Real-time automatic pollen concentration detection method and device

Abstract

The invention relates to the technical field of data processing, and particularly provides a pollen concentration real-time automatic detection method and device.The method comprises the following steps that monitoring data obtained through monitoring of an automatic pollen real-time monitoring device are obtained; based on the monitoring data, a pollen concentration detection result is obtained through a pollen concentration detection model obtained through pre-training; the pollen concentration detection model is obtained by training a pollen concentration real-time monitoring data sample set obtained by monitoring the pollen in the air through the automatic pollen real-time monitoring device and a pollen concentration standard sampling data sample set obtained by collecting the Hurst type trap in the same preset period and in the same territorial range. According to the detection method provided by the invention, the pollen concentration can be accurately and automatically detected in real time.

Description

Technical Field

[0001] The present invention belongs to the technical field of data processing, and in particular relates to a method and device for real-time automatic detection of pollen concentration. Background Art

[0002] Monitoring airborne pollen is crucial for assessing bioaerosols and allergic respiratory health. Hirst-type traps are internationally recognized as a reference method for pollen monitoring, providing volumetric concentration data (CEN / EN 16868:2019) that accurately reflects airborne pollen concentrations. However, Hirst-type trap sampling currently requires manual counting, making it difficult to automate data processing at high temporal resolution.

[0003] To address the need for more convenient, high-temporal-resolution pollen monitoring, several types of automated pollen monitoring instruments have been developed. These include optical counting methods, image recognition technology, and laser particle size measurement. These automated monitoring methods significantly reduce manual labor, improve monitoring efficiency, and enable real-time data collection. However, the monitoring accuracy of existing automated pollen monitoring instruments remains unverified, and their accuracy is significantly affected by environmental factors. Therefore, obtaining real-time monitoring data on pollen concentration while ensuring that this accuracy remains comparable to that of standard manual methods remains an urgent problem. Summary of the Invention

[0004] In view of this, an object of the present invention is to provide a method and device for real-time automatic detection of pollen concentration, which can provide real-time automatic detection of pollen concentration and improve detection accuracy.

[0005] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0006] According to a first aspect of the present invention, a method for real-time automatic detection of pollen concentration is provided, comprising the following steps:

[0007] Acquire monitoring data obtained by an automatic real-time pollen monitoring device;

[0008] Based on the monitoring data, the pollen concentration detection results are obtained using the pre-trained pollen concentration detection model;

[0009] Among them, the pollen concentration detection model is trained by using the pollen concentration real-time monitoring data sample set obtained by monitoring the pollen concentration in the air through an automatic real-time pollen monitoring device and the pollen concentration standard sampling data sample set collected by a Hurst-type trap within the same preset period and geographical range.

[0010] Furthermore, the training of the pollen concentration detection model includes:

[0011] Obtaining pollen concentration real-time monitoring data sample sets and pollen concentration standard sampling data sample sets respectively;

[0012] Construct the initial function of the linear regression model;

[0013] The sample data in the pollen concentration standard sampling sample data set is used as the dependent variable, and the independent variables include the corresponding sample data in the pollen concentration real-time monitoring sample data set. The initial function is iteratively trained to obtain the pollen concentration detection model.

[0014] Furthermore, the training of the pollen concentration detection model also includes:

[0015] After obtaining the pollen concentration real-time monitoring data sample set and the pollen concentration standard sampling data sample set, the data in the pollen concentration real-time monitoring data set and the pollen concentration standard sampling data set are preprocessed, and iterative training is performed using the data in the preprocessed data set.

[0016] Furthermore, the pre-processing includes:

[0017] The data with measurement time less than the preset time length in the pollen concentration real-time monitoring data set and the pollen concentration standard sampling data set are eliminated, and the data in the data set are arranged in natural time order.

[0018] Furthermore, the linear regression model is a simple linear regression model.

[0019] Furthermore, the linear regression model is a multiple linear regression model, and the independent variables also include: meteorological data and / or atmospheric particulate matter concentration data collected during the same period as pollen monitoring, meteorological data including one or more of temperature, relative humidity, precipitation, wind speed, and air pressure, and atmospheric particulate matter concentration data including PM 10 value.

[0020] Furthermore, the K-fold cross-validation method is used in the training process of the linear regression model. When the preset number of iterations is reached, the iteration is stopped, and a model with one or more of the coefficient of determination, root mean square error, and mean absolute error that meets the predetermined conditions is selected as the pollen concentration detection model.

[0021] Furthermore, obtaining the sample data in the pollen concentration standard sampling sample data set includes:

[0022] Pollen samples were collected using a Hoechst-type trap;

[0023] The pollen sample is stained with a toluidine blue solution to obtain a stained sample;

[0024] The stained samples were counted under a fluorescence microscope to obtain sample data of the pollen sample.

