Ultrafine particle identification counting method and device based on photoelectric effect and storage medium

Abstract

The present application relates to the field of particle detection, and specifically provides a method and device for identifying and counting ultrafine particles based on photoelectric effect, and a storage medium, comprising: obtaining photoelectric signals generated by a flow of particles to be detected passing through a photoelectric detection array; extracting original features of the photoelectric signals, combining time domain decomposition and frequency domain transformation to construct an initial signal feature matrix; performing multi-scale tensor decomposition and signal component screening on the initial signal feature matrix to generate enhanced photoelectric signals; determining a polarization coherence angle interval based on the enhanced photoelectric signals, extracting polarization energy features, analyzing polarization state conversion characteristics, and establishing a three-dimensional polarization feature fingerprint; performing adaptive segmentation processing on the enhanced photoelectric signals, extracting oscillation feature spectrum, and combining the three-dimensional polarization feature fingerprint to obtain a feature correlation matrix; and constructing a multi-dimensional classification criterion based on the feature correlation matrix, and determining a final particle count result accordingly. The present application improves the accuracy and recognition ability of ultrafine particle counting, and is particularly suitable for high-density micro-nano particle detection in complex environments.

Description

Technical Field

[0001] This invention belongs to the field of particle detection technology, and particularly relates to a method, device and readable storage medium for identifying and counting ultrafine particles based on the photoelectric effect. Background Technology

[0002] The identification and counting of ultrafine particles has significant application value in fields such as environmental monitoring, medical diagnosis, biosafety, and industrial quality control. Currently, particle detection technology based on the photoelectric effect is the mainstream method for achieving ultrafine particle counting. This technology utilizes the principle of optical scattering; when a particle passes through the detection area, it causes a change in the intensity of scattered light, which is then converted into an electrical signal for analysis and processing, enabling particle identification and counting. With the development of technology and the increasing demands of applications, higher requirements are being placed on the accuracy of ultrafine particle counting, the accuracy of identification, and adaptability to complex environments.

[0003] Existing photoelectric effect-based ultrafine particle identification and counting technologies still have shortcomings. Traditional single-detection-channel photoelectric detection systems are prone to signal overlap and occlusion when processing particle samples with high concentrations or wide size distributions, leading to a decrease in counting accuracy. Existing technologies do not adequately utilize polarization information; most systems only focus on scattering intensity while ignoring the polarization changes caused by particles, lacking in-depth analysis capabilities of particle shape, structure, and other characteristics, and failing to effectively distinguish particles with similar shapes but different compositions. Current particle counting technologies lack effective signal enhancement and noise suppression mechanisms when facing complex background interference, especially when multiple particles pass through the detection area simultaneously, making accurate single-particle separation and feature extraction difficult, thus limiting the system's performance in practical applications. Summary of the Invention

[0004] In view of this, the present invention aims to provide a method, device and readable storage medium for ultrafine particle identification and counting based on photoelectric effect, so as to improve the ability to distinguish ultrafine particles with similar shapes and sizes, and make particle counting in complex environments more resistant to interference and adaptable.

[0005] To achieve the above objectives, the technical solution created by this invention is implemented as follows:

[0006] In a first aspect, the present invention provides a method for identifying and counting ultrafine particles based on the photoelectric effect, comprising:

[0007] S100: Acquire the photoelectric signal generated by the particle stream to be tested passing through the photoelectric detection array, wherein the photoelectric detection array has multiple detection channels;

[0008] S200. Extract the original features of the photoelectric signals in each of the detection channels, and construct an initial signal feature matrix by combining time-domain decomposition and frequency-domain transformation;

[0009] S300. Perform multi-scale tensor decomposition and signal component screening on the initial signal feature matrix, and combine the detection channel collaborative compensation with wavefront propagation characteristics to generate an enhanced photoelectric signal.

[0010] S400. Based on the enhanced photoelectric signal, determine the polarization coherence angle range, extract polarization energy features in the polarization coherence angle range, analyze polarization state conversion characteristics, and establish a three-dimensional polarization feature fingerprint.

[0011] S500 performs adaptive segmentation processing on the enhanced photoelectric signal and extracts the oscillation feature spectrum;

[0012] S600. Perform deep correlation matching between the oscillation feature spectrum and the three-dimensional polarization feature fingerprint to obtain a feature correlation matrix;

[0013] S700. Construct a multidimensional classification criterion based on the feature correlation matrix, and perform feature separation on the particles to be tested accordingly to determine the final particle counting result.

[0014] Furthermore, the original features include scattering intensity distribution features and polarization state conversion features; step S200 includes:

[0015] S210. Perform time-domain decomposition on the original features to obtain the signal envelope, and simultaneously perform frequency-domain transformation to obtain a frequency-domain feature set;

[0016] S220. The signal envelopes of adjacent detection channels are correlated and matched by a spatial mapping function, and a light intensity modulation curve is generated by combining the frequency domain feature set.

[0017] S230. Dynamically segment the light intensity modulation curve, extract the frequency modulation depth, phase modulation characteristics and polarization state transition law of each segment of the signal, and construct a modulation information sequence.

[0018] S240. Construct a weight coefficient matrix based on the modulation information sequence, and establish a signal correlation matrix between the detection channels;

[0019] S250. The time-domain features, frequency-domain features and scattering intensity distribution in the signal correlation matrix are fused in a hierarchical manner to generate an initial signal feature matrix.

[0020] Furthermore, step S300 includes:

[0021] S310. Perform multi-scale tensor decomposition on the initial signal feature matrix to obtain the spatial gradient tensor of scattering intensity and the time evolution tensor of polarization state, and construct an adaptive threshold boundary accordingly.

[0022] S320. The complex domain trajectory of the signal component is projected using the adaptive threshold boundary, and the effective signal components are selected. Nonlinear weighted fusion is performed on the effective signal components to obtain the scale fusion signal.

[0023] S330. Extract wavefront propagation direction features and wavefront arrival time features from the scale fusion signal, calculate the direction deviation and time deviation between the detection channels, and construct a spatiotemporal coordination index to determine the reference channel and the channel to be compensated.

[0024] S340. Extract the spatiotemporal evolution mode of scattering intensity and the polarization state rotation evolution mode of the reference channel, construct the signal manifold topology, map the channel to be compensated to the signal manifold topology, and obtain the manifold space reconstruction result.

[0025] S350. Perform phase space alignment and amplitude normalization processing on the reconstructed manifold space result to generate an enhanced photoelectric signal.

[0026] Furthermore, step S330 includes:

[0027] S331. Perform spatial domain gradient field analysis on the scale-fused signal to extract wavefront propagation direction features;

[0028] S332. Perform time-domain peak localization on the scale-fused signal to obtain wavefront arrival time characteristics;

[0029] S333. For each of the detection channels, select the neighboring detection channels within a preset neighborhood range, and calculate the deviation index of the wavefront propagation characteristics based on the wavefront arrival time characteristics.

[0030] S334. Weighted fusion of the wavefront propagation direction deviation and the wavefront arrival time deviation is performed to construct a space-time coordination index.

[0031] S335. The detection channels whose spatiotemporal coordination index is higher than the preset coordination verification threshold are determined as reference channels, and the remaining detection channels are determined as channels to be compensated.

[0032] Furthermore, the S400 includes:

[0033] S410. Generate a polarization coherence angle range based on the polarization component of the detection channel in the enhanced photoelectric signal;

[0034] S420. Extract the intensity of linear polarization component and circular polarization component within the polarization coherence angle interval, and perform time-by-time ratio calculation to generate a polarization energy ratio time series.

[0035] S430. Construct a two-dimensional polarization energy matrix based on the polarization energy ratio time series, identify the moment when the polarization state undergoes a jump, and determine the moment when the polarization energy conversion reverses.

[0036] S440. Within a preset time window before and after the polarization energy conversion reversal moment, extract the difference in the rate of change of the polarization energy ratio to generate a polarization degree jump amplitude sequence.

[0037] S450. At the moment of polarization energy conversion reversal, calculate the coordinate values ​​of the polarization component in the Stokes parametric space and generate the polarization state displacement vector.

[0038] S460. Perform spatial orientation clustering on the polarization state displacement vector, identify valid orientation clusters, and generate polarization state transition features;

[0039] S470. Combine the polarization degree jump amplitude sequence, the polarization state transition feature, and the polarization coherence angle interval to construct a three-dimensional polarization feature fingerprint.

[0040] Furthermore, step S430 includes:

[0041] S431. Construct a two-dimensional polarization energy matrix by dividing the polarization energy ratio time series into linear polarization components and circular polarization components. Calculate the polarization state transition intensity and polarization state transition direction based on the two-dimensional polarization energy matrix to generate a polarization state transition feature sequence.

