Micro-nano grating device data analysis method and system for optical sensing

Abstract

The invention relates to the technical field of micro-nano optics, in particular to a micro-nano grating device data analysis method and system for optical sensing. If a baseline drift phenomenon exists in one or more actual spectral characteristic data of the target micro-nano grating device, defining the actual spectral characteristic data with the baseline drift phenomenon as baseline drift spectral characteristic data; the correlation spectrum regulation and control unit queries a database according to the correlation spectrum regulation and control unit to determine baseline drift spectrum characteristic data; according to the method, various real-time working parameters of the correlation spectrum regulation and control unit of the baseline drift spectrum characteristic data are obtained, and parameter analysis processing is carried out on the various real-time working parameters of the correlation spectrum regulation and control unit of the baseline drift spectrum characteristic data. And the stability and adaptability of the optical sensing system are enhanced, and the optical sensing system has important significance for improving the overall performance of the optical sensing system.

Description

Technical Field

[0001] The present invention relates to the field of micro-nano optical technology, and in particular to a method and system for analyzing data of a micro-nano grating device for optical sensing. Background Art

[0002] With the rapid development of micro-nano grating technology, optical sensing technology based on micro-nano grating structures has been widely used in biomedicine, environmental monitoring, industrial detection and other fields. Micro-nano grating devices have become an important component of building optical sensing systems due to their advantages such as high sensitivity, high resolution and miniaturization. However, in practical applications, the optical properties of micro-nano grating devices will be affected by environmental factors, device aging, working parameter drift and other factors, causing their spectral response to change, thereby affecting the sensing accuracy and stability. For example, factors such as light source intensity fluctuations, detector noise, and temperature changes may cause baseline drift in the spectral signal, causing the actual spectral data obtained to deviate from the preset spectral characteristics, thereby affecting the accurate extraction of target sensing information.

[0003] At present, the methods for optical sensing data analysis are mainly concentrated on signal processing and feature extraction, such as wavelet transform, Fourier transform, principal component analysis, etc. However, most of these methods focus on the processing of the spectral signal itself, and lack the analysis of the reasons behind the changes in spectral characteristics. In addition, when dealing with baseline drift problems, existing methods usually use simple linear or polynomial fitting methods for correction, which is difficult to effectively deal with complex and changeable baseline drift phenomena. Therefore, there is an urgent need for a data comprehensive analysis and processing method that can combine the working state of micro-nano grating devices, changes in spectral characteristics, and correlation spectral control unit parameters to achieve accurate identification and effective compensation of baseline drift phenomena, thereby improving the stability and reliability of optical sensing systems. Summary of the invention

[0004] The present invention overcomes the deficiencies of the prior art and provides a method and system for analyzing data of a micro-nano grating device for optical sensing.

[0005] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is:

[0006] The present invention discloses a data analysis method for a micro-nano grating device for optical sensing, comprising the following steps:

[0007] Acquire the actual sensing conditions of the target micro-nano grating device at several preset time nodes, and determine various preset spectral characteristic data sets of the target micro-nano grating device within a preset time period based on the actual sensing conditions and in combination with the spectral characteristic query database;

[0008] Collecting actual spectral characteristic data of the target micro-nano grating device at a number of preset time nodes; screening the actual spectral characteristic data to obtain various actual spectral characteristic data sets within a preset time period;

[0009] Determine whether there is a baseline drift phenomenon in each actual spectral characteristic data of the target micro-nano grating device according to the preset spectral characteristic data set and the actual spectral characteristic data set within the preset time period;

[0010] If one or more actual spectral characteristic data of the target micro-nano grating device have a baseline drift phenomenon, the actual spectral characteristic data with the baseline drift phenomenon is defined as baseline drift spectral characteristic data; a correlation spectral control unit query database is constructed, and a correlation spectral control unit of the baseline drift spectral characteristic data is determined according to the correlation spectral control unit query database;

[0011] Acquire various real-time working parameters of the spectrum control unit of the correlation spectrum characteristic data of the baseline drift spectrum characteristic data, perform parameter analysis and processing on various real-time working parameters of the spectrum control unit of the correlation spectrum characteristic data of the baseline drift spectrum characteristic data, acquire abnormal working parameters, and perform adjustment processing on the abnormal working parameters.