[0025] According to another aspect of the present invention, there is provided a device for automatically detecting pollen concentration in real time, comprising:

[0026] A pollen monitoring data acquisition unit is used to acquire real-time pollen concentration monitoring data obtained by the automatic real-time pollen monitoring device monitoring the pollen concentration in the air;

[0027] The data processing unit is used to process the monitoring data using the pre-trained pollen concentration detection model to obtain the pollen concentration detection result.

[0028] Furthermore, the pollen concentration detection model is a multiple linear regression model, and the real-time automatic detection device for pollen concentration further includes:

[0029] A meteorological data collection unit for collecting meteorological data collected in the same period and in the same area as the pollen automatic real-time monitoring device; and / or

[0030] Atmospheric particulate matter data collection unit, used to collect atmospheric particulate matter concentration data in the same period and in the same area as the pollen automatic real-time monitoring device;

[0031] The data processing unit processes the monitoring data, meteorological data and / or atmospheric particulate matter concentration data to obtain a pollen concentration detection result.

[0032] The above technical solution of the present invention has at least one of the following beneficial effects:

[0033] The present invention obtains monitoring data from an automatic real-time pollen monitoring device and, based on this data, utilizes a pre-trained pollen concentration detection model to obtain pollen concentration detection results. The pollen concentration detection model is trained using a set of real-time pollen concentration monitoring data samples obtained by the automatic real-time pollen monitoring device and a set of standard pollen concentration sampling data samples collected by a Hurst-type trap within the same preset time period and geographical range. The detection method of the present invention can improve the accuracy of real-time pollen concentration detection, achieving a detection result that approaches the accuracy of standard manual methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] To more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present invention, and those skilled in the art can derive other drawings based on these drawings without inventive effort.

[0035] Figure 1This is a flow chart of a method for real-time automatic detection of pollen concentration provided by one embodiment of the present invention;

[0036] Figure 2 This is a Passing-Bablok regression analysis graph of pollen counts using a fluorescent staining method provided by one embodiment of the present invention and a magenta staining method used for comparison;

[0037] Figure 3 This is a graph showing daily pollen concentration changes in a 2023 automatic real-time pollen monitoring device and a Hurst-type trap, provided as an example of an embodiment of the present invention;

[0038] Figure 4 This is a graph showing daily pollen concentration changes in a 2024 automatic real-time pollen monitoring device and a Hurst-type trap, provided as an example of an embodiment of the present invention;

[0039] Figure 5 This is an example diagram of a simple linear regression model provided by one embodiment of the present invention;

[0040] Figure 6 is a scatter plot of predicted values and actual values obtained by a multiple linear regression model provided by one embodiment of the present invention;

[0041] Figure 7 This is a schematic structural diagram of a real-time automatic detection device for pollen concentration provided by the present invention;

[0042] Figure 8 A block diagram of an electronic device provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0043] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of the present invention.

[0044] It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the numbers used in this manner are interchangeable where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or device comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0045] The embodiments of the present invention provide a method and device for real-time automatic detection of pollen concentration, which can realize real-time automatic detection of pollen concentration with high accuracy. By using the pollen concentration obtained by the real-time monitoring device and the standard manual pollen count collected by the Hurst-type trap as the independent variable and dependent variable, respectively, a pollen concentration detection model is pre-trained. The monitoring data obtained by the automatic real-time pollen monitoring device is input into the model to obtain a pollen concentration detection result with an accuracy close to that of the standard manual method. This real-time automatic detection method for pollen concentration will help provide clinicians and the public with timely airborne pollen information, promote the establishment of an early warning system, enable susceptible patients or populations to take preventive measures in advance, and optimize the allocation of medical resources to meet the emergency medication needs that may arise during the peak period of pollen allergy.

[0046] In order to make the purpose, technical solutions and advantages disclosed in the embodiments of the present invention more clearly understood, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present invention and are not intended to limit the embodiments of the present invention.

[0047] Reference Manual Figure 1 , which shows the process of a real-time automatic detection method for pollen concentration provided by an embodiment of the present invention. The method may include the following steps:

[0048] S101: Acquire monitoring data obtained by the automatic real-time pollen monitoring device.

[0049] Specifically, the pollen automatic real-time monitoring device of this embodiment can use a particle counter (such as the KH-3000 pollen automatic monitor). The particle counter uses a light beam to illuminate the particles in the air and detects the intensity of the scattered light to calculate the pollen concentration. The particle counter can calculate the pollen concentration per cubic meter of air per minute (particles / m 3 ), calculated by dividing the total number of pollen particles detected in a day by the total volume inhaled (m 3), thus obtaining the daily pollen concentration per cubic meter of air (particles / m 3 ).

[0050] Existing particle counters can measure particles ranging from 0.5–300 μm, including pollen, spores, and fine particulate matter. However, in extreme environmental conditions such as sandstorms and haze, atmospheric particulate matter can affect sensor detection accuracy. Furthermore, meteorological factors such as precipitation can cause pollen fragmentation, interfering with sensor pollen detection and leading to false alarms. Therefore, the monitoring data needs to be calibrated.