[0042] S432. Perform phase unwrapping processing on the polarization state transition feature sequence, extract the phase jump point, and determine the time corresponding to the phase jump point as the first candidate reversal time point;

[0043] S433. Calculate the covariance matrix of polarization state transition intensity and polarization state transition direction at the candidate reversal time point, determine the stability of polarization state transition based on the eigenvalues ​​of the covariance matrix, and select a second candidate reversal time point from the first candidate reversal time points.

[0044] S434. Calculate the rate of change of the scattering cross section at the second candidate reversal time point, and determine the candidate reversal time point where the rate of change of the scattering cross section is greater than a preset rate of change threshold as the polarization energy conversion reversal time point.

[0045] Furthermore, step S500 includes:

[0046] S510. Divide the enhanced photoelectric signal into multiple sub-signal segments; calculate the variance of the time interval sequence of signal zero-crossing points in each sub-signal segment, and determine the oscillation complexity of the sub-signal segment accordingly.

[0047] S520. Determine the frequency sampling density based on the oscillation complexity, and perform frequency domain decomposition on the sub-signal segment accordingly to determine the frequency domain energy distribution;

[0048] S530. In the frequency domain energy distribution of the sub-signal segment, the peak frequency is time-tracked, and the peak frequency with a frequency drift amplitude less than a preset drift threshold is identified as a stable oscillation frequency. The energy proportion corresponding to the stable oscillation frequency is extracted as a time-varying sequence to determine the energy modulation evolution trajectory. The stable oscillation frequency and the energy modulation evolution trajectory are combined to form an oscillation feature spectrum.

[0049] Furthermore, the multidimensional classification criteria in step S700 include velocity distribution, acceleration distribution, size features, and shape features; after feature separation of the particle to be tested, the method further includes generating a feature identifier for a single particle to be tested.

[0050] In a second aspect, the present invention provides a computer device comprising:

[0051] At least one processor; and

[0052] A memory communicatively connected to the at least one processor; wherein,

[0053] The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the photoelectric effect-based ultrafine particle identification and counting method provided in any embodiment of the present invention.

[0054] Thirdly, the present invention provides a non-transient computer-readable storage medium storing computer instructions, the computer instructions being used to cause the computer to execute the photoelectric effect-based ultrafine particle identification and counting method provided in any embodiment of the present invention.

[0055] Compared with existing technologies, this invention combines time-domain decomposition, frequency-domain transformation, and tensor decomposition techniques to establish a three-dimensional polarization fingerprint and oscillation feature spectrum of particles. By forming a feature correlation matrix through deep correlation matching, it improves the resolution of ultrafine particles with similar shapes and sizes, especially showing better counting accuracy in high-density particle streams. Based on the collaborative analysis method of polarization state features and oscillation features, a multi-dimensional classification criterion including velocity distribution, acceleration distribution, size features, and shape features is constructed, realizing accurate identification of submicron-level ultrafine particles. It has stronger anti-interference ability and adaptability for particle counting in complex environments. Attached Figure Description

[0056] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments and descriptions of the invention are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0057] Figure 1 A schematic flowchart of the photoelectric effect-based ultrafine particle identification and counting method described in the embodiments of the present invention;

[0058] Figure 2 A flowchart illustrating step S300 in the photoelectric effect-based ultrafine particle identification and counting method described in the embodiment of the present invention;

[0059] Figure 3 A schematic diagram of the structure of the computer device described in the embodiments of the present invention;

[0060] Explanation of reference numerals in the attached figures:

[0061] 62. Computer equipment; 64. Peripheral devices; 66. Processor / processing unit; 68. Bus; 70. Network adapter; 72. Input / output (I / O) interface; 74. Display; 78. System memory; 80. Random access memory (RAM); 82. Cache memory; 84. Storage system; 92. Program module; 90. Program / utility. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and do not constitute a limitation thereof. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the invention. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, some operations related to the invention are not shown or described in the specification. This is to avoid obscuring the core parts of the invention with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0063] It should be noted that, unless otherwise specified, the embodiments and features described in this invention can be combined to form various implementations. Furthermore, the order of the steps or actions in the method description can be changed or adjusted in a manner readily apparent to those skilled in the art. Therefore, the various orders in the specification and drawings are merely for the clear description of a particular embodiment and do not imply a mandatory order, unless otherwise stated that a particular order must be followed.

[0064] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this invention. The term "based on" should be understood as "at least partially based on." Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more, and the term "including" means "including but not limited to." Various embodiments of the present invention may exist in the form of a range; it should be understood that the description in the form of a range is merely for convenience and brevity and should not be construed as a rigid limitation on the scope of the invention; therefore, it should be considered that the range description has specifically disclosed all possible sub-ranges and single numerical values ​​within that range; for example, it should be considered that the range description from 1 to 6 has specifically disclosed sub-ranges, such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6, etc., and single numbers within the range, such as 1, 2, 3, 4, 5, and 6, regardless of the range. Furthermore, whenever a numerical range is referred to herein, it means including any referenced number (fraction or integer) within the range referred to.

[0065] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0066] The invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0067] Example 1

[0068] like Figure 1 As shown, this invention provides a method for identifying and counting ultrafine particles based on the photoelectric effect, comprising:

[0069] S100: Acquire the photoelectric signal generated by the particle stream to be tested passing through the photoelectric detection array, wherein the photoelectric detection array has multiple detection channels;

[0070] A photoelectric detection array is used to acquire the photoelectric signal generated by the particle stream under test. The photoelectric detection array contains multiple linearly arranged detection channels, each consisting of a light source, an optical focusing element, and a photodetector. When the particle stream passes through the detection area, each detection channel independently acquires the particle scattered light signal, with a sampling frequency set to 100 kHz and a sampling duration of 5 seconds.

[0071] To ensure signal quality, the incident light wavelength of the photoelectric detection array can be selected as 532nm, the light source power is 5mW, the detection channel spacing is 1mm, and the detector sensitivity reaches 0.1mV / μW.

[0072] In practice, after the acquisition is completed, a signal quality check can be performed; if it is qualified, proceed to step S200; if it is unqualified or the noise is too high, the photoelectric signal is reacquired.

[0073] S200. Extract the original features of the photoelectric signals in each of the detection channels, and construct an initial signal feature matrix by combining time-domain decomposition and frequency-domain transformation;

[0074] The photoelectric signals from adjacent detection channels are processed to extract scattering intensity and polarization state characteristics. Scattering intensity characteristics are characterized by calculating the signal peak value, integral area, and peak width; typical values ​​are a peak value range of 3.5–5.2 V, an integral area of ​​30–150 V·μs, and a peak width of 2–15 μs. Polarization state characteristics are determined by measuring the relative intensity ratio of the horizontal and vertical polarization components, with a normal range of 0.2–5.0.

[0075] Through time-domain decomposition, the photoelectric signal of each detection channel is divided into rising, peak, and falling segments, with typical durations of 0.5–2 μs, 1–3 μs, and 2–10 μs, respectively. Frequency-domain transformation employs a 512-point fast calculation method to extract spectral features within the 0–50 kHz range, with a focus on the 1–10 kHz band. An initial signal feature matrix is ​​established by calculating the time delay correlation and spectral similarity of the photoelectric signals from adjacent detection channels. This initial signal feature matrix contains feature data in four dimensions: intensity, time delay, spectrum, and polarization of the photoelectric signal from each detection channel.

[0076] S300. Perform multi-scale tensor decomposition and signal component screening on the initial signal feature matrix, and combine the detection channel collaborative compensation with wavefront propagation characteristics to generate an enhanced photoelectric signal.

[0077] For the initial signal feature matrix, multi-scale tensor decomposition and signal component filtering operations are performed. Multi-scale tensor decomposition decomposes the initial signal feature matrix into three core components: time pattern, detection channel pattern, and feature pattern. For the time pattern, the decomposition order is set to 5, retaining the main components with a contribution rate exceeding 85%. For the detection channel pattern, a threshold of 10% of the original signal energy is set, filtering out noise components below this threshold. For the feature pattern, the components are sorted according to their eigenvalues, retaining the top few with a cumulative contribution rate reaching 95%. Combined with a detection channel collaborative compensation technique based on wavefront propagation characteristics, and based on the signal propagation delay patterns of adjacent detection channels (typical delay values ​​of 5–20 μs), missing or distorted signals are repaired, achieving a compensation accuracy of over 90% of the original signal, ultimately generating an enhanced photoelectric signal.

