CN104535465A - PM2.5 concentration detection method and device based on neural network - Google Patents

Abstract

The invention provides a method and device for detecting PM2.5 concentration based on a neural network. First, a laser detection system is established based on the principle of Fraunhofer diffraction, and then according to the light intensity digital signal collected by the laser detection system, the Fraunhofer Based on the principle of diffraction, it is verified that there is a corresponding functional relationship between the light intensity digital signal and the PM2.5 concentration value, and the light intensity digital signal is used as an input to establish a regularized neural network model to output the PM2.5 concentration value. The shortcomings of the low automation degree of the PM2.5 concentration detection method in the prior art are overcome, and repeated detection can be realized, the detection accuracy is high, and the calculation is simple and convenient.

Description

PM2.5 Concentration Detection Method and Device Based on Neural Network

Technical field:

The invention relates to the field of concentration detection, in particular to a neural network-based PM2.5 concentration detection method and device.

Background technique:

In recent years, severe smog and polluted air have appeared in the central and eastern regions of my country, and related studies have shown that PM2.5 is the chief culprit of smog. PM2.5 refers to particulate matter with an aerodynamic diameter less than or equal to 2.5 μm in the atmosphere, also known as particulate matter that can enter the lungs. Compared with coarser atmospheric particles, PM2.5 has a smaller particle size and is rich in a large number of toxic and harmful substances And the residence time in the atmosphere is long, the transportation distance is long, pollutes the atmospheric environment, and will pose a serious threat to people's health.

The existing PM2.5 concentration detection methods mostly use manual gravimetric method, β-ray attenuation method and micro-vibration balance method. However, the method of manual gravimetric detection of PM2.5 has problems such as low automation, poor detection repeatability, easy to accumulate errors, and medium consumption.

Invention content:

Aiming at the defects of the prior art, the present invention provides a method and device for detecting PM2.5 concentration based on a neural network, which overcomes the defect of low automation of the PM2.5 concentration detection method in the prior art, and can realize repeated detection, detection High precision and easy calculation.

On the one hand, the present invention provides a kind of PM2.5 concentration detection method based on neural network, comprising:

Establish a laser detection system based on the principle of Fraunhofer diffraction;

According to the light intensity digital signal collected by the laser detection system, based on the Fraunhofer diffraction principle, it is verified that there is a corresponding functional relationship between the light intensity digital signal and the PM2.5 concentration value;

Using the light intensity digital signal as an input, a regularized neural network model is established to output a PM2.5 concentration value.

Optionally, the laser detection system includes: a power supply unit, a He-Ne laser, a filter lens, a beam expander lens, an air pump, an air chamber, a signal receiving unit, a detection unit and a calculation unit;

Wherein, the power supply unit is used to provide power for the He-Ne laser, the air pump, the detection unit and the calculation unit;

The signal receiving unit includes a Fourier lens and 70 photodetectors, which are used to receive the light signal diffracted by air particles and convert it into an analog signal of light intensity;

The detection unit is used to amplify the light intensity analog signal and convert it into a light intensity digital signal;

The calculation unit is used for modeling according to the light intensity digital signal, and calculates the PM2.5 concentration value;

The power supply unit is respectively connected with the He-Ne laser, the air pump, the detection unit and the calculation unit, the He-Ne laser is connected with the filter lens, and the air chamber is respectively connected with the The beam expander lens, the air pump are connected to the signal receiving unit, the signal receiving unit is connected to the detection unit, and the detection unit is connected to the computing unit.

Optionally, according to the light intensity digital signal collected by the laser detection system, based on the principle of Fraunhofer diffraction, it is verified that there is a corresponding functional relationship between the light intensity digital signal and the PM2.5 concentration value, including:

According to the light energy distribution vector measured by the photodetector, the particle size distribution of the particle group to be measured is obtained after calculation and processing, and the particle size distribution percentages with a diameter less than 2.5 μm are added to obtain the PM2.5 concentration value, and the light intensity number is verified. There is a corresponding relationship between the signal and the PM2.5 concentration value.