[0051] S102: Based on the monitoring data, a pollen concentration detection model obtained in advance is used to obtain a pollen concentration detection result.

[0052] Among them, the pollen concentration detection model is trained by using the pollen concentration real-time monitoring data sample set obtained by monitoring the pollen concentration in the air through an automatic real-time pollen monitoring device and the pollen concentration standard sampling data sample set collected by a Hurst-type trap within the same preset period and geographical range.

[0053] Pollen detection using an automatic real-time pollen monitoring device is highly timely, but the real-time monitoring data of pollen concentration obtained is insufficient in accuracy.

[0054] To reliably calibrate monitoring data, the inventors conducted extensive research, collecting a large amount of sample data from a Hirst-type trap using standard metrology and a large amount of sample data from a particle counter for training. This resulted in a detection model (also known as a model for calibrating monitoring data). This pollen concentration detection model, trained using sample data from both methods, produced highly accurate and timely pollen detection results.

[0055] Next, the training of the pollen concentration detection model is explained in detail.

[0056] The training of the pollen concentration detection model according to an embodiment of the present invention includes the following steps:

[0057] (1) Collecting samples to obtain a sample set

[0058] The pollen concentration real-time monitoring data sample set and the pollen concentration standard sampling data sample set are obtained respectively.

[0059] 1. Sampling of pollen concentration standard sampling data sample set

[0060] According to an embodiment of the present invention, pollen is collected using a Hirst-type trap in accordance with European standards. The Hirst-type trap is a classic device in the field of ambient air pollen monitoring, and its design principles and methods have been widely used for more than half a century. The Hirst-type pollen trap is a classic volumetric single-stage slit impact sampler, which is mainly used for continuous monitoring of pollen and spores in ambient air. Its design principle is based on aerodynamics. By extracting a fixed volume of air sample, the particulate matter in the air collides with and adheres to a mobile carrier, and then the pollen in the obtained sample is manually identified and counted, thereby obtaining standard sampling data.

[0061] The Hoechst trap collects a pollen sample strip once a week and cuts it into seven sections, corresponding to seven consecutive days of pollen sampling. To ensure proper operation, the entire instrument needs to be thoroughly cleaned at least once a month.

[0062] 2. Manual counting

[0063] According to some embodiments of this embodiment, the operating standards for manual pollen counting are as follows:

[0064] First, all pollen samples collected by Hoechst traps were stained with toluidine blue and counted under a fluorescence microscope.

[0065] The specific counting operation, according to the European standard (EN16868:2019), is to count pollen grains in a Hoechst trap sample at 400x magnification in four vertical lines, covering approximately 15.8% of the total area of the slide, which exceeds the standard requirement of at least 10%. Two of the lines are located in the upper half of the slide, and the other two are located in the lower half, all symmetrically distributed about the center line. The daily pollen count is then multiplied by the sampling surface area for the entire day.

[0066] Here, it should be noted that the inventors have found that the staining and counting of pollen concentration standard sampling samples (i.e., Hoechst-type trap marking) can obtain more accurate values by staining with toluidine blue solution and counting under a fluorescence microscope.

[0067] Specifically, in some embodiments of the present invention, obtaining sample data from a pollen concentration standard sampling sample data set includes:

[0068] 1) Collect pollen samples using a Hoechst trap;

[0069] 2) staining the pollen sample with a toluidine blue solution to obtain a stained sample;

[0070] 3) Count the stained sample under a fluorescence microscope to obtain sample data of the pollen sample.

[0071] In an embodiment of the present invention, toluidine blue solution is used to dye pollen. The toluidine blue dye solution can be prepared by diluting 0.5 g of powdered toluidine blue to 100 ml with distilled water. The pollen dyed with toluidine blue appears green or yellow-green under blue-violet light of a fluorescence microscope.

[0072] In contrast, the traditional pollen sample counting method is fuchsin staining. The same sample band counted by the two staining methods is compared, and the results are as follows: Figure 2 As shown. By comparing the two staining methods, a strong positive linear relationship was observed between them (Pearson correlation coefficient was 0.99). Paired t-test analysis showed no statistically significant difference in the number of pollen grains measured by the two methods (P = 0.25). The calculated root mean square error (RMSE) was 17.21. After using Passing-Bablok regression analysis, a quantitative relationship model between the two staining methods was obtained:

[0073] N 品红染色法测得的花粉粒数(参考方法) =5.50+1.00×N 荧光染色法测得的花粉粒数

[0074] The slope of the regression equation is 1.00, close to the ideal theoretical identity line (dashed orange line in the figure), indicating good agreement between the two methods, almost perfect agreement. In the Bland-Altman plot, the 95% confidence interval of the regression line is narrow (grey shaded area in the figure), indicating minimal difference in the values measured by the two methods across the entire range of pollen grain numbers, demonstrating good agreement between the two methods.