[0078] S400. Based on the enhanced photoelectric signal, determine the polarization coherence angle range, extract polarization energy features in the polarization coherence angle range, analyze polarization state conversion characteristics, and establish a three-dimensional polarization feature fingerprint.

[0079] When determining the polarization coherence angle range, the rate of change of the polarization direction can be calculated by analyzing the polarization state changes of signals from adjacent detection channels, thus identifying the angle range with the strongest polarization coherence, typically 15°–75°. In practice, this can be achieved through multi-channel polarization detection, with each channel responsible for acquiring polarization components at a specific angle. Within this angle range, polarization energy features are extracted, including degree of polarization (range 0.1–0.9), polarization direction (0°–180°), and polarization ellipticity (0.1–5.0). By analyzing the polarization state transition characteristics during particle passage, the dynamic change trajectory of polarization parameters is recorded, constructing a three-dimensional polarization fingerprint composed of three dimensions: degree of polarization, polarization direction, and polarization ellipticity. The number of feature points is typically 10–30.

[0080] S500 performs adaptive segmentation processing on the enhanced photoelectric signal and extracts the oscillation feature spectrum;

[0081] Segmentation points are set based on changes in signal smoothness, and adaptive segmentation processing is performed on the enhanced photoelectric signal, typically with 3 to 7 segments. Smoothness is calculated based on the rate of change of adjacent data points, and the threshold is set to 20% of the signal standard deviation. Frequency domain analysis based on oscillation complexity is performed on each segment to extract the oscillation characteristic spectrum, which includes the dominant frequency, harmonic distribution, and spectral energy distribution. The typical range of the dominant frequency is 2–15 kHz, the first 5 harmonic components are generally retained, and the spectral energy distribution focuses on the frequency band below 10 kHz.

[0082] S600. Perform deep correlation matching between the oscillation feature spectrum and the three-dimensional polarization feature fingerprint to obtain a feature correlation matrix;

[0083] The oscillation feature spectrum and the three-dimensional polarization feature fingerprint are deeply correlated and matched. The feature distance calculation method is used, and the threshold is set to 0.3 to generate a feature correlation matrix characterizing the particle properties.

[0084] S700. Construct a multidimensional classification criterion based on the feature correlation matrix, and perform feature separation on the particles to be tested accordingly to determine the final particle counting result.

[0085] A multidimensional classification criterion is constructed based on the feature correlation matrix, which includes velocity distribution, acceleration distribution, size features, and shape features. The velocity distribution is calculated based on the time delay between adjacent detection channels, typically ranging from 0.5 to 10 m / s; the acceleration distribution is determined by the rate of change of velocity, ranging from ±50 m / s²; the size feature is estimated based on the width of the scattered signal, achieving a resolution of 0.5 μm; and the shape feature is identified through polarization scattering patterns, distinguishing basic shapes such as spheres, rods, and sheets. The multidimensional classification criterion adopts a hierarchical structure, first grouping by velocity range, and then further subdividing based on size, acceleration, and shape features, forming a decision tree structure.

[0086] Based on the established multidimensional classification criteria, the particles under test are feature-separated, and the particle signals are divided into three categories according to velocity range: fast (>5m / s), medium (1-5m / s), and slow (<1m / s). Within each velocity category, they are further classified according to size characteristics (small <2μm, medium 2-10μm, large >10μm). For particle overlap phenomena, identification is achieved through the number of peaks in the oscillation characteristic spectrum and abrupt changes in polarization characteristics, with a success rate exceeding 85%. The final particle counting results show an accuracy rate exceeding 95%. In practical implementation, a feature identifier for each individual particle under test can also be generated, i.e., for each identified individual particle, a feature identifier containing timestamp, velocity, size, shape parameters, and polarization characteristics is generated.

[0087] The photoelectric effect-based ultrafine particle identification and counting method provided in this embodiment can effectively process complex signals generated by multi-channel photoelectric detection arrays, achieve high-precision particle feature separation and counting, and is suitable for applications requiring high-precision particle counting, such as air quality monitoring and biological sample analysis.

[0088] Example 2

[0089] Based on Example 1, the present invention also provides a method for identifying and counting ultrafine particles based on the photoelectric effect, which also includes steps S100 to S700, and will not be described in detail here.

[0090] In this embodiment, step S200 includes:

[0091] S210. Perform time-domain decomposition on the original features to obtain the signal envelope, and simultaneously perform frequency-domain transformation to obtain a frequency-domain feature set;

[0092] The aforementioned original features include scattering intensity distribution characteristics and polarization state transition characteristics; among which, the scattering intensity distribution characteristics include statistical parameters such as signal mean, variance, and peak value. For example, for the signal acquired by detection channel 1, its average scattering intensity is 2.36 mW / cm². 2 The variance is 0.18mW. 2 / cm 4 The polarization state conversion characteristics include the rate of change of polarization degree and the change of polarization direction angle. For example, the rate of change of polarization degree of detection channel 1 is 0.032 / ms, and the change of polarization direction angle is 0.85° / ms. These characteristics are combined to form the original features.

[0093] The original feature was decomposed in the time domain, and the signal envelope was extracted using the Hilbert transform method. An analytical signal was constructed for each detection channel's original signal, and its magnitude was calculated to obtain the signal envelope. For example, the signal envelope extraction result for detection channel 2 showed that its peak value was 3.21 mW / cm². 2 The envelope period is 25ms. A Fast Fourier Transform (FFT) is performed simultaneously to obtain frequency domain features, with a transform window length of 1024 points, an overlap rate of 50%, and Hanning window weighting. The frequency domain feature set includes parameters such as the dominant frequency component, harmonic component intensity ratio, and spectral width. Actual analysis shows that the dominant frequency of the signal in detection channel 2 is 42Hz, the subharmonic intensity ratio is 0.37, and the spectral width is 12Hz.

[0094] S220. The signal envelopes of adjacent detection channels are correlated and matched by a spatial mapping function, and a light intensity modulation curve is generated by combining the frequency domain feature set.

[0095] A spatial mapping function is used to process the signal envelopes of adjacent detection channels to establish a correlation matching relationship. This mapping function employs a Gaussian kernel design with a kernel width parameter set to 0.75 and a matching threshold of 0.82. Taking detection channel 1 and detection channel 2 as an example, the correlation coefficient calculated through spatial mapping is 0.88, exceeding the matching threshold, confirming a strong correlation between the two. Combined with the frequency domain feature set, an intensity modulation curve is generated, containing signal amplitude changes and phase modulation information. The intensity modulation curve has 500 sampling points and a time resolution of 0.5 ms. The modulation curve clearly shows an amplitude jump at 32 ms, with the amplitude increasing from 2.1 mW / cm². 2 Increased to 2.9 mW / cm 2 A phase abrupt change occurs at 57ms, with a phase difference of 42°.

[0096] S230. Dynamically segment the light intensity modulation curve, extract the frequency modulation depth, phase modulation characteristics and polarization state transition law of each segment of the signal, and construct a modulation information sequence.

[0097] Dynamic segmentation is performed based on the characteristics of the intensity modulation curve to detect amplitude jump points and phase abrupt changes. The amplitude jump threshold is set to 15% of the signal average, and the phase abrupt change threshold is set to 30°. Within each detected segment, the frequency modulation depth, phase modulation characteristics, and polarization state transition law are extracted. Taking the third segment as an example, its frequency modulation depth is 0.24, the phase modulation characteristics exhibit linear changes with a slope of 0.67° / ms, and the polarization state transition law shows that the polarization degree is stable at 0.83, while the polarization direction angle changes periodically with a period of 38ms. These parameters are organized into a modulation information sequence, with a sequence length equal to the number of segments, and each segment contains 8 feature parameters.

[0098] S240. Construct a weight coefficient matrix based on the modulation information sequence, and establish a signal correlation matrix between the detection channels;

[0099] A weighted coefficient matrix is ​​constructed based on the modulation information sequence, with dimensions equal to the number of detection channels multiplied by the number of features. The initial weights are all set to 0.5. For adjacent detection channels 1 and 2, the weight coefficients are dynamically adjusted based on their signal correlation. The adjustment formula is based on the correlation coefficient and signal-to-noise ratio. After adjustment, the weight of the scattering intensity feature of detection channel 1 is increased to 0.72, and the weight of the polarization state feature is decreased to 0.35; the corresponding weights for detection channel 2 are 0.65 and 0.42, respectively. A signal correlation matrix between the detection channels is constructed using these weights. The matrix element values ​​range from 0 to 1, representing the correlation strength of signal features between different detection channels.