[0012] Preferably, the actual sensing conditions of the target micro-nano grating device at several preset time nodes are obtained, and the preset spectral characteristic data sets of the target micro-nano grating device within a preset time period are determined based on the actual sensing conditions and in combination with the spectral characteristic query database, specifically:

[0013] Obtain various preset spectral characteristic data of the target micro-nano grating device under various preset sensing conditions through the big data network;

[0014] Construct a blank database, divide the blank database into several sub-databases, and import various preset spectral characteristic data of the target micro-nano grating device under various preset sensing conditions into corresponding sub-databases for storage, so as to obtain a spectral characteristic query database;

[0015] Acquire the actual sensing conditions of the target micro-nano grating device at a number of preset time nodes, and import the actual sensing conditions of each preset time node into the spectral characteristic query database;

[0016] Calculate the mutual information value between the actual sensing condition at each preset time node and the preset sensing condition in each sub-storage repository; and sort the mutual information values ​​between the actual sensing condition at each preset time node and the preset sensing condition in each sub-storage repository to obtain the sorting result;

[0017] Obtain a maximum mutual information value according to the sorting result, obtain a sub-repository corresponding to the maximum mutual information value, and extract various preset spectral characteristic data of the target micro-nano grating device when it works at a corresponding preset time node from the sub-repository corresponding to the maximum mutual information value;

[0018] And so on, until the preset spectral characteristic data of the target micro-nano grating device when it works at each preset time node are obtained, and the preset spectral characteristic data sets within the preset time period are generated according to the preset spectral characteristic data of the target micro-nano grating device when it works at each preset time node.

[0019] Preferably, the actual spectral characteristic data is subjected to screening processing to obtain various actual spectral characteristic data sets within a preset time period, specifically:

[0020] Acquire characteristic data corresponding to various spectral noise data and characteristic data corresponding to various spectral characteristic data through a big data network; and acquire the number of data types of the spectral noise data and the number of data types of the spectral characteristic data;

[0021] Constructing a binary tree model, dividing a corresponding number of first-class branches in the binary tree model according to the number of data types of the spectral noise data; dividing a corresponding number of second-class branches in the binary tree model according to the number of data types of the spectral characteristic data;

[0022] Mapping the characteristic data corresponding to each type of spectral noise data to a type of branch corresponding to the binary tree model; and mapping the characteristic data corresponding to each type of spectral characteristic data to a type of branch corresponding to the binary tree model;

[0023] Obtain the collected actual spectral characteristic data, and calculate the cosine similarity between each actual spectral characteristic data and each characteristic data in the binary tree model;

[0024] Allocate each actual spectral characteristic data to the first or second branch with the largest cosine similarity;

[0025] After completing the initial allocation of each actual spectral characteristic data, the contour coefficient of the actual spectral characteristic data in each of the first-class branches and the second-class branches is calculated; the contour coefficient of the actual spectral characteristic data in each of the first-class branches and the second-class branches is compared with a preset coefficient threshold;

[0026] If the silhouette coefficients of the actual spectral characteristic data in each of the first-class branches and the second-class branches are greater than the preset coefficient threshold, the initial allocation result is output as the final allocation result;

[0027] If there is a situation where the profile coefficient of the actual spectral characteristic data in at least one of the first-class branches or the second-class branches is not greater than the preset coefficient threshold, the actual spectral characteristic data of the first-class branches or the second-class branches whose profile coefficient is not greater than the preset coefficient threshold are redistributed until the profile coefficients of the actual spectral characteristic data in each of the first-class branches and the second-class branches are greater than the preset coefficient threshold, and the last distribution result is output as the final distribution result;

[0028] After the final allocation result is obtained, each of the first-class branches and the second-class branches of the binary tree model is pruned to obtain an actual spectral characteristic data set and a spectral noise data set.

[0029] Preferably, judging whether there is a baseline drift phenomenon in each actual spectral characteristic data of the target micro-nano grating device according to the preset spectral characteristic data set and the actual spectral characteristic data set within the preset time period is specifically as follows:

[0030] Sorting each actual spectral characteristic data set based on the acquisition timestamp to obtain an actual spectral characteristic data set based on the time series; constructing a curve diagram of each actual spectral characteristic data change according to the actual spectral characteristic data set based on the time series;

[0031] Acquire various preset spectral characteristic data sets of the target micro-nano grating device within a preset time period, and construct various preset spectral characteristic data change curves according to the various preset spectral characteristic data sets of the target micro-nano grating device within the preset time period;

[0032] Performing a coincidence analysis on each actual spectral characteristic data change curve graph and the corresponding preset spectral characteristic data change curve graph to obtain the degree of coincidence between each actual spectral characteristic data change curve graph and the corresponding preset spectral characteristic data change curve graph;

[0033] If the degree of overlap between a certain actual spectral characteristic data change curve and a corresponding preset spectral characteristic data change curve is not greater than a preset overlap threshold, it indicates that the actual spectral characteristic data in the micro-nano grating device has a baseline drift phenomenon;

[0034] If the degree of overlap between a certain actual spectral characteristic data change curve and the corresponding preset spectral characteristic data change curve is greater than the preset overlap threshold, it means that there is no baseline drift phenomenon in the actual spectral characteristic data in the micro-nano grating device.