Optionally, the particle size distribution of the particle group to be measured is calculated by the following formula,

E=TW

Among them, E is the light energy distribution vector measured by the photodetector, $T=\left[\begin{array}{ccc}{t}_{11}& ...& t_{1m}\\ ...& & \\ t_{N1}& ...& t_{\mathrm{N\; M}}\end{array}\right]$ is the light energy contribution matrix, and W is the particle size distribution vector of the particle group to be measured;

in, $t_{\mathrm{N\; M}}=C\frac{1}{{\mathrm{D.}}_m}[J_0^2\left(x_{m,N}\right)+J_1^2\left(x_{m,N}\right)-J_0^2\left(x_{m,N-1}\right)-J_1^2\left(x_{m,N+1}\right)]$ is the light energy contribution of a particle with a diameter of D _M to the Nth detection ring of the detector, is a fixed constant, M is the type of particles with different diameters, I ₀ is the intensity of incident light, J ₀ (X _M,N ) is the zero-order Bessel function, and J ₁ (X _M,N ) is the first-order Bessel function.

Optionally, the PM2.5 concentration value is calculated by the following formula,

Wherein, β _{air density} is 1.205 kg/m ³ at standard atmospheric pressure at 20°C, and α is the particle size distribution with a diameter less than 2.5 μm.

Optionally, the described light intensity digital signal is used as an input quantity, a regularized neural network model is established, and the step of outputting the PM2.5 concentration value includes:

Initialize the neural network parameters;

The least mean square algorithm is used to train the weights of the neural network model and adjust the size of the weights.

Optionally, the initialization of neural network parameters is specifically:

The number of nodes in the hidden layer of the neural network model is selected to be equal to the number of training samples, the Gaussian function is selected as the activation function, the center of the activation function is selected as the sample data center, and the weights of the neural network and the expansion constants of the activation function are initialized.

On the other hand, the present invention provides a kind of PM2.5 concentration detection device based on neural network, comprising:

A laser detection system establishment unit, used to establish a laser detection system based on the principle of Fraunhofer diffraction;

A functional relationship verification unit, used to verify that there is a corresponding functional relationship between the light intensity digital signal and the PM2.5 concentration value based on the light intensity digital signal collected by the laser detection system and based on the Fraunhofer diffraction principle;

The neural network model calculation unit is used to use the light intensity digital signal as an input quantity, establish a regularized neural network model, and output the PM2.5 concentration value.

Optionally, the neural network model calculation unit includes:

A parameter initialization module for initializing neural network parameters;

The weight training unit is used to train the weight of the neural network model by adopting the least mean square algorithm, and adjust the size of the weight.

It can be seen from the above technical scheme that the neural network-based PM2.5 concentration detection method and device of the present invention first establishes a laser detection system based on the Fraunhofer diffraction principle, and then according to the light intensity digital signal collected by the laser detection system, Based on the principle of Fraunhofer diffraction, it is verified that there is a corresponding functional relationship between the light intensity digital signal and the PM2.5 concentration value, and the light intensity digital signal is used as an input to establish a regularized neural network model to output the PM2.5 concentration value , the method overcomes the shortcomings of the low automation degree of the PM2.5 concentration detection method in the prior art, and can realize repeated detection, high detection accuracy, and simple calculation.

Description of drawings:

Fig. 1 is the schematic flow chart of the PM2.5 concentration detection method based on neural network that the first embodiment of the present invention provides;

Fig. 2 is a schematic diagram of the neural network structure provided by the first embodiment of the present invention;

Fig. 3 is the schematic flow chart of the PM2.5 concentration detection method based on the neural network provided by the second embodiment of the present invention;

Fig. 4 is a schematic structural diagram of the laser detection system provided by the second embodiment of the present invention;

Fig. 5 is a schematic diagram of the neural network model provided by the second embodiment of the present invention for fitting 5 groups of data;

Fig. 6 is the structure diagram of the PM2.5 concentration detection device based on neural network provided by the third embodiment of the present invention;

FIG. 7 provides the comparison results of PM2.5 concentration values according to the third embodiment of the present invention.