[0100] To eliminate signal distortion, a polarization compensation correction mechanism is introduced. This mechanism calculates polarization compensation parameters by detecting abnormal changes in polarization state. For example, when detection channel 3 detects a sudden drop in polarization degree from 0.78 to 0.42, a compensation coefficient of 1.32 is calculated and applied to signal recovery processing, effectively eliminating signal distortion caused by polarization instability, achieving a signal recovery rate of 94.7%.

[0101] S250. The time-domain features, frequency-domain features and scattering intensity distribution in the signal correlation matrix are fused in a hierarchical manner to generate an initial signal feature matrix.

[0102] A hierarchical feature fusion method was performed, fusing the time-domain features, frequency-domain features, and scattering intensity distribution from the signal correlation matrix. A weighted averaging method was used, with a weight of 0.4 for time-domain features, 0.35 for frequency-domain features, and 0.25 for scattering intensity distribution features. The fusion generated an initial signal feature matrix with dimensions equal to the number of detection channels multiplied by the number of integrated features. Fifteen integrated features were extracted from each detection channel. Experimental results show that the feature matrix constructed using this method effectively characterizes the time-frequency characteristics and spatial correlation of photoelectric signals from multiple detection channels, achieving a feature discrimination ratio of 92.3%, providing a reliable foundation for subsequent signal processing and feature recognition.

[0103] In this embodiment, a correlation feature model of multi-detection channel photoelectric signals is established, which can effectively extract scattering intensity and polarization state features. Correlation features between detection channels are established through time-domain decomposition and frequency-domain transformation, and finally a highly discriminative initial signal feature matrix is ​​constructed, providing a new technical solution for photoelectric signal processing in complex environments.

[0104] Example 3

[0105] Based on Example 1 or Example 2, the present invention also provides a method for identifying and counting ultrafine particles based on the photoelectric effect, which also includes steps S100 to S700, and will not be described in detail here.

[0106] In this embodiment, step S300 specifically includes:

[0107] S310. Perform multi-scale tensor decomposition on the initial signal feature matrix to obtain the spatial gradient tensor of scattering intensity and the time evolution tensor of polarization state, and construct an adaptive threshold boundary accordingly.

[0108] The matrix contains the signal data collected by a multi-detection-channel photodetector. When performing multi-scale tensor decomposition on this initial signal feature matrix, the original matrix is divided into multiple sub-matrices through the sliding window technique. The window sizes are set to 16×16, 32×32, and 64×64 pixels respectively, and the window sliding step is 1 / 4 of the original window size. Apply high-order singular value decomposition to each sub-matrix to extract the scattering intensity spatial gradient tensor G and the polarization state time evolution tensor P. In actual tests, taking the data collected by a 4-detection-channel 16-bit quantization photodetector as an example, process the original data of 1024×1024 pixels. The dimension of the obtained tensor G is 4×1024×1024×3, representing the scattering intensity gradient of 4 detection channels in three-dimensional space; the dimension of the tensor P is 4×1024×1024×60, representing the change of the polarization state of 4 detection channels at 60 time sampling points.

[0109] Determine the coupling relationship between the two by calculating the cross-correlation coefficient matrix of tensor G and tensor P. When the cross-correlation coefficient in a certain area is greater than 0.75, it indicates that the scattering intensity change and the polarization state change are highly correlated, and a higher threshold is constructed at this time; when the cross-correlation coefficient is between 0.35 and 0.75, a medium threshold is set; when the cross-correlation coefficient is less than 0.35, a lower threshold is set. Based on this adaptive threshold mechanism, construct the threshold boundary surface B(x, y, t), where x and y are spatial coordinates and t is the time coordinate. In actual applications, for areas with strong signal intensity (signal-to-noise ratio SNR>20dB), the threshold is set to 0.85; for medium signal areas (10dB<SNR<20dB), the threshold is set to 0.65; for weak signal areas (SNR<10dB), the threshold is set to 0.45.

[0110] S320. Perform complex-domain trajectory projection of the signal components using the adaptive threshold boundary, screen out the effective signal components, and perform non-linear weighted fusion on the effective signal components to obtain a scale fusion signal;

[0111] It should be noted that the detection channels are distributed at different spatial positions. Therefore, scale fusion processing needs to be performed on the original signals of each detection channel to generate scale fusion signals. The scale fusion processing includes performing multi-scale decomposition of the original signal using Gaussian wavelets, extracting signal features at different scales, and fusing these features according to the weight coefficients. For example, 5 different scale parameter values can be selected: 1.0, 2.5, 4.0, 5.5, and 7.0, and the corresponding weight coefficients are 0.15, 0.25, 0.3, 0.2, and 0.1 respectively. In this way, the scale fusion signal can retain the key features in the original signal while suppressing noise interference.

[0112] In this embodiment, the complex-domain trajectory projection of the signal component is performed by using the constructed adaptive threshold boundary. The original signal is transformed into the complex domain through the Hilbert transform to form a complex trajectory. For the complex trajectory of each detection channel, its curvature feature K and closure feature C are calculated. The curvature feature K is represented by the average curvature of the arc formed by adjacent three points. In actual processing, the curvature is calculated every 10 sampling points and then averaged. The closure feature C is characterized by the ratio of the Euclidean distance between the starting point and the ending point of the trajectory to the total length of the trajectory. In actual tests, when K > 0.12 and C < 0.25, it is determined as a valid signal component; when 0.05 < K < 0.12 and C < 0.35, it is determined as a sub-valid signal component; in other cases, it is determined as an invalid signal component. A weight of 0.8 is assigned to the valid signal component, a weight of 0.2 is assigned to the sub-valid signal component, and a weight of 0 is assigned to the invalid signal component. The scale fusion signal S is obtained by weighted summation.

[0113] S330. Extract the wavefront propagation direction feature and the wavefront arrival time feature from the scale fusion signal, calculate the direction deviation and the time deviation between the detection channels, and construct a spatio-temporal cooperation index to determine the reference channel and the compensation channel to be compensated;

[0114] For the scale fusion signal S, wavefront feature extraction is performed, and the wavefront propagation direction feature D and the arrival time feature T are calculated. The direction feature D is determined by calculating the gradient direction of the signal energy in space and is represented by an angle; the arrival time feature T is determined by detecting the time when the signal energy exceeds three standard deviations of the background noise mean. For each detection channel pair (i, j), the direction deviation ΔDij = |Di - Dj| and the time deviation ΔTij = |Ti - Tj| are calculated. The spatio-temporal cooperation index STIij = w1·ΔDij + w2·ΔTij is constructed, where the weights w1 = 0.6 and w2 = 0.4. In a four-detection-channel system, the STI values of all detection channel pairs are calculated to form 6 STI values. The cooperation degree of the detection channel i is defined as the reciprocal of the average STI value of it and other detection channels. The detection channel with the highest cooperation degree is used as the reference channel R, and the detection channel with the lowest cooperation degree is used as the compensation channel C to be compensated.

[0115] S340. Extract the spatio-temporal evolution pattern of the scattering intensity and the polarization state rotation evolution pattern of the reference channel, construct the signal manifold topological structure, and map the compensation channel to be compensated to the signal manifold topological structure to obtain the manifold space reconstruction result;

[0116] The spatiotemporal evolution pattern M_I of scattering intensity and the polarization state rotation evolution pattern M_P are extracted from the reference channel R. M_I describes the variation pattern of signal intensity with space and time, and in actual processing, an 8×8×8 three-dimensional block is used to represent the local evolution pattern. M_P describes the rotation direction and rate of polarization state, and is characterized by the change of polarization ellipse parameters over time. Based on these two patterns, a signal manifold topology F is constructed to map the signal distribution in the multidimensional feature space onto the low-dimensional manifold. In practical applications, the isometric mapping algorithm (ISOMAP) is used to reduce the high-dimensional features to a 3D representation. The signal of the channel C to be compensated is mapped onto the manifold F, and the signal is reconstructed in the manifold space using the nearest neighbor interpolation method to obtain the manifold space reconstruction result R_C.

[0117] S350. Perform phase space alignment and amplitude normalization processing on the reconstructed manifold space result to generate an enhanced photoelectric signal.

[0118] Phase spatial alignment and amplitude normalization are performed on the reconstructed manifold result R_C. Phase spatial alignment is achieved by calculating the cross-correlation function between the reference channel R and the reconstructed result R_C, determining the phase shift corresponding to the position of maximum correlation, and then adjusting the phase of R_C accordingly. Amplitude normalization uses the Z-score method to standardize the signal amplitude to a distribution with a mean of 0 and a variance of 1. In practical applications, the enhanced signal detection probability is increased by 27.4%, the false detection rate is reduced by 18.6%, and the signal-to-noise ratio is improved by an average of 6.5 dB. The final enhanced photoelectric signal retains the physical characteristics of the original signal while significantly improving signal quality and detection performance.