[0035] Preferably, a correlation spectrum control unit query database is constructed, and a correlation spectrum control unit of the baseline drift spectrum characteristic data is determined according to the correlation spectrum control unit query database, specifically:

[0036] Obtaining a product structure specification of a target micro-nano grating device, determining each spectral control unit of the target micro-nano grating device according to the product structure specification, and obtaining control function characteristic information of each spectral control unit;

[0037] According to the control function characteristic information of each spectral control unit, correlation analysis is performed on each spectral control unit and each spectral characteristic data, and spectral control units having correlation with each spectral characteristic data are obtained to obtain correlation spectral control units of each spectral characteristic data;

[0038] Constructing a correlation spectrum regulation unit query database according to the correlation spectrum regulation unit of each spectrum characteristic data;

[0039] The baseline drift spectral characteristic data in the target micro-nano grating device is obtained, and the baseline drift spectral characteristic data in the target micro-nano grating device is imported into the correlation spectrum control unit query database for query matching to obtain the correlation spectrum control unit of the baseline drift spectral characteristic data.

[0040] Preferably, various real-time working parameters of the correlation spectrum control unit of the baseline drift spectrum characteristic data are obtained, parameter analysis and processing are performed on various real-time working parameters of the correlation spectrum control unit of the baseline drift spectrum characteristic data, abnormal working parameters are obtained, and the abnormal working parameters are adjusted, specifically:

[0041] Acquire various real-time working parameters of the spectrum control unit of the correlation spectrum characteristic data of the baseline drift, and calculate the difference between each real-time working parameter and the corresponding preset working parameter to obtain the parameter deviation value of each real-time working parameter;

[0042] Preset the deviation value threshold of each real-time working parameter, and compare the parameter deviation value of each real-time working parameter with the corresponding deviation value threshold;

[0043] If the parameter deviation value of a certain real-time working parameter is greater than the corresponding deviation value threshold, the real-time working parameter is defined as an abnormal working parameter;

[0044] If the parameter deviation value of a certain real-time working parameter is not greater than the corresponding deviation value threshold, the real-time working parameter is defined as a normal working parameter;

[0045] A parameter deviation value defined as an abnormal operating parameter is obtained, and the abnormal operating parameter is adjusted based on the corresponding parameter deviation value to adjust the corresponding real-time operating parameter to within a normal range.

[0046] Among them, the sensing conditions include light intensity, dust concentration, temperature, humidity and air pressure; the spectral characteristic data include reflection spectrum data, transmission spectrum data, absorption spectrum data, Raman spectrum data, ellipsometric spectrum data and dispersion spectrum data; the real-time working parameters include light source wavelength, light source intensity, polarization state of light source light, incident angle, bias voltage and current.

[0047] On the other hand, the present invention discloses a micro-nano grating device data analysis system for optical sensing, the micro-nano grating device data analysis system includes a memory and a processor, the memory stores a micro-nano grating device data analysis method program, when the micro-nano grating device data analysis method program is executed by the processor, any one of the steps of the micro-nano grating device data analysis method is implemented.

[0048] The present invention solves the technical defects existing in the background technology, and the present invention has the following beneficial effects: First, in terms of data processing, by analyzing the preset spectral characteristic data set and the actual spectral characteristic data set, it is possible to accurately determine whether there is a baseline drift phenomenon, thereby improving the measurement accuracy. Secondly, by constructing a correlation spectrum control unit query database, the control unit that affects the spectral characteristics can be accurately located, and the possible factors that cause the baseline drift can be quickly found, thereby enhancing the pertinence of system maintenance. Furthermore, the analysis and processing of the real-time working parameters of the correlation spectrum control unit can accurately identify abnormal working parameters and make adjustments, effectively correct the anomalies, and improve the stability and accuracy of the system. In addition, the present method can also achieve early fault prevention, correct problems before abnormal parameters cause serious impacts, and help optimize system performance, extend the service life of micro-nano grating devices, reduce maintenance costs, and improve their adaptability to different environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0050] Figure 1 This is the overall method flow chart of the data analysis method of the micro-nano grating device;

[0051] Figure 2 This is a partial method flow chart of the data analysis method of the micro-nano grating device;