[0075] Pollen identification and verification are typically based on the following: pollen shape (usually spherical, subspherical, or prolate), pollen size (polar range × equatorial range), pollen germinal organs (cracks, pores, grooves, and sulcus-like pores), and pollen wall thickness, presence, and ornamentation (e.g., granular, cerebellar, verrucous, or tumorous). For both staining methods, both light microscope images clearly reveal the basic structure, shape, germinal organs, and ornamentation of the pollen grains. However, the germinal organs of pollen grains captured by fluorescence microscopy are clearer than those captured by conventional light microscopy, allowing for more accurate identification and reducing errors.

[0076] 3. Sampling of pollen concentration real-time monitoring data sample set

[0077] An automatic real-time pollen monitoring device (such as the Yamatronics KH-3000) can generate pollen concentration data every minute. The monitoring time can be set to the entire day or to a period that includes a key monitoring period with higher pollen concentrations each day.

[0078] During monitoring, the standard inlet flow rate of the Yamatronics KH-3000 was set to 4.1 liters / minute. The monitoring equipment should be regularly checked to ensure proper operation and calibrated to maintain data accuracy. This application does not impose any restrictions on this. The following description is only an example of daily data expansion.

[0079] For example, the KH-3000 instrument, used as an automated real-time pollen monitoring device, was used to monitor pollen on the rooftop of the third floor of a building, approximately 12 meters above the ground. Simultaneously, pollen was collected using standard Hoechst-type traps placed on the same rooftop, 3 meters apart to minimize turbulent interference.

[0080] Reference Manual Figure 3 and 4 , the daily pollen concentration changes of the automatic real-time pollen monitoring device and the Hurst-type trap were recorded from April 4 to December 31, 2023 and from April 1 to November 30, 2024, respectively.

[0081] As can be seen from these two figures, although the values are different, the curves are similar and have a good correlation. The Pearson correlation coefficient and Spearman rank correlation coefficient between the data of the two sample sets were calculated, and the results showed that there was a significant correlation between the two methods.

[0082] In other words, the automatic real-time pollen monitoring device can be used to monitor pollen in real time. However, there is a certain deviation between the monitored value and the actual value (i.e., the value obtained by the Hurst-type trap annotation). To use it for automatic and accurate pollen detection, it should be calibrated accordingly. Precisely based on this, the inventors used the automatic real-time pollen monitoring device as the independent variable and the value obtained by the Hurst-type trap annotation as the dependent variable for training, thus obtaining a detection model (i.e., a model used to calibrate the automatic real-time pollen monitoring device).

[0083] In some embodiments of the present invention, after obtaining a pollen concentration real-time monitoring data sample set and a pollen concentration standard sampling data sample set, the data in the pollen concentration real-time monitoring data set and the pollen concentration standard sampling data set are preprocessed, and iterative training is performed using the data in the preprocessed data set.

[0084] Specifically, data preprocessing can include detecting outliers that significantly deviate from the standard deviation and correcting or deleting them. These outliers may be caused by accidental factors, such as extreme weather or collector failure. These data are inherently irregular, and removing them can improve the model's detection accuracy.

[0085] In some embodiments of the present invention, the specific preprocessing includes:

[0086] Data with measurement time less than the preset duration in the pollen concentration real-time monitoring dataset (sometimes also referred to as the KH-3000 dataset in the following examples) and the pollen concentration standard sampling dataset (sometimes also referred to as the Hirst-type trap dataset in the following examples) are removed, and the data in the dataset are arranged in natural time order.

[0087] For example, when preprocessing the dataset, data corresponding to days with less than 16 hours of daily monitoring time are removed. The processed data in the real-time pollen concentration monitoring dataset and the standard pollen concentration sampling dataset are sorted by day, and the two sets of data are matched and aligned according to date.

[0088] Specifically, the results of the preprocessing are shown in Table 1 below.

[0089] Table 1. Summary of data collection using Hoechst traps and automated real-time monitoring in 2023 and 2024

[0090]

[0091] In addition, this application also considers the impact of meteorological factors and specific substances in the air on the real-time monitoring data of pollen concentration. When obtaining pollen concentration monitoring data, meteorological data and / or atmospheric particulate matter concentration data are collected. Meteorological data may include one or more of temperature, relative humidity, precipitation, wind speed, and air pressure; atmospheric particulate matter concentration data may include PM 2.5 Value, PM 10 Value, etc.

[0092] The meteorological data and / or atmospheric particulate matter concentration data collected during the same period have the same time unit as the pollen concentration data, which can be daily averages. For example, the temperature can be the average temperature of the day (in degrees Celsius), and the humidity can be the daily relative humidity. In addition, the atmospheric particulate matter concentration data, except for PM 10 In addition to the value, PM 2.5 , due to the positive correlation between atmospheric particulate matter concentration data and pollen concentration calculated based on actual detection data, PM 10 The positive correlation is higher, so PM is preferred. 10 value.

