The overall effect of this design is: The electronic circuit part of this system is mainly divided into a power supply module, a laser driver module, a modulation and demodulation module, and a detector receiving module. The power supply module supplies power to the entire system, mainly providing ±12V voltage. The power supply circuits of each part are analyzed, and chips are selected in combination with the system situation. The laser driver module provides temperature and current control for the laser, and STM32F405 can collect and process data. The function of the modulation and demodulation module is to superimpose a flow control modulation signal and demodulate the received signal, and filter the signal before and after demodulation to reduce noise. The detector collects and processes the laser passing through the sample, and an AC scan analysis is performed on the filter circuit in the detector.

[0034] Example 3: This example provides a flour moisture detection sensor based on the absorption spectrum of a scattering medium gas, as Figures 1-4 shown, including the following installation method: 1. Hardware Assembly (1) Fix the DFB laser (1392 nm) on the heat dissipation base, and connect the MAX1978 temperature controller and the OP213 constant current source.

[0035] (2) The modulation signal is generated by 74HC4046, superimposed with a sawtooth wave after being frequency-divided by CPLD, and input into the laser drive circuit.

[0036] (3) The photodetector (G12180-020A) receives the transmitted optical signal, which is transmitted to the STM32F405 after being amplified by the transimpedance amplifier OPA828 and demodulated by the AD630.

[0037] (4) The 3D printed cuvette container is installed at the center of the optical path to ensure that the sample thickness is 5 mm.

[0038] 2. Software Configuration (1) Burn the FIR filtering program in the STM32F405 and set the sampling frequency to 1 kHz.

[0039] (2) Calibration stage: Measure 5 groups of samples with known water content (3.75% - 8.49%) and fit a linear model.

[0040] (3) Detection stage: Input the signal of the unknown sample, calculate the water content through the model, and the result is displayed on the PC side in real time.

[0041] 3. Verification Test (1) By comparing eight groups of samples with the drying results, it is found that the maximum relative error is 9.3% and the average relative error is 4.6%; Continuously detect a 4.28% sample for 70 minutes, and the fluctuation range is 0.48%.

[0042] Example 4: This example provides a method for detecting the moisture content of flour based on the gas absorption spectrum of a scattering medium, which is characterized by including the following steps: Calibration stage: Dynamically adjust the optical path by the incremental method, measure the equivalent average optical path (Leq) of multiple groups of flour samples with known water content, and establish a linear calibration model between Leq and the water content; Detection stage: Load the flour sample to be measured into the scattering medium cavity, transmit the 1392 nm laser through the sample, and use a photodetector to collect the transmitted optical signal after multiple scattering; Signal processing: Use wavelength modulation spectroscopy (WMS) to extract the second harmonic signal, calculate the normalized peak-to-peak value after FIR filtering; Result output: Input the processed signal into the calibration model, display the water content in real time. After the comparative test of 8 groups of samples with different water contents (3.75% - 8.49%), the maximum relative error is 9.3% and the average error is 4.6%; The least squares fitting formula in the calibration stage is: Water content = k * Leq + b, where the slope k = 0.152 ± 0.005% / cm and the intercept b = 6.8 ± 0.3%; A fifth-order Butterworth bandpass filter is used in the signal processing, with a passband range of 2.25 - 2.35 kHz and a stopband attenuation of ≥ 30 dB / decade.

[0043] The above content is a further detailed description of the present invention in combination with specific preferred implementation embodiments, which is convenient for those skilled in the art of this technology to understand and apply the present invention. It cannot be determined that the specific implementation of the present invention is only limited to these descriptions.

Claims

1. A flour moisture detection sensor based on the gas absorption spectrum of a scattering medium, characterized in that, It includes a tunable semiconductor laser light source module, a signal processing module, an equivalent average optical path (Leq) calculation module, a portable integrated structure, and an optical detection module.

2. The flour moisture detection sensor based on the gas absorption spectrum of a scattering medium according to claim 1, wherein The tunable semiconductor laser light source module is used to emit near-infrared laser in the 1392 nm band. The laser light source is a distributed feedback semiconductor laser (DFB-LD), configured with a temperature control module and a current drive circuit with sawtooth wave superimposed sine modulation. The output power is 10 mW, and the modulation frequency is 2.3 kHz.

3. The flour moisture detection sensor based on the gas absorption spectrum of a scattering medium according to claim 2, wherein The tunable semiconductor laser light source module also includes a drive and modulation circuit and an optical component used in conjunction with it. The drive and modulation circuit includes a low-noise constant current source that provides a stable injection current (accuracy ±0.1 mA) to avoid wavelength drift caused by current fluctuations, slow start protection to prevent current mutations from damaging the laser, a high-frequency scanning signal, a sawtooth wave or sine wave signal (frequency 1 Hz - 10 kHz) to drive the laser wavelength to periodically scan the absorption peak, and a lock-in amplifier interface to synchronize with the detection module and extract weak absorption signals (suppress noise through frequency locking); The optical component includes an optical isolator to prevent reflected light from returning to the laser cavity and avoid wavelength instability or mode hopping caused by optical feedback. Parameters: isolation degree > 30 dB, operating wavelength matching the laser, a collimating lens, type: aspherical lens or gradient refractive index (GRIN) lens, to collimate the laser beam into a low-divergence beam (divergence angle < 1°), and coating: anti-reflection (AR) coating to reduce interface loss (transmittance > 99%).

4. The flour moisture detection sensor based on the gas absorption spectrum of the scattering medium according to claim 3, characterized in that, The signal processing module includes a modulation and demodulation circuit, an analog-to-digital converter (ADC), and a microcontroller. The modulation and demodulation circuit uses lock-in amplification technology to extract the second harmonic (2f) signal and performs noise reduction processing on the signal through the FIR filtering algorithm.