[0119] Example 4

[0120] Based on Example 3, step S330 of the photoelectric effect-based ultrafine particle identification and counting method provided by the present invention specifically includes:

[0121] S331. Perform spatial domain gradient field analysis on the scale-fused signal to extract wavefront propagation direction features;

[0122] Spatial domain gradient field analysis is performed on the scale-fused signal to extract wavefront propagation direction features. Specifically, the rising edge of the signal in each detection channel is sampled, with sampling points acquired at 0.1 ms intervals within a 20 ms time window. The amplitude difference between adjacent sampling points is calculated to construct a phase gradient vector. For example, if the signal amplitude of detection channel A at t=10.5 ms is 0.45 and the signal amplitude at t=10.6 ms is 0.52, then the gradient value of this sampling point is 0.07. Spatial interpolation is performed on the gradient vectors of all detection channels to construct a wavefront propagation spatial surface. The wavefront propagation direction feature is defined as the normal vector direction of this surface at each detection channel position. For example, the wavefront propagation direction vector at detection channel A is [0.707, 0.707, 0], indicating that the wavefront propagates at a 45-degree angle in the xy plane.

[0123] S332. Perform time-domain peak localization on the scale-fused signal to obtain wavefront arrival time characteristics;

[0124] Temporal peak localization is performed on the scale-fused signal to obtain wavefront arrival time characteristics. The signal is resampled using cubic spline interpolation, increasing the sampling rate to 10 times the original to improve peak localization accuracy. In the resampled signal, the point with the maximum amplitude is identified as the peak point, and its corresponding time is recorded. For example, the peak of detection channel A occurs at t=12.34ms, and the peak of detection channel B occurs at t=12.57ms. The difference in peak arrival times between adjacent detection channels is calculated, and a spatiotemporal correlation mapping is constructed based on the spatial relationship between the detection channels. For example, if the spatial distance between detection channels A and B is 2.3mm and the time difference is 0.23ms, the wavefront propagation velocity can be estimated to be 10.0mm / ms.

[0125] S333. For each of the detection channels, select the neighboring detection channels within a preset neighborhood range, and calculate the deviation index of the wavefront propagation characteristics based on the wavefront arrival time characteristics.

[0126] For each detection channel, the consistency of wavefront propagation characteristics is calculated among neighboring detection channels within a preset neighborhood. The neighborhood can be defined as a circular area with a radius of 5 mm centered on the current detection channel. The angle between the wavefront propagation directions of the current detection channel and its neighboring channels is calculated, resulting in a set of angle values. For example, the angles between the wavefront propagation directions of detection channel A and its four neighboring detection channels are 5°, 7°, 12°, and 9°, respectively. The standard deviation of this set of angle values ​​is calculated as an indicator of wavefront propagation direction deviation. In the example above, the standard deviation is 2.94°. The wavefront propagation velocity is calculated based on the spatial distance between detection channels and the peak arrival time difference. For example, the wavefront propagation velocities of detection channel A and its four neighboring detection channels are 9.8 mm / ms, 10.2 mm / ms, 9.5 mm / ms, and 10.5 mm / ms, respectively. The standard deviation of this set of velocity values ​​is calculated as an indicator of wavefront arrival time deviation. In the example above, the standard deviation is 0.42 mm / ms.

[0127] S334. Weighted fusion of the wavefront propagation direction deviation and the wavefront arrival time deviation is performed to construct a space-time coordination index.

[0128] A spatiotemporal coordination index is constructed by weighting and fusing the wavefront propagation direction deviation and the wavefront arrival time deviation. The weighting coefficients can be adjusted according to the actual application scenario; for example, the weight of the direction deviation is 0.6, and the weight of the time deviation is 0.4. The direction deviation and the time deviation are normalized to the interval [0, 1], and then the weighted sum is calculated according to the weights. For example, if the normalized direction deviation is 0.15 and the time deviation is 0.21, then the spatiotemporal coordination index is 0.15 × 0.6 + 0.21 × 0.4 = 0.174.

[0129] S335. The detection channels whose spatiotemporal coordination index is higher than the preset coordination verification threshold are determined as reference channels, and the remaining detection channels are determined as channels to be compensated.

[0130] A preset co-verification threshold of 0.25 is set. Detection channels with a spatiotemporal co-verification index below the threshold are identified as reference channels, indicating that these detection channels have high consistency with the wavefront propagation characteristics of neighboring detection channels. Detection channels with a spatiotemporal co-verification index above the threshold are identified as channels to be compensated, indicating that the signal quality of these detection channels is poor or interfered with, requiring compensation processing.

[0131] For example, assuming 16 detection channels are arranged in a 4×4 grid, the spatiotemporal coordination index of each detection channel is calculated using the method described above. The calculation results show that the index values ​​of 12 detection channels are between 0.05 and 0.22, and these are identified as reference channels; the index values ​​of 4 detection channels are between 0.31 and 0.48, and these are identified as channels to be compensated. The signals of the channels to be compensated can be corrected using the signal characteristics of the reference channels, improving the reliability and accuracy of the overall signal processing system.

[0132] The spatiotemporal collaborative feature extraction method described above can effectively identify detection channels with poor signal quality, providing a basis for subsequent signal processing and system optimization, and has important value in practical applications.

[0133] Example 5

[0134] Based on any of the above embodiments, the present invention also provides a method for identifying and counting ultrafine particles based on the photoelectric effect, which also includes steps S100 to S700, and will not be described in detail here.

[0135] like Figure 2 In the embodiment shown, step S400 includes:

[0136] S410. Generate a polarization coherence angle range based on the polarization component of the detection channel in the enhanced photoelectric signal;

[0137] To determine the polarization coherence angle range, after inputting the enhanced photoelectric signal, it is necessary to calculate the phase difference between the polarization components acquired by each detection channel. This calculation involves selecting signals from adjacent detection channels and calculating the short-time phase difference sequence using a sliding window method. For example, in a system with eight detection channels, each channel corresponds to detection angles of 0°, 22.5°, 45°, 67.5°, 90°, 112.5°, 135°, and 157.5°, respectively. In this way, the phase difference between adjacent detection channels can be calculated.

[0138] By analyzing the phase difference variation amplitude, stable detection channel combinations can be identified. In practice, it can be determined whether the phase difference variation amplitude is less than a preset amplitude threshold. If not, the detection channel group is discarded; if so, a qualified detection channel group is identified and the corresponding detection angle range is extracted. The detection angle range corresponding to the qualified detection channel is the polarization coherence angle interval.

[0139] Practice shows that setting the phase difference variation threshold to 5° can effectively identify stable coherent detection channel groups. For example, when the phase difference variations between the three detection channels at 45°, 67.5°, and 90° are all less than 5°, the angular range of 45° to 90° can be defined as the polarization coherence angle interval. In a certain test sample, the polarization coherence angle interval was determined to be 62° to 108°, and the signal within this interval exhibited obvious phase stability characteristics.

[0140] S420. Extract the intensity of linear polarization component and circular polarization component within the polarization coherence angle interval, and perform time-by-time ratio calculation to generate a polarization energy ratio time series.

[0141] Within a defined polarization coherence angle range, the intensities of linear polarization and circular polarization components are extracted. Linear polarization components are obtained by selecting signals from a specific detection channel within the coherence angle range, while circular polarization components are obtained by adding a quarter-wave plate to the detection channel. For a specific case, within the coherence angle range of 62° to 108°, signals from detection angles of 70° and 100° can be used as linear polarization intensity references, and the corresponding circular polarization detection channel signals can be used as circular polarization intensity references.

[0142] The ratios of the extracted linear and circular polarization component intensities are calculated time-by-time to generate a polarization energy ratio time series. For example, in a set of test data lasting 10 seconds, the ratios are calculated at 10-millisecond intervals, resulting in a time series of 1000 data points. In actual testing, the polarization energy ratio sequence of a certain sample showed a significant change from 0.82 to 0.41 at 2.35 seconds, indicating a polarization state transition.

[0143] S430. Construct a two-dimensional polarization energy matrix based on the polarization energy ratio time series, identify the moment when the polarization state undergoes a jump, and determine the moment when the polarization energy conversion reverses.

[0144] A two-dimensional polarization energy matrix was constructed based on the time series of polarization energy ratios. Rows represent different time points, columns represent different polarization angles, and matrix elements represent the polarization energy at the corresponding time and angle. By analyzing the changes in energy distribution within the matrix, the moments when polarization state transitions occur can be identified. In practical applications, by setting the polarization energy change rate threshold to 0.15 / second, polarization state transitions were successfully detected at 2.35 seconds, 4.68 seconds, and 7.22 seconds.