[0093] The sampling location for meteorological data and atmospheric particulate matter concentration can be close to the sampling location of pollen concentration, or data from a meteorological monitoring station asymptotically close to the sampling location of pollen concentration can be directly used. The present invention does not impose any limitation on this.

[0094] (2) Constructing the initial function of the linear regression model

[0095] Combine Figure 3 and Figure 4 Assuming that there is a stable linear deviation between the sample data obtained by the two methods, linear regression can directly fit this relationship. The linear regression model can still be trained stably with small samples, can avoid overfitting, and has the advantage of fast training speed.

[0096] The sample data corresponding to the pollen concentration real-time monitoring sample data set is equivalent to the input features of the model, and the sample data in the pollen concentration standard sampling sample data set is equivalent to the true label. A linear regression model is used to establish a mapping model based on the corresponding relationship between the two.

[0097] 1. Simple Linear Regression Model

[0098] In some embodiments of the present invention, only the real-time monitoring data of pollen concentration is considered as the independent variable of the linear regression model, that is, the linear regression model is a simple linear regression model.

[0099] The initial function of the simple linear regression model was constructed as follows: Y = A + BX, where Y represents the pollen data collected using the standard Hurst trap method and X represents the real-time monitoring data of pollen concentration.

[0100] 2. Multiple Linear Regression Model

[0101] In other embodiments of the present invention, as independent variables of the linear regression model, in addition to the real-time monitoring data of pollen concentration, meteorological data and / or atmospheric particulate matter concentration data are also considered, that is, the linear regression model is a multiple linear regression model. For example, in addition to the real-time monitoring data of pollen concentration, the independent variables also include: meteorological data and / or atmospheric particulate matter concentration data collected at the same time as the pollen monitoring, the meteorological data includes one or more of temperature, relative humidity, precipitation, wind speed, and air pressure, and the atmospheric particulate matter concentration data includes PM 10 value.

[0102] That is to say, the initial function of constructing the multiple linear regression model is: Y=A'+B1X1+B2X2+

[0103] B3X3+B4X4+B5X5+B6X6, where X1, X2, X3, X4, X5, and X6 represent the real-time monitoring data of pollen concentration, temperature, relative humidity, precipitation, wind speed, air pressure, and PM2.5 collected simultaneously. 10 value.

[0104] (3) For the initial function, use the sample set for training

[0105] That is to say, the sample data in the pollen concentration standard sampling sample data set is used as the dependent variable, and the independent variables include the corresponding sample data in the pollen concentration real-time monitoring sample data set. The initial function is iteratively trained to obtain the pollen concentration detection model.

[0106] Specifically, the sample data set collected in (1) is used to iteratively train the initial function constructed in (2).

[0107] In some embodiments of the present invention, the K-fold cross-validation method is used in the training process of the linear regression model. When the preset number of iterations is reached, the iteration is stopped, and a model that meets the predetermined conditions of one or more of the coefficient of determination, root mean square error, and mean absolute error is selected as the pollen concentration detection model.

[0108] Specifically, the steps of training using the K-fold cross-validation method include:

[0109] The dataset consisting of the daily real-time pollen concentration monitoring dataset and the daily standard pollen concentration sampling dataset is divided into K subsets of equal size. In each iteration, one subset is used as the validation set, and the remaining subsets are used as the training set to train the initial monitoring model. Training ends when all subsets are used as validation sets.

[0110] Among them, a model that meets one or more predetermined conditions among the coefficient of determination, root mean square error and mean absolute error is selected as the pollen concentration detection model. 2 ), root mean square error (RMSE) and mean absolute error (MAE) were used to comprehensively evaluate the models and determine the model with the best performance.

[0111] 1. Simple Linear Regression Model

[0112] For example, the instructions are attached Figure 5 It is a simple linear regression model established by training using the real-time monitoring data set of pollen concentration and the standard sampling data set of pollen concentration collected for multiple consecutive months in 2023 and 2024 through the above-mentioned verification method.

[0113] The model obtained after training is Y = 0.95 + 0.94X, where X represents the real-time monitoring data of pollen concentration and Y is the corrected value (i.e., the value that should be obtained by sampling and counting using a standard Hurst trap). The coefficient of determination of this model is R 2 =0.76. The p-value is a test of statistical significance, used to measure whether the coefficient of the independent variable in a linear regression model is significantly different from zero. A p-value of <0.01 for this model indicates that the results are not due to random chance but rather demonstrate a real correlation and are statistically significant.

[0114] 2. Multiple linear regression model.

[0115] In order to study meteorological data, atmospheric particulate matter concentration data PM 10 The inventors constructed different multiple linear regression models for the specific effects of pollen concentration, using the sample data obtained in (1) above and the K-fold cross-validation method for training, and used the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and adjusted R 2 The adaptability of the regression model to the daily pollen concentration was compared with the variance inflation factor (VIF), and the results are shown in Table 2 below.