[0052] Figure 3 This is the system block diagram of the micro-nano grating device data analysis system. DETAILED DESCRIPTION

[0053] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0054] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0055] The present invention discloses a data analysis method for a micro-nano grating device for optical sensing, such as Figure 1 As shown, the following steps are included:

[0056] S102: Acquire actual sensing conditions of the target micro-nano grating device at a number of preset time nodes, and determine various preset spectral characteristic data sets of the target micro-nano grating device within a preset time period based on the actual sensing conditions and in combination with a spectral characteristic query database;

[0057] S104: collecting actual spectral characteristic data of the target micro-nano grating device at a number of preset time nodes; screening the actual spectral characteristic data to obtain various actual spectral characteristic data sets within a preset time period;

[0058] S106: judging whether there is a baseline drift phenomenon in each actual spectral characteristic data of the target micro-nano grating device according to the preset spectral characteristic data set and the actual spectral characteristic data set within the preset time period;

[0059] S108: If one or more actual spectral characteristic data of the target micro-nano grating device have a baseline drift phenomenon, the actual spectral characteristic data with the baseline drift phenomenon is defined as baseline drift spectral characteristic data; a correlation spectral control unit query database is constructed, and a correlation spectral control unit of the baseline drift spectral characteristic data is determined according to the correlation spectral control unit query database;

[0060] S110: Acquire various real-time working parameters of the correlation spectrum control unit of the baseline drift spectrum characteristic data, perform parameter analysis on various real-time working parameters of the correlation spectrum control unit of the baseline drift spectrum characteristic data, acquire abnormal working parameters, and perform adjustment processing on the abnormal working parameters.

[0061] Among them, the sensing conditions include light intensity, dust concentration, temperature, humidity and air pressure; the spectral characteristic data include reflection spectrum data, transmission spectrum data, absorption spectrum data, Raman spectrum data, ellipsometric spectrum data and dispersion spectrum data; the real-time working parameters include light source wavelength, light source intensity, polarization state of light source light, incident angle, bias voltage and current.

[0062] This method is mainly used to monitor and correct the changes in the spectral characteristics of micro-nano grating devices under different environmental conditions to ensure the accuracy and stability of its measurement results. First, it is necessary to obtain the actual sensing conditions of the micro-nano grating device at multiple preset time points, including light intensity, dust concentration, temperature, humidity, and air pressure. According to the actual sensing conditions, the expected spectral characteristic data set of the target micro-nano grating device within the preset time period is determined in combination with the spectral characteristic database. The actual spectral characteristic data of the target micro-nano grating device is collected at the preset time point, including reflection spectrum, transmission spectrum, absorption spectrum, Raman spectrum, ellipsometry spectrum, and dispersion spectrum. The collected actual spectral characteristic data are screened and processed to extract the actual spectral characteristic data set within the preset time period. The expected spectral characteristic data set is compared with the actual spectral characteristic data set to determine whether there is a baseline drift phenomenon. Baseline drift refers to the phenomenon that the spectral data systematically shifts over time. If baseline drift is detected in the actual spectral characteristic data, it is necessary to identify the relevant spectral control unit and its operating parameters (such as light source wavelength, light source intensity, polarization state of the light source, incident angle, bias voltage and current, etc.) that cause this phenomenon in order to eliminate the baseline drift.

[0063] By monitoring and correcting the baseline drift phenomenon in the spectral characteristic data, the measurement accuracy of micro-nano grating devices under different environmental conditions can be significantly improved. It helps to identify and correct various factors that affect the spectral characteristics, thereby enhancing the stability and reliability of the long-term operation of micro-nano grating devices. Since a variety of environmental factors (such as temperature, humidity, etc.) are taken into account, this method has strong environmental adaptability and is suitable for a variety of application scenarios. In summary, this data analysis method not only improves the measurement accuracy of micro-nano grating devices in complex environments, but also enhances its stability and adaptability, which is of great significance for improving the overall performance of optical sensing systems.