Detailed ways:

The specific implementation manners of the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

Fig. 1 shows a schematic flowchart of a neural network-based PM2.5 concentration detection method provided by the first embodiment of the present invention. As shown in Fig. 1 , the method of this embodiment is as follows.

101. Establish a laser detection system based on the principle of Fraunhofer diffraction.

In this step, the laser detection system includes: a power supply unit, a He-Ne laser, a filter lens, a beam expander lens, an air pump, an air chamber, a signal receiving unit, a detection unit and a calculation unit;

Wherein, the power supply unit is used to provide power for the He-Ne laser, the air pump, the detection unit and the calculation unit;

The signal receiving unit includes a Fourier lens and 70 photodetectors, which are used to receive the light signal diffracted by air particles and convert it into an analog signal of light intensity;

The detection unit is used to amplify the light intensity analog signal and convert it into a light intensity digital signal;

The calculation unit is used for modeling according to the light intensity digital signal, and calculates the PM2.5 concentration value;

The power supply unit is respectively connected with the He-Ne laser, the air pump, the detection unit and the calculation unit, the He-Ne laser is connected with the filter lens, and the air chamber is respectively connected with the The beam expander lens, the air pump are connected to the signal receiving unit, the signal receiving unit is connected to the detection unit, and the detection unit is connected to the computing unit.

102. According to the light intensity digital signal collected by the laser detection system and based on the Fraunhofer diffraction principle, verify that there is a corresponding functional relationship between the light intensity digital signal and the PM2.5 concentration value.

In this step, it should be noted that before the neural network model is established, the performance of the neural network model needs to be verified, and then the validity of the method is strictly explained. Therefore, in order to obtain such a complex function mapping relationship through the neural network, it is necessary to obtain the only PM2. There is a functional correspondence between the strong digital signal and the PM2.5 concentration value.

103. Using the light intensity digital signal as an input, establish a regularized neural network model, and output a PM2.5 concentration value.

In this step, FIG. 2 shows a schematic diagram of the neural network structure provided by the first embodiment of the present invention. As shown in FIG. 2, the number of nodes in the hidden layer is selected as the number of samples, and the light intensity digital signal X _iN =(x _i1 , x _i2 ,...,x _iN ) are neural network input quantities, where i=1,2,...p are training samples, and N is the dimension of input nodes. In this embodiment, N=70, so that each Each training sample corresponds to a hidden layer node, that is, p hidden layer neurons are taken. The Gaussian function is used as the activation function of the hidden layer and the output layer to reflect the nonlinear mapping ability of the regularized neural network model, and the expression is as follows:

${\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; \&delta;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; p</span>}$

in, $\left|\right|x-C_i\left|\right|=\sqrt{{(x_1-c_{i1})}^2+{(x_2-c_{i2})}^2+...+{(x_N-c_i)}^2}$ is the two-dimensional norm, and δ _i is the expansion constant of the radial odd function;

Let the weight matrix be W=(w ₁ ,w ₂ ,...,w _p ) ^T , and the output value is PM2.5 concentration, then there is a p-dimensional hidden layer for each node output, the expression is as follows:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; p</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>$

It should be noted that in this embodiment, in order to simplify calculations, facilitate the transplantation of single-chip microcomputers, and make the system miniaturized, the center of each hidden layer basis function is selected as C _iN =X _iN , that is, the data center of each basis function corresponds to the corresponding sample itself , the radial expansion function is taken as where d _max is the maximum distance between samples.

Therefore, the present invention only needs to train the weight matrix, adjust the size of the weight to fit the mapping relationship, determine the parameters of the neural network model, and perform online or offline fitting to estimate system performance.