5. The flour moisture detection sensor based on the gas absorption spectrum of a scattering medium according to claim 4, wherein The signal processing module also includes a photoelectric conversion circuit, an analog filtering circuit, and a reference signal synchronization channel. The photoelectric conversion circuit includes a detector, a high-speed InGaAs photodiode (response wavelength 900 - 1700 nm) or an avalanche photodiode (APD), which converts the optical signal into an electrical current signal, and a transimpedance amplifier (TIA) that converts the electrical current signal into a voltage signal (gain adjustable range 1 kΩ - 10 MΩ) to suppress high-frequency noise; The analog filtering circuit: a low-pass filter with a cut-off frequency of 1 - 10 kHz (matching the laser modulation frequency) to suppress high-frequency noise, and a band-pass filter for the characteristic frequency of the absorption peak (such as 1 - 100 Hz) to eliminate power frequency interference (50 / 60 Hz).

6. The flour moisture detection sensor based on the gas absorption spectrum of a scattering medium according to claim 5, wherein The equivalent average optical path (Leq) calculation module includes measuring the normalized second harmonic peak-to-peak value of the water vapor absorption signal in the scattering medium by the incremental method and obtaining a linear relationship model between the equivalent optical path and the moisture content by fitting with the least squares method.

7. The flour moisture detection sensor based on the gas absorption spectrum of a scattering medium according to claim 6, characterized in that, The equivalent average optical path (Leq) calculation module further includes a Monte Carlo simulation unit to pre-calculate the mapping relationship between Leq and scattering parameters through photon transmission simulation, input parameters, flour optical parameters (μs, μa, anisotropy factor g), geometric parameters (sample thickness d, light source-detector distance), an experimental calibration module, a standard sample library, prepare flour samples with known moisture content (0.5% - 15%) and density, and measure the absorption signal intensity through a transmission experiment.

8. The flour moisture detection sensor based on the gas absorption spectrum of a scattering medium according to claim 7, wherein The portable integrated structure has an overall size of 16.3×14.8×8.5 cm³ and includes a driving circuit board, a modulation and demodulation circuit board, and a power supply module with a stacked design, supporting in-situ detection within 10 seconds; the optical detection module includes a scattering medium cavity for accommodating the flour sample, a G12180-020A photodetector, and a transimpedance amplifier circuit. The scattering medium cavity uses a 3D-printed polylactic acid (PLA) cuvette container, and the optical path is designed as a transmission-type multiple scattering structure.

9. The flour moisture detection sensor based on the gas absorption spectrum of a scattering medium according to claim 8, characterized in that, The operating parameters of the tunable semiconductor laser light source module are as follows: central wavelength 1392.53 nm, output power 10 mW, modulation frequency 2.3 kHz, sawtooth wave scanning frequency 0.5 Hz, temperature control accuracy ±0.1 °C; the modulation and demodulation circuit includes: a 2.3 kHz square wave modulation signal generated by a voltage-controlled oscillator (VCO); a 1f reference signal and a 2f frequency-doubled signal generated by a complex programmable logic device (CPLD); a fifth-order Butterworth low-pass filter for smoothing the demodulated signal; a balanced modulator-demodulator for synchronously demodulating the output signal of the photodetector. FIR algorithm filtering is used to reduce the noise of the demodulated signal; the implementation method of the FIR filtering algorithm is: , the FIR filtering algorithm is a 50th-order finite impulse response filter, the passband cut-off frequency is 500 Hz, the stopband attenuation ≥ 40 dB, the passband ripple ≤ 0.1 dB, and the stopband starting frequency is 600 Hz; the calculation formula for the equivalent average optical path (Leq) is: , where by continuously increasing the air cavity with a known optical path, the linear relationship between the normalized second harmonic signal intensity and the optical path is measured, and the goodness of fit R² ≥ 0.996; the power supply module includes: an input stage using an IRW-60-24V switching power supply to convert 220V AC into 24V DC; a secondary voltage stabilizing circuit using LT3045 / LT3094 chips to generate ±15V and ±12V low-noise power supplies; the power supply ripple ≤ 1 mV, and the load regulation rate ≤ 0.01%; the optical path length of the 3D printed cuvette container is 5 mm, the light transmittance of the quartz window ≥ 98%, it supports rapid sample replacement and the air interference suppression rate ≥ 90%.

10. A method for detecting the moisture content of flour based on the gas absorption spectrum of a scattering medium, characterized in that, It includes the following steps: Calibration stage: Dynamically adjust the optical path by the incremental method, measure the equivalent average optical path (Leq) of multiple groups of flour samples with known water content, and establish a linear calibration model between Leq and water content. Detection stage: Load the flour sample to be measured into the scattering medium cavity, transmit the 1392nm laser through the sample, and use the photodetector to collect the transmitted light signal after multiple scattering. Signal processing: Use wavelength modulation spectroscopy (WMS) to extract the second harmonic signal, and calculate the normalized peak-to-peak value after FIR filtering. Result output: Input the processed signal into the calibration model, display the water content in real time. After comparative tests on 8 groups of samples with different water contents (3.75% - 8.49%), the maximum relative error is 9.3%, and the average error is 4.6%. The least squares fitting formula in the calibration stage is: Moisture content = , where the slope k = 0.152 ± 0.005% / cm and the intercept b = 6.8 ± 0.3%; In the signal processing, a fifth-order Butterworth band-pass filter is used, with a passband range of 2.25 - 2.35kHz and a stopband attenuation of ≥30dB / decade.

Images