[0064] Preferably, the actual sensing conditions of the target micro-nano grating device at several preset time nodes are obtained, and the preset spectral characteristic data sets of the target micro-nano grating device within a preset time period are determined based on the actual sensing conditions and in combination with the spectral characteristic query database, such as Figure 2 As shown, specifically:

[0065] S202: Acquire various preset spectral characteristic data of the target micro-nano grating device under various preset sensing conditions through a big data network;

[0066] S204: construct a blank database, divide the blank database into a plurality of sub-databases, and import various preset spectral characteristic data of the target micro-nano grating device under various preset sensing conditions into corresponding sub-databases for storage, thereby obtaining a spectral characteristic query database;

[0067] S206: Acquire actual sensing conditions of the target micro-nano grating device at a number of preset time nodes, and import the actual sensing conditions of each preset time node into the spectral characteristic query database;

[0068] S208: Calculate the mutual information value between the actual sensing condition at each preset time node and the preset sensing condition in each sub-storage repository; and sort the mutual information values ​​between the actual sensing condition at each preset time node and the preset sensing condition in each sub-storage repository to obtain a sorting result;

[0069] S210: Obtaining a maximum mutual information value according to the sorting result, obtaining a sub-repository corresponding to the maximum mutual information value, and extracting various preset spectral characteristic data of the target micro-nano grating device when it works at a corresponding preset time node from the sub-repository corresponding to the maximum mutual information value;

[0070] S212: And so on, until the preset spectral characteristic data of the target micro-nano grating device when it works at each preset time node are obtained, and the preset spectral characteristic data sets within the preset time period are generated according to the preset spectral characteristic data of the target micro-nano grating device when it works at each preset time node.

[0071] The steps of this method realize efficient management and query of spectral characteristic data by constructing a database, importing data, calculating mutual information values, etc. Specifically, various preset spectral characteristic data of the target micro-nano grating device under various preset sensing conditions are obtained through a big data network; a blank database is constructed and divided into several sub-repositories. Various preset spectral characteristic data of the target micro-nano grating device under various preset sensing conditions are respectively imported into corresponding sub-repositories for storage to form a spectral characteristic query database. The actual sensing conditions of the target micro-nano grating device at several preset time nodes are obtained, and these actual sensing conditions are imported into the spectral characteristic query database. The mutual information value between the actual sensing conditions at each preset time node and the preset sensing conditions in each sub-repository is calculated. The mutual information value is a statistic that measures the degree of mutual dependence between two random variables. The mutual information values ​​between the actual sensing conditions at each preset time node and the preset sensing conditions in each sub-repository are sorted to obtain the sorting result. The maximum mutual information value is obtained according to the sorting result, and the sub-repository corresponding to the maximum mutual information value is found, and the various preset spectral characteristic data of the target micro-nano grating device when working at the corresponding preset time node are extracted from it. Repeat the above steps until the preset spectral characteristic data of the target micro-nano grating device when it works at each preset time node are obtained. According to the preset spectral characteristic data of the target micro-nano grating device when it works at each preset time node, various preset spectral characteristic data sets within a preset time period are generated.

[0072] By constructing a spectral characteristic query database, the preset spectral characteristic data are classified and stored according to the preset sensing conditions, which is convenient for rapid retrieval and use. The mutual information value calculation method can accurately match the similarity between the actual sensing conditions and the preset sensing conditions, thereby improving the accuracy of data matching. By sorting and selecting the maximum mutual information value, the preset spectral characteristic data closest to the actual sensing conditions can be quickly found, reducing the query time and computing resource consumption. The database design is flexible and can be expanded and adjusted according to different application scenarios and requirements to improve the adaptability and flexibility of the system. By accurately matching and extracting the preset spectral characteristic data, the measurement accuracy and reliability of micro-nano grating devices in practical applications can be improved. In summary, this method can not only efficiently manage and query spectral characteristic data, but also improve the accuracy of data matching through mutual information value calculation, thereby improving data quality, data analysis accuracy and system flexibility.

[0073] Preferably, the actual spectral characteristic data is subjected to screening processing to obtain various actual spectral characteristic data sets within a preset time period, specifically:

[0074] Acquire characteristic data corresponding to various spectral noise data and characteristic data corresponding to various spectral characteristic data through a big data network; and acquire the number of data types of the spectral noise data and the number of data types of the spectral characteristic data;

[0075] Constructing a binary tree model, dividing a corresponding number of first-class branches in the binary tree model according to the number of data types of the spectral noise data; dividing a corresponding number of second-class branches in the binary tree model according to the number of data types of the spectral characteristic data;

[0076] Mapping the characteristic data corresponding to each type of spectral noise data to a type of branch corresponding to the binary tree model; and mapping the characteristic data corresponding to each type of spectral characteristic data to a type of branch corresponding to the binary tree model;

[0077] Obtain the collected actual spectral characteristic data, and calculate the cosine similarity between each actual spectral characteristic data and each characteristic data in the binary tree model;

[0078] Allocate each actual spectral characteristic data to the first or second branch with the largest cosine similarity;