[0116] Table 2. Fitness results of different multiple linear regression models for pollen concentration

[0117] Model Independent variables included in the multiple linear regression model AIC BIC <![CDATA[Adjusted R 2 ]]> VIF Model 1 <![CDATA[KH3000,TEM,AP,RH,PM 10 ]]> 1946.25 3700.62 0.62 ＜10 Model 2 <![CDATA[KH3000,TEM,Precip.,RH,PM 2.5 ]]> 1684.04 3450.65 0.67 ＜10 Model 3 <![CDATA[KH3000,TEM,WS,RH,PM 10 ]]> 1591.84 3358.45 0.76 ＜10 Model 4 <![CDATA[KH3000,TEM,AP,RH,PM 2.5 ]]> 1794.37 3560.98 0.57 ＜10

[0118] Among them, TEM is temperature, AP is air pressure, RH is relative humidity, WS is wind speed, and Precip is precipitation.

[0119] According to the results in Table 2, Model 3 performs best, with the lowest AIC and BIC, and the adjusted R 2 Higher.

[0120] In other words, the independent variables also include real-time monitoring data of pollen concentration, temperature, relative humidity, wind speed, PM 10 The model obtained by ∑

[0121] In the following, the independent variables include real-time monitoring data of pollen concentration, temperature, relative humidity, wind speed, PM 10 It should be noted that the present invention is not limited to this, and the independent variables can be appropriately selected in combination with the specific local climate conditions.

[0122] For example, the pollen concentration real-time monitoring dataset and pollen concentration standard sampling dataset collected in multiple consecutive months in 2023 and 2024 (i.e. Figure 3 and Figure 4The sample data shown in the figure above is used to build a multiple linear regression model. The data set used in this example is the same as the data set used in the simple linear regression model. After training using Model 3 in Table 2 above, the multiple linear regression relationship is as follows:

[0123] Y 花粉(粒数 / m3) =2.575+0.93×KH3000+0.01×PM 10 +0.04×TEM-0.02×RH-0.79×WS

[0124] Among them, KH3000 represents the real-time monitoring data of daily pollen concentration, PM 10 Represents daily atmospheric particulate matter concentration data PM 10 value, TEM represents the daily mean temperature, RH represents the daily relative humidity, and WS represents the daily mean wind speed.

[0125] The coefficient of determination R of the model 2 =0.76, all variables were statistically significant (P<0.01).

[0126] Attachment Figure 6 This figure shows the relationship between the model predictions and the actual observed values for a multiple linear regression model. Each point on the graph represents a pair of predicted and observed values for a given observation point. The dotted line is the reference line y = x. The model shows a fairly high prediction accuracy, R 2 The value is 0.76, indicating that about 76% of the variance in actual pollen concentration can be explained by the multiple regression model (p < 0.01). Figure 6 As shown, the predicted values match the actual values well, as evidenced by the alignment of the data points along the diagonal reference line.

[0127] The linear regression model obtained through this training uses the real-time pollen concentration data obtained by the automatic real-time pollen monitoring device as the independent variable. By entering this data into the model, the actual pollen concentration detection value in the air (i.e., the corrected value) can be obtained. Of course, in the case of a multivariate linear regression model, meteorological data and atmospheric particulate matter concentration data are also required.

[0128] Reference Manual Figure 7 , which shows a real-time automatic detection device 700 for pollen concentration provided by an embodiment of the present invention, and its structure is as follows Figure 7 As shown, the automatic detection device 700 includes:

[0129] The pollen monitoring data acquisition unit 710 is used to acquire the real-time pollen concentration monitoring data obtained by the automatic real-time pollen monitoring device monitoring the pollen concentration in the air;

[0130] The data processing unit 720 is used to process the monitoring data using a pre-trained pollen concentration detection model to obtain a pollen concentration detection result.

[0131] Furthermore, the pollen concentration detection model is a multiple linear regression model, and the real-time automatic pollen concentration detection device 700 may further include:

[0132] A meteorological data collection unit for collecting meteorological data collected in the same period and in the same area as the pollen automatic real-time monitoring device; and / or

[0133] Atmospheric particulate matter data collection unit, used to collect atmospheric particulate matter concentration data in the same period and in the same area as the pollen automatic real-time monitoring device;

[0134] The data processing unit processes the monitoring data, meteorological data and / or atmospheric particulate matter concentration data to obtain a pollen concentration detection result.

[0135] It should be noted that the devices provided in the above embodiments are only illustrated by the division of the above functional modules when implementing their functions. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the devices provided in the above embodiments and the corresponding method embodiments are based on the same concept. The specific implementation process is detailed in the corresponding method embodiments and will not be repeated here.