[0145] S440. Within a preset time window before and after the polarization energy conversion reversal moment, extract the difference in the rate of change of the polarization energy ratio to generate a polarization degree jump amplitude sequence.

[0146] After determining the polarization energy conversion reversal moment, the difference in the rate of change of the polarization energy ratio within a 200-millisecond time window before and after this moment is extracted to generate a polarization degree jump amplitude sequence. Taking a certain sample as an example, at the polarization energy conversion reversal moment at 2.35 seconds, the rate of change in the first 200 milliseconds is ~0.08 / second, and the rate of change in the next 200 milliseconds is 0.22 / second, with a calculated difference in rate of change of 0.30 / second. Similarly, the differences in rate of change at 4.68 seconds and 7.22 seconds are 0.28 / second and 0.33 / second, respectively. These values ​​together constitute the polarization degree jump amplitude sequence, reflecting the evolution law of polarization degree.

[0147] S450. At the moment of polarization energy conversion reversal, calculate the coordinate values ​​of the polarization component in the Stokes parametric space and generate the polarization state displacement vector.

[0148] At the moment of polarization energy conversion reversal, the coordinate values ​​of the polarization components in Stokes parametric space are calculated to generate the polarization state displacement vector. In Stokes parametric space, the three coordinate axes represent horizontal / vertical linear polarization, 45° / 135° linear polarization, and right-handed / left-handed circular polarization, respectively. For the three conversion reversal moments in the above sample, the polarization state displacement vectors are (0.42, 0.18, 0.31), (0.38, 0.22, 0.29), and (0.45, 0.15, 0.33), respectively.

[0149] S460. Perform spatial orientation clustering on the polarization state displacement vector, identify valid orientation clusters, and generate polarization state transition features;

[0150] Spatial orientation clustering of polarization state displacement vectors can be performed using an improved K-means clustering algorithm. The magnitude threshold is set to 0.3, and the orientation dispersion threshold is set to 0.15. In the above samples, the average magnitude of the three polarization state displacement vectors is 0.41, and the orientation dispersion is 0.12, meeting the threshold requirements. Therefore, they can be identified as a valid orientation cluster, generating polarization state transition features.

[0151] S470. Combine the polarization degree jump amplitude sequence, the polarization state transition feature, and the polarization coherence angle interval to construct a three-dimensional polarization feature fingerprint.

[0152] A three-dimensional polarization fingerprint is constructed by combining the polarization degree jump amplitude sequence, polarization state transition features, and polarization coherence angle interval. For the above sample, the three-dimensional polarization fingerprint can be represented as: polarization coherence angle interval [62°, 108°], polarization degree jump amplitude sequence [0.30 / s, 0.28 / s, 0.33 / s], and polarization state transition feature vector (0.42, 0.18, 0.31). This three-dimensional polarization fingerprint has uniqueness and stability, and can be applied to scenarios such as target recognition and feature extraction.

[0153] Example 6

[0154] Based on the above embodiment 5, the present invention also provides a method for identifying and counting ultrafine particles based on the photoelectric effect, which also includes steps S100 to S700, and will not be described in detail here.

[0155] In this embodiment, step S430 includes:

[0156] S431. Construct a two-dimensional polarization energy matrix by dividing the polarization energy ratio time series into linear polarization components and circular polarization components. Calculate the polarization state transition intensity and polarization state transition direction based on the two-dimensional polarization energy matrix to generate a polarization state transition feature sequence.

[0157] A two-dimensional polarization energy matrix was constructed by analyzing the polarization energy ratio time series based on linear and circular polarization components. The polarization energy ratio time series was acquired from a multi-channel photoelectric detection array with a sampling rate of 100 kHz and a time series length typically ranging from 512 to 1024 points. The linear polarization component was calculated by dividing the difference between the horizontal and vertical polarization intensities by the total intensity, with values ​​ranging from ~1 to 1, where ~1 represents pure vertical polarization and 1 represents pure horizontal polarization. The circular polarization component was calculated by dividing the difference between the left-handed and right-handed polarization intensities by the total intensity, also ranging from ~1 to 1, where ~1 represents pure left-handed polarization and 1 represents pure right-handed polarization. For typical ultrafine particles, the linear polarization component varies in the range of 0.3 to 0.7, and the circular polarization component varies in the range of ~0.4 to 0.4. The constructed two-dimensional polarization energy matrix contains time points and the corresponding two polarization components, with a matrix size of twice the number of time points. The polarization state transition intensity and direction were calculated based on this matrix. The polarization state transition intensity is defined as the Euclidean distance between polarization energy vectors at adjacent time points, typically ranging from 0.01 to 0.15. The polarization state transition direction is defined as the azimuth angle of the change in the polarization energy vector, ranging from 0 to 360 degrees. By calculating the transition intensity and direction at each time point, a polarization state transition characteristic sequence is generated, containing both transition intensity and direction components.

[0158] S432. Perform phase unwrapping processing on the polarization state transition feature sequence, extract the phase jump point, and determine the time corresponding to the phase jump point as the first candidate reversal time point;

[0159] Phase unwrapping is performed on the polarization state transition feature sequence using the cumulative phase difference method. The window length for phase unwrapping is set to 32 points, and the threshold is set to 180 degrees. When the phase difference between adjacent time points exceeds the threshold, it is determined to be a phase jump. The phase value after phase unwrapping is the cumulative phase, i.e., the continuous phase after adding or subtracting an integer multiple of 360 degrees. In regions of drastic polarization state changes, the phase change usually exceeds 360 degrees, requiring 3 to 5 wrapping corrections. Through phase unwrapping, phase jump points are extracted, i.e., points where the phase change rate exceeds a preset threshold. The threshold is usually set to 3 times the average phase change rate; for ultrafine particles, this threshold is approximately 45 degrees per sampling point. The time corresponding to the phase jump point is determined as the candidate reversal time point. In practical applications, 5 to 15 candidate reversal time points are usually identified, representing moments when the polarization state may undergo significant changes.

[0160] S433. Calculate the covariance matrix of polarization state transition intensity and polarization state transition direction at the candidate reversal time point, determine the stability of polarization state transition based on the eigenvalues ​​of the covariance matrix, and select a second candidate reversal time point from the first candidate reversal time points.

[0161] The covariance matrix of polarization state transition intensity and direction at candidate reversal time points is calculated. The covariance matrix is ​​calculated using a sliding window method, with a window length of 17 points (8 sampling points before and after the candidate point). The covariance matrix is ​​2×2 in size and describes the relationship between the transition intensity and direction in a local region. The stability of the polarization state transition is determined based on the eigenvalues ​​of the covariance matrix. The two eigenvalues ​​represent the intensity of the primary and secondary directions of change. The eigenvalue ratio is defined as the larger eigenvalue divided by the smaller eigenvalue. A larger ratio indicates a more stable polarization state transition, meaning the change is mainly concentrated in one direction. A preset stability threshold is typically set to 5, requiring the intensity of the primary direction of change to be at least 5 times that of the secondary direction. For ultrafine particles with diameters between 0.5 and 2.5 μm, a stable polarization state transition eigenvalue ratio is typically in the range of 8 to 25. After stability screening, 2–5 candidate reversal time points are usually remaining, and candidate reversal time points with eigenvalue ratios less than the preset stability threshold are discarded.

[0162] S434. Calculate the rate of change of the scattering cross section at the second candidate reversal time point, and determine the candidate reversal time point where the rate of change of the scattering cross section is greater than a preset rate of change threshold as the polarization energy conversion reversal time point.

[0163] Calculate the rate of change of the scattering cross section at the remaining candidate reversal time points. The scattering cross section is proportional to the total scattering intensity of the particles and is obtained by integrating the omnidirectional scattered light intensity. The rate of change of the scattering cross section is calculated as the slope of the scattering cross section at five sampling points before and after the candidate point, i.e., the change in the scattering cross section per unit time. For ultrafine particles, the scattering cross section changes significantly when their shape or orientation changes, with a rate of change reaching 20-50% / μs. The preset rate of change threshold is usually set at 15% / μs of the average scattering cross section value. Candidate reversal time points with a scattering cross section rate of change greater than the preset rate of change threshold are identified as polarization energy conversion reversal times. In practical applications, 1-3 polarization energy conversion reversal times are usually ultimately determined. These times correspond to key moments when the particle scattering characteristics change significantly and are important indicators for identifying particle characteristics.