In the neural network-based PM2.5 concentration detection method and device of this embodiment, a laser detection system is first established based on the Fraunhofer diffraction principle, and then according to the light intensity digital signal collected by the laser detection system, based on Fraunhofer Diffraction principle, verify that there is a corresponding functional relationship between the light intensity digital signal and the PM2.5 concentration value, use the light intensity digital signal as an input, establish a regularized neural network model, and output the PM2.5 concentration value, this method overcomes the The PM2.5 concentration detection method in the prior art has the disadvantage of low degree of automation, and can realize repeated detection, with small accumulation error and simple calculation.

Fig. 3 shows a schematic flow chart of the neural network-based PM2.5 concentration detection method provided by the second embodiment of the present invention. As shown in Fig. 3, the method of this embodiment is as follows.

301. According to the laser detection system established based on the principle of Fraunhofer diffraction, collect 70 channels of photoelectric signal data.

In this step, specifically, the power supply unit converts AC power into DC power through a switching power supply, and outputs regulated 3V DC power through a three-terminal voltage stabilizing integrated circuit LM317 with a potentiometer for use by the He-Ne laser, and the He-Ne laser generates a wavelength of λ =0.6328μm red monochromatic laser, the laser emitted from the He-Ne laser is sorted by the filter lens and the beam expander lens to form a beam of parallel monochromatic light, which is used as the incident light for laser particle size measurement. The air in the air chamber is fully and evenly dispersed in the sample box, Fraunhofer diffraction occurs when the incident light passes through the air chamber, the signal receiving unit adopts a Fourier lens with a focal length of 180mm to receive the diffracted light, and converges the light to the photodetector. The detector is a fan-shaped structure, which is composed of 70 fan rings corresponding to the same central angle. Each ring detects the scattered light energy within its corresponding scattering angle range, and a small hole is opened at the center of the fan-shaped circle for optical path alignment. , the light intensity analog signal received by the detector is converted into a light intensity digital signal through the detection unit, and the calculation unit uses a computing single-chip microcomputer for modeling and calculation to obtain air particle size and PM2.5 concentration values, as shown in Figure 4, Figure 4 A schematic structural diagram of the laser detection system provided by the second embodiment of the present invention is shown.

302. According to the principle of Fraunhofer diffraction, verify that there is a corresponding functional relationship between the 70 channels of photoelectric signal data and the PM2.5 concentration value.

In this step, according to the principle of Fraunhofer diffraction, when light passes through a particle with a diameter of D, the distribution of diffracted light intensity at any angle is calculated by the following formula,

$\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\frac{{\mathrm{<span\; class="notranslate">\; \&pi;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; D.</span>}}^{\mathrm{<span\; class="notranslate">\; 4</span>}}}{\mathrm{<span\; class="notranslate">\; 16</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{<span\; class="notranslate">\; [</span>\frac{{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; ]</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}$

Among them, I(θ) is the intensity distribution of diffracted light under any angle of light intensity digital signal, I ₀ is the intensity of incident light, f is the focal length of Fourier lens, D is the diameter of the air particle, λ is the wavelength of the incident light, J ₁ (X) is the first-order Bessel function, then the diffracted light is distributed in the nth detection ring of the detector (the ring radius is from S _n to S _n+1 , The light energy on the corresponding angle from θ _n to θ _n+1 ) is calculated by the following formula,

${\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; \&Integral;</span>}_{{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}}^{{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}}\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; \&pi;</span>}\mathrm{<span\; class="notranslate">\; sin</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; d\&theta;</span>}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; \&Integral;</span>}_{{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}}^{{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}}\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; \&pi;SdS</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; )</span>$

Since θ is small, then Putting it into the above formula and recursing the formula by the Bessel function, the expression is as follows,

${\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; \&pi;D</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 8</span>}}{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>$

in, J ₀ represents the zero-order Bessel function.