[0136] One embodiment of the present invention also provides an electronic device, which includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the real-time pollen concentration detection method provided in the above method embodiment.

[0137] The memory can be used to store software programs and modules. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory. The memory can mainly include a program storage area and a data storage area. The program storage area can store the operating system, application programs required for the function, etc.; the data storage area can store data created based on the use of the device, etc. In addition, the memory can include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory can also include a memory controller to provide the processor with access to the memory.

[0138] Combined with the reference manual Figure 8, which is a block diagram of an electronic device 800 according to one embodiment of the present invention. The electronic device 800 may include one or more processors 802, a system control logic 808 connected to at least one of the processors 802, a system memory 804 connected to the system control logic 808, a non-volatile memory (NVM) 806 connected to the system control logic 808, and a network interface 810 connected to the system control logic 808.

[0139] The processor 802 may include one or more single-core or multi-core processors. The processor 802 may include any combination of general-purpose processors and specialized processors (e.g., graphics processors, application processors, baseband processors, etc.). In the embodiments herein, the processor 802 may be configured to execute Figure 1 The embodiment shown.

[0140] In some embodiments, system control logic 808 may include any suitable interface controller to provide any suitable interface to at least one of processors 802 and / or any suitable device or component in communication with system control logic 808 .

[0141] In some embodiments, the system control logic 808 may include one or more memory controllers to provide an interface to the system memory 804. The system memory 804 may be used to load and store data and / or instructions. In some embodiments, the memory 804 of the electronic device 800 may include any suitable volatile memory, such as a suitable dynamic random access memory (DRAM).

[0142] NVM / memory 806 may include one or more tangible, non-transitory computer-readable media for storing data and / or instructions. In some embodiments, NVM / memory 806 may include any suitable non-volatile memory such as flash memory and / or any suitable non-volatile storage device, such as at least one of an HDD (Hard Disk Drive), a CD (Compact Disc) drive, and a DVD (Digital Versatile Disc) drive.

[0143] NVM / storage 806 may include a portion of storage resources installed on a device of electronic device 800, or it may be accessible to the device but not necessarily a part of the device. For example, NVM / storage 806 may be accessed over a network via network interface 810.

[0144] In particular, system memory 804 and NVM / storage 806 may include, respectively, a temporary copy and a permanent copy of instructions 820. Instructions 820 may include instructions that, when executed by at least one of processors 802, cause electronic device 800 to perform the following operations: Figure 1 In some embodiments, the instructions 820, hardware, firmware, and / or software components thereof may be additionally or alternatively placed in the system control logic 808, the network interface 810, and / or the processor 802.

[0145] The network interface 810 may include a transceiver for providing a radio interface for the electronic device 800 to communicate with any other suitable devices (such as a front-end module, an antenna, etc.) via one or more networks. In some embodiments, the network interface 810 may be integrated with other components of the electronic device 800. For example, the network interface 810 may be integrated with at least one of a communication module of the processor 802, a system memory 804, an NVM / storage 806, and a firmware device (not shown) having instructions. When at least one of the processors 802 executes the instructions, the electronic device 800 implements Figure 1 The detection method of the embodiment shown.

[0146] The network interface 810 may further include any suitable hardware and / or firmware to provide a multiple-input multiple-output radio interface. For example, the network interface 810 may be a network adapter, a wireless network adapter, a telephone modem, and / or a wireless modem.

[0147] In one embodiment, at least one of the processors 802 may be packaged together with logic for one or more controllers of the system control logic 808 to form a system-in-package (SiP). In one embodiment, at least one of the processors 802 may be integrated on the same die with logic for one or more controllers of the system control logic 808 to form a system-on-chip (SoC).

[0148] Electronic device 800 may further include an input / output (I / O) device 812. I / O device 812 may include a user interface that enables a user to interact with electronic device 800; peripheral component interfaces may also be designed to enable peripheral components to interact with electronic device 800. In some embodiments, device 800 may also include a sensor for determining at least one of environmental conditions and location information related to electronic device 800.

[0149] In some embodiments, the user interface may include, but is not limited to, a display (e.g., an LCD display, a touch screen display, etc.), a speaker, a microphone, one or more cameras (e.g., a still image camera and / or a video camera), a flashlight (e.g., an LED flash), and a keyboard.

[0150] In some embodiments, the peripheral component interface may include, but is not limited to, a non-volatile memory port, an audio jack, and a power interface.

[0151] In some embodiments, the sensors may include, but are not limited to, a gyroscope sensor, an accelerometer, a proximity sensor, an ambient light sensor, and a positioning unit. The positioning unit may also be part of or interact with the network interface 810 to communicate with components of a positioning network (e.g., a Global Positioning System (GPS) satellite).