[0079] After completing the initial allocation of each actual spectral characteristic data, the contour coefficient of the actual spectral characteristic data in each of the first-class branches and the second-class branches is calculated; the contour coefficient of the actual spectral characteristic data in each of the first-class branches and the second-class branches is compared with a preset coefficient threshold;

[0080] If the silhouette coefficients of the actual spectral characteristic data in each of the first-class branches and the second-class branches are greater than the preset coefficient threshold, the initial allocation result is output as the final allocation result;

[0081] If there is a situation where the profile coefficient of the actual spectral characteristic data in at least one of the first-class branches or the second-class branches is not greater than the preset coefficient threshold, the actual spectral characteristic data of the first-class branches or the second-class branches whose profile coefficient is not greater than the preset coefficient threshold are redistributed until the profile coefficients of the actual spectral characteristic data in each of the first-class branches and the second-class branches are greater than the preset coefficient threshold, and the last distribution result is output as the final distribution result;

[0082] After the final allocation result is obtained, each of the first-class branches and the second-class branches of the binary tree model is pruned to obtain an actual spectral characteristic data set and a spectral noise data set.

[0083] It should be noted that first, two types of feature data are obtained through the big data network, namely, the feature data corresponding to various spectral noise data and the feature data corresponding to various spectral characteristic data (such as reflection spectrum, transmission spectrum, etc.). At the same time, the number of data types of spectral noise data and the number of data types of spectral characteristic data are also obtained. This step provides basic data for the subsequent construction of the binary tree model. These data are the basis for accurate screening of spectral characteristics. Different types of spectral noise and spectral characteristics have their own unique characteristics. By obtaining their feature data, the actual spectral characteristic data can be classified according to these characteristics in the subsequent steps. According to the number of data types of spectral noise data, a corresponding number of first-class branches are cut out in the binary tree model, and similarly, according to the number of data types of spectral characteristic data, a corresponding number of second-class branches are cut out. The binary tree model is a data structure in which each node has at most two child nodes. This structure helps to classify and organize data. The feature data corresponding to each type of spectral noise data is mapped to the first-class branch corresponding to the binary tree model, and the feature data corresponding to each type of spectral characteristic data is mapped to the corresponding second-class branch. In this way, the binary tree model becomes a framework for classifying and storing spectral noise and spectral characteristic feature data. After obtaining the collected actual spectral characteristic data, the cosine similarity between each actual spectral characteristic data and each characteristic data in the binary tree model is calculated respectively. Cosine similarity is an indicator to measure the similarity between two vectors. In this scenario, it is used to judge the similarity between the actual spectral characteristic data and the characteristic data stored in the binary tree model. According to the cosine similarity, each actual spectral characteristic data is respectively assigned to the first-class branch or the second-class branch with the largest cosine similarity. This step is to perform a preliminary classification of the actual spectral characteristic data based on similarity. After the initial allocation of each actual spectral characteristic data, the silhouette coefficient of the actual spectral characteristic data in each first-class branch and the second-class branch is calculated. The silhouette coefficient is an indicator for evaluating the clustering effect. It measures the closeness of a data point with other data points in the cluster to which it belongs and the degree of separation from other cluster data points. Then the silhouette coefficient is compared with the preset coefficient threshold. If the silhouette coefficients of the actual spectral characteristic data in each first-class branch and the second-class branch are greater than the preset coefficient threshold, it means that the clustering effect of the initial allocation is good, and the initial allocation result is output as the final allocation result. If there is a situation where the silhouette coefficient of the actual spectral characteristic data in at least one of the first or second branches is not greater than the preset coefficient threshold, it means that the clustering effect is not ideal, and the actual spectral characteristic data of the first or second branches whose silhouette coefficient is not greater than the preset coefficient threshold need to be redistributed. This process will be repeated until the silhouette coefficients of the actual spectral characteristic data in each of the first and second branches are greater than the preset coefficient threshold, and the last allocation result will be output as the final allocation result.After obtaining the final allocation result, the first-class branches and the second-class branches of the binary tree model are pruned to obtain the actual spectral characteristic data set and the spectral noise data set. The pruned process is to extract the data from the branch to form the final actual spectral characteristic data set and the spectral noise data set.