Assuming that the collected air is composed of particles with M diameters, and assuming that there are Q _i particles with a diameter of D _i , the total diffracted light energy of the particle group at the nth ring is calculated by the following formula,

${\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; \&pi;I</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; 8</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>$

The usual particle distribution is expressed in parts by weight, and the relationship between weight and number is calculated by the following formula,

${\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; 6</span>}\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{\mathrm{<span\; class="notranslate">\; \&pi;\&rho;</span>}{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; 3</span>}}}$

Wherein, ρ is the density of the particulate matter, and further, the total diffracted light energy of the nth ring of the particle group is calculated by the following formula,

${\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}\mathrm{<span\; class="notranslate">\; \&rho;</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\frac{{\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>$

Assuming that the detector consists of N rings, a system of linear equations consisting of N equations can be established

$\begin{array}{c}{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 3</span>}{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}\mathrm{<span\; class="notranslate">\; \&rho;</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\frac{{\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\\ {\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}\mathrm{<span\; class="notranslate">\; \&rho;</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\frac{{\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\\ <span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\\ {\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; N</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}\mathrm{<span\; class="notranslate">\; \&rho;</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\frac{{\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{{\mathrm{<span\; class="notranslate">\; D.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; J</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\end{array}$

The above formula can be written in matrix form as follows,

E=TW

Among them, E is the light energy distribution vector measured by the photodetector, $T=\left[\begin{array}{ccc}{t}_{11}& ...& t_{1m}\\ ...& & \\ t_{N1}& ...& t_{\mathrm{N\; M}}\end{array}\right]$ is the light energy contribution matrix, and W is the particle size distribution vector of the particle group to be measured;

in, $t_{\mathrm{N\; M}}=C\frac{1}{{\mathrm{D.}}_m}[J_0^2\left(x_{m,N}\right)+J_1^2\left(x_{m,N}\right)-J_0^2\left(x_{m,N-1}\right)-J_1^2\left(x_{m,N+1}\right)]$ is the light energy contribution of a particle with a diameter of D _M to the Nth detection ring of the detector, is a fixed constant, M is the type of particles with different diameters, I ₀ is the intensity of incident light, J ₀ (X _M,N ) is the zero-order Bessel function, and J ₁ (X _M,N ) is the first-order Bessel function.

According to the above linear equations, the particle size distribution of the particle group to be tested can be obtained, and the PM2.5 particle size α can be obtained by adding the particle sizes with a diameter less than 2.5 μm. Then, the PM2.5 concentration can be calculated by the following formula,

Wherein, β _{air density} is 1.205 kg/m ³ at standard atmospheric pressure at 20°C, and α is the particle size distribution with a diameter less than 2.5 μm.

Thus, according to the light energy distribution vector, the particle size distribution of the particle group to be measured is obtained, and the particle size distribution percentages with diameters less than 2.5 μm are added and multiplied by the air density to obtain the PM2.5 concentration value, which verifies the light intensity There is a corresponding functional relationship between the digital signal and the PM2.5 concentration value.

303. Using the light intensity digital signal as an input, establish a regularized neural network model.

304. Initialize neural network parameters.

In this step, specifically: select the number of nodes in the hidden layer of the neural network model equal to the number of training samples, select the Gaussian function as the activation function, select the center of the activation function as the sample data center, initialize the weights of the neural network and expand the activation function constant.

305. Using the standard PM2.5 concentration value as the output, the least mean square algorithm is used to train the weights of the neural network model, and the size of the weights is adjusted.

In this step, it should be explained that the photoelectric data of the laser detection system and the corresponding real-time PM2.5 concentration value need to be collected, and the data can be collected every one hour continuously, and at the same time, the corresponding real-time PM2. 5 data, a total of 200 sets of data were collected. In order to prevent the accidental and inaccurate detection results and enhance the representativeness of the data, 100 sets of data with a large dynamic range and a certain interval were selected as the training set to establish a regularized neural network. In the remaining 100 sets of data, 30 data were selected as the test set to verify the established regularized neural network model.

In this embodiment, the weight value can be initialized to any value. Using the sample data and the known standard PM2.5 value, the least mean square algorithm learning rule is used for training. The learning signal expression is as follows,

e _j ＝d _j -W ^T G _j (x) j＝1,2,...p

The expression of weight vector adjustment is as follows,

△W=ηe _j G _j (x)

Then the expression of the right to adjust the amount is as follows,

W(k+1)=W(k)+△W

Wherein, η is the learning rate, and η is initialized to any positive number from 0 to 1, as shown in FIG. 5 , which shows a schematic diagram of the neural network model provided by the second embodiment of the present invention for fitting 5 sets of data.