[0164] Experimental verification shows that when detecting standard spherical particles with a diameter of 1.2 μm, the deviation between the polarization energy conversion reversal time identified by this method and the theoretical prediction is less than 2 μs, achieving an accuracy of over 95%. For ellipsoidal particles with a shape ratio of 3:1, two polarization energy conversion reversal times can be accurately identified, corresponding to the moments when the particle's major and minor axes pass through the detection area, with a time interval of approximately 10–15 μs.

[0165] In this embodiment, by accurately identifying the moment of polarization energy conversion reversal, key technical support is provided for the precise counting and characteristic analysis of ultrafine particles, which can effectively improve the accuracy of particle size, shape and structure identification.

[0166] Example 7

[0167] Based on any of the above embodiments, the present invention also provides a method for identifying and counting ultrafine particles based on the photoelectric effect, which also includes steps S100 to S700, and will not be described in detail here.

[0168] In this embodiment, step S500 includes:

[0169] S510. Divide the enhanced photoelectric signal into multiple sub-signal segments; calculate the variance of the time interval sequence of signal zero-crossing points in each sub-signal segment, and determine the oscillation complexity of the sub-signal segment accordingly.

[0170] When calculating the product of the amplitude change rate and the second-order rate of change of adjacent sampling points in an enhanced photoelectric signal, a differential method is used to process the original sampled data. Enhanced photoelectric signals typically have a sampling rate of 100kHz, a quantization precision of 16 bits, and a signal dynamic range of 0–5V. The amplitude change rate is defined as the difference in amplitude between two adjacent sampling points divided by the sampling time interval, with units of V / μs. For ultrafine particles, this typically varies within the range of ±0.5V / μs. The second-order rate of change is defined as the difference in amplitude between two adjacent rates of change divided by the sampling time interval, with units of V / μs.2 The typical range is ±0.2V / μs. 2 The absolute value of the product of the two values ​​is taken. When this value exceeds a preset smoothness threshold, the corresponding time point is marked as the segmentation anchor point. The smoothness threshold is dynamically adjusted according to the signal characteristics and is usually set to 0.05 times the standard deviation of the signal amplitude. For ultrafine particles with a diameter of 0.5–2.5 μm, this threshold is approximately 0.005–0.02V. 2 / μs 3 In practical applications, a typical particle event generates 3 to 7 segmentation anchor points, corresponding to rapidly changing regions of the signal, such as rising edges, falling edges, and regions of significant fluctuation. Based on these anchor points, the enhanced photoelectric signal is segmented to obtain multiple sub-signal segments. Each sub-signal segment contains 25 to 150 sampling points, with a time span of approximately 0.25 to 1.5 ms.

[0171] When calculating the variance of the time interval sequence of zero-crossing points within each sub-signal segment, the sub-signal segment must first undergo mean-reduction processing to ensure the signal fluctuates around the zero point. A zero-crossing point is defined as the moment the signal changes from negative to positive or vice versa. The zero-crossing moment is precisely determined using linear interpolation, achieving a time resolution better than the sampling interval, reaching 0.2 μs. Each sub-signal segment typically contains 5–30 zero-crossing points, forming a zero-crossing time interval sequence. The variance of this sequence is calculated; a larger variance indicates stronger irregularity in signal oscillation, i.e., higher oscillation complexity. For regularly oscillating signals, the variance is typically less than 10% of the average of the total intervals; for complex oscillating signals, the variance can reach 30–80% of the average. In actual tests, the oscillation complexity of sub-signal segments generated by standard spherical particles with a diameter of 1.0 μm is typically 15–25%, while the oscillation complexity generated by non-spherical or aggregated particles can reach 40–65%.

[0172] S520. Determine the frequency sampling density based on the oscillation complexity, and perform frequency domain decomposition on the sub-signal segment accordingly to determine the frequency domain energy distribution;

[0173] A dynamic adjustment strategy is employed when determining the frequency sampling density based on oscillation complexity. Oscillation complexity is positively correlated with the richness of the signal spectrum; signals with high complexity require higher frequency resolution to accurately characterize their properties. Frequency sampling density is defined as the number of sampling points per kilohertz frequency range, with a baseline of 10 points / kHz. When the oscillation complexity is below 20%, the baseline frequency sampling density is used; when the complexity is between 20% and 40%, the frequency sampling density is increased to 1.5 times the baseline value; and when the complexity exceeds 40%, the frequency sampling density is increased to twice the baseline value. Frequency domain decomposition is performed on sub-signal segments at the specified frequency sampling density using a short-time Fourier transform method with a Hanning window and a 50% overlap rate. For a sub-signal segment of 100 points, a 128-point Fourier transform is used, achieving a frequency resolution of 0.8 kHz. The analysis frequency range is 0–50 kHz, covering the main frequency components of ultrafine particle scattering signals. The peak frequency and energy percentage of frequency components whose energy contribution exceeds a preset contribution threshold are extracted; this threshold is typically set to 5% of the total energy. For ultrafine particles with diameters of 0.5–2.5 μm, 2–6 main frequency components are usually extracted, with peak frequencies distributed in the range of 3–25 kHz and energy percentages between 5% and 40%.

[0174] S530. In the frequency domain energy distribution of the sub-signal segment, the peak frequency is time-tracked, and the peak frequency with a frequency drift amplitude less than a preset drift threshold is identified as a stable oscillation frequency. The energy proportion corresponding to the stable oscillation frequency is extracted as a time-varying sequence to determine the energy modulation evolution trajectory. The stable oscillation frequency and the energy modulation evolution trajectory are combined to form an oscillation feature spectrum.

[0175] When performing time-series tracking of peak frequencies in the frequency domain energy distribution of multiple sub-signal segments, a nearest neighbor matching algorithm is employed. For peak frequencies in adjacent sub-signal segments, the frequency difference between them is calculated. If the difference is less than a preset drift threshold, they are considered to belong to the same stable oscillation frequency. The preset drift threshold is set according to signal characteristics, typically 8% of the average peak frequency. For the main frequency component around 15kHz, the drift threshold is approximately 1.2kHz. For particle passage events lasting longer than 100μs, 1–3 stable oscillation frequencies can usually be identified. The dominant frequency is usually inversely proportional to the particle size; the dominant frequency of a standard spherical particle with a diameter of 1.0μm is approximately 12kHz, while the dominant frequency of a particle with a diameter of 0.5μm can reach 22kHz. The energy proportion corresponding to the stable oscillation frequency is extracted as a sequence of changes over time, forming an energy modulation evolution trajectory. The energy modulation evolution trajectory reflects the time-varying characteristics of the oscillation intensity. For spherical particles, the energy proportion usually exhibits a single-peak symmetrical distribution; for non-spherical particles, the energy proportion may show a multi-peak or asymmetrical distribution. The number of trajectory sampling points is the same as the number of sub-signal segments, typically 3 to 7 points for a particle passage event. The stable oscillation frequency and the energy-modulated evolution trajectory are combined to form an oscillation characteristic spectrum, which serves as an important feature for particle identification.

[0176] In this embodiment, by accurately extracting the oscillation characteristics of particle photoelectric scattering signals and combining time-domain and frequency-domain analysis methods, high-precision identification and counting of ultrafine particles are achieved. This method is particularly suitable for the precise measurement of particles with a diameter of less than 2.5 μm in fields such as air quality monitoring and micro / nano particle analysis.

[0177] Example 8

[0178] Accordingly, according to embodiments of the present invention, the present invention also provides a computer device, a readable storage medium, and a computer program product.

[0179] Figure 3 This is a schematic diagram of the structure of a computer device 62 provided in an embodiment of the present invention. Figure 3 A block diagram of an exemplary computer device 62 suitable for implementing embodiments of the present invention is shown. Figure 3 The computer device 62 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of the present invention.

[0180] like Figure 3As shown, computer device 62 is represented in the form of a general-purpose computing device. Computer device 62 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0181] The components of computer device 62 may include, but are not limited to: one or more processors or processing units 66, system memory 78, and bus 68 connecting different system components (including system memory 78 and processing unit 66).

[0182] Bus 68 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Detection Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0183] Computer device 62 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by computer device 62, including volatile and non-volatile media, removable and non-removable media.

[0184] System memory 78 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 80 and / or cache memory 82. Computer device 62 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 84 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 3 Not shown; usually referred to as a "hard drive"). Although Figure 3Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 68 via one or more data media interfaces. System memory 78 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention.

[0185] A program / utility 90 having a set (at least one) of program modules 92 may be stored, for example, in system memory 78. Such program modules 92 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 92 typically perform the functions and / or methods described in the embodiments of the present invention.