[0084] By constructing a binary tree model and performing data allocation and adjustment based on cosine similarity and silhouette coefficient, the actual spectral characteristic data can be accurately separated from the spectral noise data to obtain an accurate actual spectral characteristic data set. This precise classification helps to accurately analyze the spectral characteristics in the subsequent analysis and avoids the interference of noise on the analysis results. By calculating the silhouette coefficient and comparing it with the preset threshold, the allocation results are evaluated and adjusted multiple times to optimize the clustering effect, which makes the final data set have a high similarity internally and a high difference with other data sets externally, thus improving the quality of the data. High-quality data sets are of great significance for subsequent spectral characteristic research, optical sensing analysis, etc. For example, in optical sensing, the performance of micro-nano grating devices can be more accurately analyzed to improve the accuracy and reliability of measurement. Using the binary tree model for data mapping, allocation and adjustment, this data structure has efficient search and classification capabilities. When processing a large amount of spectral data, data processing can be performed quickly, reducing computing time and resource consumption. The entire process is automatically performed based on data features and algorithms, reducing the need for manual intervention. This not only improves the efficiency of data processing, but also reduces the impact of human factors on data processing results and improves the objectivity of the results.

[0085] Preferably, judging whether there is a baseline drift phenomenon in each actual spectral characteristic data of the target micro-nano grating device according to the preset spectral characteristic data set and the actual spectral characteristic data set within the preset time period is specifically as follows:

[0086] Sorting each actual spectral characteristic data set based on the acquisition timestamp to obtain an actual spectral characteristic data set based on the time series; constructing a curve diagram of each actual spectral characteristic data change according to the actual spectral characteristic data set based on the time series;

[0087] Acquire various preset spectral characteristic data sets of the target micro-nano grating device within a preset time period, and construct various preset spectral characteristic data change curves according to the various preset spectral characteristic data sets of the target micro-nano grating device within the preset time period;

[0088] Performing a coincidence analysis on each actual spectral characteristic data change curve graph and the corresponding preset spectral characteristic data change curve graph to obtain the degree of coincidence between each actual spectral characteristic data change curve graph and the corresponding preset spectral characteristic data change curve graph;

[0089] If the degree of overlap between a certain actual spectral characteristic data change curve and a corresponding preset spectral characteristic data change curve is not greater than a preset overlap threshold, it indicates that the actual spectral characteristic data in the micro-nano grating device has a baseline drift phenomenon;

[0090] If the degree of overlap between a certain actual spectral characteristic data change curve and the corresponding preset spectral characteristic data change curve is greater than the preset overlap threshold, it means that there is no baseline drift phenomenon in the actual spectral characteristic data in the micro-nano grating device.

[0091] It should be noted that the spectral characteristic data change curve diagram intuitively shows the changing trend of the spectral characteristic data over time. The actual spectral characteristic data change curve diagram and the corresponding preset spectral characteristic data change curve diagram are analyzed for coincidence, and the coincidence degree between the two is obtained. The coincidence degree can be understood as the degree of coincidence between the actual data and the preset data in the changing trend. Whether there is a baseline drift phenomenon is judged according to the coincidence degree. By comparing the change curve diagram of the actual spectral characteristic data with the preset spectral characteristic data, it can be accurately judged whether there is a baseline drift phenomenon. The baseline drift phenomenon will affect the stability of the measurement results. By timely detecting and correcting the baseline drift phenomenon, the long-term working stability of the micro-nano grating device can be enhanced. The detection of the baseline drift phenomenon by an automated method reduces the need for manual intervention and improves the accuracy and efficiency of the detection. In summary, this method can effectively detect and judge the baseline drift phenomenon by comparing the change curve diagram of the actual spectral characteristic data with the preset spectral characteristic data, thereby improving the measurement accuracy and system stability.

[0092] Preferably, a correlation spectrum control unit query database is constructed, and a correlation spectrum control unit of the baseline drift spectrum characteristic data is determined according to the correlation spectrum control unit query database, specifically:

[0093] Obtaining a product structure specification of a target micro-nano grating device, determining each spectral control unit of the target micro-nano grating device according to the product structure specification, and obtaining control function characteristic information of each spectral control unit;

[0094] According to the control function characteristic information of each spectral control unit, correlation analysis is performed on each spectral control unit and each spectral characteristic data, and spectral control units having correlation with each spectral characteristic data are obtained to obtain correlation spectral control units of each spectral characteristic data;

[0095] Constructing a correlation spectrum regulation unit query database according to the correlation spectrum regulation unit of each spectrum characteristic data;

[0096] The baseline drift spectral characteristic data in the target micro-nano grating device is obtained, and the baseline drift spectral characteristic data in the target micro-nano grating device is imported into the correlation spectrum control unit query database for query matching to obtain the correlation spectrum control unit of the baseline drift spectral characteristic data.