In order to guide the modeling and calculation in the computing unit, it is necessary to learn and train the data as efficiently as possible while ensuring convergence. Therefore, the number of iterations j and the fitting error δ are used to evaluate the performance of the neural network. Appropriately increasing the school rate η can reduce the number of iterations j and speed up data fitting, but too much η will cause the neural network fitting to diverge and the system to be unstable, that is, the number of iterations represents the computational complexity of the system, and the fitting error δ is The amount of deviation allowed for each set of data fitting. The larger the fitting error δ, the fitting accuracy decreases, but the corresponding number of iterations decreases, and the fitting error δ represents the calculation accuracy of the neural network algorithm. In the present invention, since the official PM2.5 data accuracy is an integer number of digits, it is sufficient to set the fitting error δ to be less than or equal to 1 in the training termination condition.

306. Input the collected light intensity digital signal, and obtain the PM2.5 concentration value output according to the neural network model.

In this embodiment, based on the principle of Fraunhofer diffraction, a laser detection system is established, and it is verified that there is an actual mapping relationship between the collected photoelectric data and the PM2. Learning and training to obtain the required mapping relationship, so as to efficiently and accurately detect PM2.5 concentration, so that the entire monitoring system has self-learning ability and has high practical value.

Fig. 6 shows the structural representation of the PM2.5 concentration detection device based on the neural network provided by the third embodiment of the present invention, as shown in Figure 6, the PM2.5 concentration detection device based on the neural network in the present embodiment includes: laser A detection system establishment unit 61 , a functional relationship verification unit 62 , and a neural network model calculation unit 63 .

The laser detection system establishment unit 61 is used to establish a laser detection system based on the principle of Fraunhofer diffraction;

The functional relationship verification unit 62 is used to verify that there is a corresponding functional relationship between the light intensity digital signal and the PM2.5 concentration value based on the light intensity digital signal collected by the laser detection system and based on the Fraunhofer diffraction principle;

The neural network model calculation unit 63 is used to use the light intensity digital signal as an input quantity, establish a regularized neural network model, and output a PM2.5 concentration value.

Wherein, the neural network model calculation unit also includes:

A parameter initialization module for initializing neural network parameters;

The weight training module is used to train the weight of the neural network model by adopting the least mean square algorithm, and adjust the size of the weight.

Finally, by combining the standard PM2.5 concentration value of the domestic environmental monitoring official website, the PM2.5 concentration value calculated based on the Fraunhofer diffraction model, and the PM2.5 concentration value calculated based on the neural network in the present invention, Comparing the PM2.5 concentration data in a certain month in Shenyang City, the comparison results of the above PM2.5 concentration values are shown in Figure 7. It can be seen that the PM2.5 concentration value obtained by using the neural network algorithm is consistent with the results of domestic official websites. The small deviation shows that the neural network-based PM2.5 concentration detection method and device of the present invention is feasible, efficient in calculation, and high in detection accuracy.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: It is still possible to modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions depart from the scope defined by the claims of the present invention .

Claims (9)

1. A PM2.5 concentration detection method based on a neural network is characterized by comprising the following steps:

establishing a laser detection system based on the Fraunhofer diffraction principle;

verifying that the light intensity digital signal has a corresponding function relation with a PM2.5 concentration value based on a Fraunhofer diffraction principle according to the light intensity digital signal acquired by the laser detection system;

and establishing a regularization neural network model by taking the light intensity digital signal as an input quantity, and outputting a PM2.5 concentration value.