[0152] It should be understood that the structure illustrated in the embodiment of the present invention does not constitute a specific limitation on the electronic device 800. In other embodiments of the present invention, the electronic device 800 may include more or fewer components than shown, or may combine or separate certain components, or arrange the components differently. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0153] One embodiment of the present invention also provides a computer-readable storage medium, which can be set in an electronic device to store at least one instruction or at least one program related to implementing the real-time automatic detection method of pollen concentration. The at least one instruction or the at least one program is loaded and executed by the processor to implement the real-time automatic detection method of pollen concentration provided by the above method embodiment.

[0154] Optionally, in an embodiment of the present invention, the storage medium may include, but is not limited to, various media capable of storing program code, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard drive, a magnetic disk, or an optical disk. One embodiment of the present invention further provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the real-time automatic detection method for pollen concentration provided in the various optional implementations described above.

[0155] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not represent the superiority or inferiority of the embodiments. The above description is of specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0156] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences from other embodiments. In particular, the device embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0157] Those skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or by a program to instruct the relevant hardware, and the program may be stored in a computer-readable storage medium, which may be a read-only memory, a disk, or an optical disk, etc.

[0158] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A real-time automatic detection method for pollen concentration, characterized in that: The following steps are involved: Acquire monitoring data obtained by an automatic real-time pollen monitoring device; Based on the monitoring data, a pollen concentration detection result is obtained using a pre-trained pollen concentration detection model; The pollen concentration detection model is trained by respectively monitoring the pollen concentration in the air in the same preset period and geographical range using a pollen concentration real-time monitoring data sample set obtained by monitoring the pollen concentration in the air by the automatic real-time pollen monitoring device and a pollen concentration standard sampling data sample set collected by a Hurst-type trap.

2. The method for real-time automatic detection of pollen concentration according to claim 1, characterized in that: The training of the pollen concentration detection model includes: Respectively obtaining the pollen concentration real-time monitoring data sample set and the pollen concentration standard sampling data sample set; Construct the initial function of the linear regression model; The pollen concentration detection model is obtained by iteratively training the initial function using the sample data in the pollen concentration standard sampling sample data set as the dependent variable and the independent variable including the corresponding sample data in the pollen concentration real-time monitoring sample data set.

3. The method for real-time automatic detection of pollen concentration according to claim 2, characterized in that: The training also includes: After obtaining the pollen concentration real-time monitoring data sample set and the pollen concentration standard sampling data sample set, the data in the pollen concentration real-time monitoring data set and the pollen concentration standard sampling data set are preprocessed, and the iterative training is performed using the data in the preprocessed data set.

4. The method for real-time automatic detection of pollen concentration according to claim 3, characterized in that: The pretreatment includes: Data with a measurement time shorter than a preset time period in the pollen concentration real-time monitoring data set and the pollen concentration standard sampling data set are eliminated, and the data in the data set are arranged in natural time order.

5. The method for real-time automatic detection of pollen concentration according to claim 2, characterized in that: The linear regression model is a simple linear regression model.

6. The method for real-time automatic detection of pollen concentration according to claim 2, characterized in that: The linear regression model is a multiple linear regression model, and the independent variables also include: meteorological data and / or atmospheric particulate matter concentration data collected during the same period as pollen monitoring, the meteorological data including one or more of temperature, relative humidity, precipitation, wind speed, and air pressure, and the atmospheric particulate matter concentration data including PM 10 value.

7. The method for real-time automatic detection of pollen concentration according to claim 2, characterized in that: The K-fold cross-validation method is used in the training process of the linear regression model. When the preset number of iterations is reached, the iteration is stopped, and a model with one or more of the coefficient of determination, root mean square error and mean absolute error meeting the predetermined conditions is selected as the pollen concentration detection model.

8. The method for real-time automatic detection of pollen concentration according to claim 2, characterized in that: The acquisition of sample data in the pollen concentration standard sampling sample data set includes: collecting pollen samples using the Hoechst trap; staining the pollen sample with a toluidine blue solution to obtain a stained sample; The stained sample is counted under a fluorescence microscope to obtain sample data of the pollen sample.

9. A real-time automatic detection device for pollen concentration, characterized in that: include: A pollen monitoring data acquisition unit is used to acquire real-time pollen concentration monitoring data obtained by the automatic real-time pollen monitoring device monitoring the pollen concentration in the air; The data processing unit is used to process the monitoring data using a pre-trained pollen concentration detection model to obtain a pollen concentration detection result.

10. The real-time automatic detection device for pollen concentration according to claim 9, characterized in that: in, The pollen concentration detection model is a multiple linear regression model, and the real-time automatic detection device for pollen concentration further includes: A meteorological data collection unit for collecting meteorological data collected during the same period and in the same area as the pollen automatic real-time monitoring device; and / or An atmospheric particulate matter data collection unit, used to collect atmospheric particulate matter concentration data in the same period and in the same area as the pollen automatic real-time monitoring device; The data processing unit processes the monitoring data, meteorological data and / or atmospheric particulate matter concentration data to obtain the pollen concentration detection result.

Images