[0186] Computer device 62 can also communicate with one or more external devices 64 (e.g., keyboard, pointing device, display 74, etc.), and with one or more devices that enable a user to interact with computer device 62, and / or with any device that enables computer device 62 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed via input / output (I / O) interface 72. Furthermore, computer device 62 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 70. As shown, network adapter 70 communicates with other modules of computer device 62 via bus 68. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with computer device 62, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0187] The processing unit 66 executes various functional applications and data processing by running programs stored in the system memory 78, such as implementing the photoelectric effect-based ultrafine particle identification and counting method provided in the embodiments of the present invention.

[0188] This invention also provides a non-transitory computer-readable storage medium storing computer instructions, on which a computer program is stored, wherein when the program is executed by a processor, it implements the photoelectric effect-based ultrafine particle identification and counting method provided in all embodiments of this invention.

[0189] The computer storage medium of this invention can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. More specific examples (a non-exhaustive list) of computer-readable storage media include: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0190] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0191] The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof. The computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0192] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described photoelectric effect-based ultrafine particle identification and counting method.

[0193] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0194] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for identifying and counting ultrafine particles based on the photoelectric effect, characterized in that, include: S100: Acquire the photoelectric signal generated by the particle stream to be tested passing through the photoelectric detection array, wherein the photoelectric detection array has multiple detection channels; S200. Extract the original features of the photoelectric signals in each of the detection channels, and construct an initial signal feature matrix by combining time-domain decomposition and frequency-domain transformation; S300. Perform multi-scale tensor decomposition and signal component screening on the initial signal feature matrix, and combine the detection channel collaborative compensation with wavefront propagation characteristics to generate an enhanced photoelectric signal. S400. Based on the enhanced photoelectric signal, determine the polarization coherence angle range, extract polarization energy features in the polarization coherence angle range, analyze polarization state conversion characteristics, and establish a three-dimensional polarization feature fingerprint. S500 performs adaptive segmentation processing on the enhanced photoelectric signal and extracts the oscillation feature spectrum; S600. Perform deep correlation matching between the oscillation feature spectrum and the three-dimensional polarization feature fingerprint to obtain a feature correlation matrix; S700. Construct a multidimensional classification criterion based on the feature correlation matrix, and perform feature separation on the particles to be tested accordingly to determine the final particle counting result.

2. The ultrafine particle identification and counting method based on the photoelectric effect according to claim 1, characterized in that, The original features include scattering intensity distribution features and polarization state conversion features; Step S200 includes: S210. Perform time-domain decomposition on the original features to obtain the signal envelope, and simultaneously perform frequency-domain transformation to obtain a frequency-domain feature set; S220. The signal envelopes of adjacent detection channels are correlated and matched by a spatial mapping function, and a light intensity modulation curve is generated by combining the frequency domain feature set. S230. Dynamically segment the light intensity modulation curve, extract the frequency modulation depth, phase modulation characteristics and polarization state transition law of each segment of the signal, and construct a modulation information sequence. S240. Construct a weight coefficient matrix based on the modulation information sequence, and establish a signal correlation matrix between the detection channels; S250. The time-domain features, frequency-domain features and scattering intensity distribution in the signal correlation matrix are fused in a hierarchical manner to generate an initial signal feature matrix.

3. The ultrafine particle identification and counting method based on the photoelectric effect according to claim 1, characterized in that, Step S300 includes: S310. Perform multi-scale tensor decomposition on the initial signal feature matrix to obtain the spatial gradient tensor of scattering intensity and the time evolution tensor of polarization state, and construct an adaptive threshold boundary accordingly. S320. The complex domain trajectory of the signal component is projected using the adaptive threshold boundary, and the effective signal components are selected. Nonlinear weighted fusion is performed on the effective signal components to obtain the scale fusion signal. S330. Extract wavefront propagation direction features and wavefront arrival time features from the scale fusion signal, calculate the direction deviation and time deviation between the detection channels, and construct a spatiotemporal coordination index to determine the reference channel and the channel to be compensated. S340. Extract the spatiotemporal evolution mode of scattering intensity and the polarization state rotation evolution mode of the reference channel, construct the signal manifold topology, map the channel to be compensated to the signal manifold topology, and obtain the manifold space reconstruction result. S350. Perform phase space alignment and amplitude normalization processing on the reconstructed manifold space result to generate an enhanced photoelectric signal.

4. The ultrafine particle identification and counting method based on the photoelectric effect according to claim 3, characterized in that, Step S330 includes: S331. Perform spatial domain gradient field analysis on the scale-fused signal to extract wavefront propagation direction features; S332. Perform time-domain peak localization on the scale-fused signal to obtain wavefront arrival time characteristics; S333. For each of the detection channels, select the neighboring detection channels within a preset neighborhood range, and calculate the deviation index of the wavefront propagation characteristics based on the wavefront arrival time characteristics. S334. Weighted fusion of the wavefront propagation direction deviation and the wavefront arrival time deviation is performed to construct a space-time coordination index. S335. The detection channels whose spatiotemporal coordination index is higher than the preset coordination verification threshold are determined as reference channels, and the remaining detection channels are determined as channels to be compensated.

5. The ultrafine particle identification and counting method based on the photoelectric effect according to claim 1, characterized in that, Step S400 includes: S410. Generate a polarization coherence angle range based on the polarization component of the detection channel in the enhanced photoelectric signal; S420. Extract the intensity of linear polarization component and circular polarization component within the polarization coherence angle interval, and perform time-by-time ratio calculation to generate a polarization energy ratio time series. S430. Construct a two-dimensional polarization energy matrix based on the polarization energy ratio time series, identify the moment when the polarization state undergoes a jump, and determine the moment when the polarization energy conversion reverses. S440. Within a preset time window before and after the polarization energy conversion reversal moment, extract the difference in the rate of change of the polarization energy ratio to generate a polarization degree jump amplitude sequence. S450. At the moment of polarization energy conversion reversal, calculate the coordinate values ​​of the polarization component in the Stokes parametric space and generate the polarization state displacement vector. S460. Perform spatial orientation clustering on the polarization state displacement vector, identify valid orientation clusters, and generate polarization state transition features; S470. Combine the polarization degree jump amplitude sequence, the polarization state transition feature, and the polarization coherence angle interval to construct a three-dimensional polarization feature fingerprint.

6. The ultrafine particle identification and counting method based on the photoelectric effect according to claim 5, characterized in that, Step S430 includes: S431. Construct a two-dimensional polarization energy matrix by dividing the polarization energy ratio time series into linear polarization components and circular polarization components. Calculate the polarization state transition intensity and polarization state transition direction based on the two-dimensional polarization energy matrix to generate a polarization state transition feature sequence. S432. Perform phase unwrapping processing on the polarization state transition feature sequence, extract the phase jump point, and determine the time corresponding to the phase jump point as the first candidate reversal time point; S433. Calculate the covariance matrix of polarization state transition intensity and polarization state transition direction at the candidate reversal time point, determine the stability of polarization state transition based on the eigenvalues ​​of the covariance matrix, and select a second candidate reversal time point from the first candidate reversal time points. S434. Calculate the rate of change of the scattering cross section at the second candidate reversal time point, and determine the candidate reversal time point where the rate of change of the scattering cross section is greater than a preset rate of change threshold as the polarization energy conversion reversal time point.

7. The ultrafine particle identification and counting method based on the photoelectric effect according to claim 1, characterized in that, Step S500 includes: S510. Divide the enhanced photoelectric signal into multiple sub-signal segments; calculate the variance of the time interval sequence of signal zero-crossing points in each sub-signal segment, and determine the oscillation complexity of the sub-signal segment accordingly. S520. Determine the frequency sampling density based on the oscillation complexity, and perform frequency domain decomposition on the sub-signal segment accordingly to determine the frequency domain energy distribution; S530. In the frequency domain energy distribution of the sub-signal segment, the peak frequency is time-tracked, and the peak frequency with a frequency drift amplitude less than a preset drift threshold is identified as a stable oscillation frequency. The energy proportion corresponding to the stable oscillation frequency is extracted as a time-varying sequence to determine the energy modulation evolution trajectory. The stable oscillation frequency and the energy modulation evolution trajectory are combined to form an oscillation feature spectrum.

8. The ultrafine particle identification and counting method based on the photoelectric effect according to claim 1, characterized in that, The multidimensional classification criteria in step S700 include velocity distribution, acceleration distribution, size features, and shape features; After performing feature separation on the particles to be tested, the method further includes: generating a feature identifier for each individual particle to be tested.

9. A computer device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the photoelectric effect-based ultrafine particle identification and counting method according to any one of claims 1 to 8.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to execute the photoelectric effect-based ultrafine particle identification and counting method according to any one of claims 1 to 8.

Images