2. The method of claim 1, wherein the laser detection system comprises: the device comprises a power supply unit, a He-Ne laser, a filter lens, a beam expanding lens, an air pump, an air chamber, a signal receiving unit, a detection unit and a calculation unit;

wherein the power supply unit is used for providing power supply for the He-Ne laser, the air pump, the detection unit and the calculation unit;

the signal receiving unit comprises a Fourier lens and 70 paths of photodetectors and is used for receiving the optical signals diffracted by the air particles and converting the optical signals into light intensity analog signals;

the detection unit is used for converting the light intensity analog signal into a light intensity digital signal after amplification processing;

the calculation unit is used for modeling according to the light intensity digital signal and calculating to obtain a PM2.5 concentration value;

the power supply unit is respectively connected with the He-Ne laser, the air pump, the detection unit and the calculation unit, the He-Ne laser is connected with the filter lens, the air chamber is respectively connected with the beam expanding lens, the air pump and the signal receiving unit, the signal receiving unit is connected with the detection unit, and the detection unit is connected with the calculation unit.

3. The method of claim 1, wherein the verifying the digital light intensity signal according to the collected digital light intensity signal of the laser detection system based on Fraunhofer diffraction principle has a corresponding function relationship with the concentration value of PM2.5 comprises:

and calculating and processing the light energy distribution vector measured by the photoelectric detector to obtain the particle size distribution of the particle group to be measured, adding the particle size distribution percentages with the diameters smaller than 2.5 mu m to obtain a PM2.5 concentration value, and verifying that the light intensity digital signal and the PM2.5 concentration value have a corresponding relation.

4. The method according to claim 3, wherein the particle size distribution of the particle group to be measured is calculated by the following formula,

E＝TW

wherein E is the light energy distribution vector measured by the photoelectric detector, $T=\left[\begin{array}{ccc}{t}_{11}& ...& t_{1M}\\ ...& & \\ t_{N1}& ...& t_{\mathrm{NM}}\end{array}\right]$ w is a particle size distribution vector of the particle group to be detected;

wherein, $t_{\mathrm{NM}}=C\frac{1}{D_M}[J_0^2\left(X_{M,N}\right)+J_1^2\left(X_{M,N}\right)-J_0^2\left(X_{M,N+1}\right)-J_1^2\left(X_{M,N+1}\right)]$ is of diameter D_MThe light energy of the nth detection ring of the detector,for a fixed constant, M is the type of particles of different diameter, I₀Is the intensity of incident light, J₀(X_M,N) Is a zero order Bessel function, J₁(X_M,N) Is a first order Bessel function.

5. The method of claim 1, wherein the PM2.5 concentration value is calculated by the following formula,

wherein, beta_{Density of air}At 20 deg.C under normal atmospheric pressureThe value is 1.205kg/m³And alpha is the particle size distribution with a diameter of less than 2.5 μm.

6. The method according to claim 1, wherein the step of establishing a regularized neural network model by using the light intensity digital signal as an input quantity and outputting a value of the concentration of PM2.5 comprises:

initializing neural network parameters;

and training the weight of the neural network model by adopting a least mean square algorithm, and adjusting the weight.

7. The method according to claim 6, wherein the initializing neural network parameters are specifically:

selecting the number of hidden layer nodes of the neural network model to be equal to the number of training samples, selecting a Gaussian function as an excitation function, selecting the center of the excitation function as a sample data center, and initializing the weight and excitation function expansion constant of the neural network.

8. A PM2.5 concentration detection device based on a neural network is characterized by comprising:

the laser detection system establishing unit is used for establishing a laser detection system based on the Fraunhofer diffraction principle;

the functional relationship verification unit is used for verifying that the light intensity digital signal and the PM2.5 concentration value have a corresponding functional relationship based on the Fraunhofer diffraction principle according to the light intensity digital signal acquired by the laser detection system;

and the neural network model calculation unit is used for establishing a regularization neural network model by taking the light intensity digital signal as an input quantity and outputting a PM2.5 concentration value.

9. The apparatus of claim 8, wherein the neural network model computation unit comprises:

the parameter initialization module is used for initializing neural network parameters;

and the weight training module is used for training the weight of the neural network model by adopting a least mean square algorithm and adjusting the size of the weight.