[0097] It should be noted that the product structure specification contains important information about the structural design and components of the micro-nano grating device, which is the basis for subsequent analysis. The spectral control unit is a component or functional module that can affect the spectral characteristics. In addition, the control function characteristic information of each spectral control unit is obtained, which describes how each spectral control unit affects the spectral characteristics, such as the ability to change wavelength, intensity, etc. Using the control function characteristic information of each spectral control unit, a correlation analysis is performed on each spectral control unit and each spectral characteristic data (such as reflection spectrum, transmission spectrum, etc.), such as through gray correlation analysis, or combined with manual experience. This analysis aims to find out which spectral control units will affect specific spectral characteristic data. Through correlation analysis, the correlation spectral control unit of each spectral characteristic data is determined. For example, a spectral control unit may have a strong correlation with the reflection spectrum data, but a weak correlation with the transmission spectrum data.

[0098] A correlation spectrum control unit query database is constructed based on the correlation spectrum control units of each spectral characteristic data obtained previously. This database is constructed to facilitate subsequent query matching. It organizes the spectral characteristic data and the spectral control units related thereto in a way that is easy to query. The baseline drift spectral characteristic data in the target micro-nano grating device is obtained. These data are spectral characteristic data that have been previously determined to have baseline drift phenomena. The baseline drift spectral characteristic data are imported into the correlation spectrum control unit query database for query matching. Through the associated information in the database, the correlation spectrum control units corresponding to these baseline drift spectral characteristic data are found.

[0099] In summary, by querying and matching the relevant spectral control units of the baseline drift spectral characteristic data, it is possible to quickly locate which spectral control units may be the cause of the baseline drift, which is of great significance for timely solving the performance problems of micro-nano grating devices and improving their working stability. Once the spectral control units related to the baseline drift are determined, these units can be maintained, adjusted or optimized in a targeted manner, which can more accurately solve the baseline drift problem, thereby improving the overall performance of the micro-nano grating device.

[0100] Preferably, various real-time working parameters of the correlation spectrum control unit of the baseline drift spectrum characteristic data are obtained, parameter analysis and processing are performed on various real-time working parameters of the correlation spectrum control unit of the baseline drift spectrum characteristic data, abnormal working parameters are obtained, and the abnormal working parameters are adjusted, specifically:

[0101] Acquire various real-time working parameters of the spectrum control unit of the correlation spectrum characteristic data of the baseline drift, and calculate the difference between each real-time working parameter and the corresponding preset working parameter to obtain the parameter deviation value of each real-time working parameter;

[0102] Preset the deviation value threshold of each real-time working parameter, and compare the parameter deviation value of each real-time working parameter with the corresponding deviation value threshold;

[0103] If the parameter deviation value of a certain real-time working parameter is greater than the corresponding deviation value threshold, the real-time working parameter is defined as an abnormal working parameter;

[0104] If the parameter deviation value of a certain real-time working parameter is not greater than the corresponding deviation value threshold, the real-time working parameter is defined as a normal working parameter;

[0105] A parameter deviation value defined as an abnormal operating parameter is obtained, and the abnormal operating parameter is adjusted based on the corresponding parameter deviation value to adjust the corresponding real-time operating parameter to within a normal range.

[0106] It should be noted that, first, the various real-time working parameters of the correlation spectrum control unit of the baseline drift spectrum characteristic data are obtained. These real-time working parameters are key indicators reflecting the current working state of the correlation spectrum control unit, such as light source wavelength, light source intensity, polarization state of light source light, incident angle, bias voltage and current. Then the parameter deviation values ​​of various real-time working parameters are calculated. The preset working parameters are the values ​​that these parameters should have under normal working conditions. By calculating the deviation values, the degree of deviation between the current real-time working parameters and the ideal state can be intuitively understood. The deviation value thresholds of various real-time working parameters are pre-set. This threshold is the basis for judging whether the working parameters are abnormal. Its setting needs to take into account various factors such as the accuracy requirements, stability requirements and actual working environment of the system. If the parameter deviation value of a real-time working parameter is greater than the corresponding deviation value threshold, it means that the parameter deviates greatly from the normal range, and the real-time working parameter is defined as an abnormal working parameter; conversely, if the parameter deviation value of a real-time working parameter is not greater than the corresponding deviation value threshold, the real-time working parameter is defined as a normal working parameter. For the case defined as an abnormal working parameter, its parameter deviation value is obtained. This deviation value contains information about the degree and direction of abnormality, which is an important basis for adjustment processing. The abnormal working parameters are adjusted based on the corresponding parameter deviation values, with the purpose of adjusting the corresponding real-time working parameters to the normal range. The adjustment method varies according to the specific working parameter type. For example, the light source intensity can be achieved by adjusting the power supply voltage or current, and the incident angle can be achieved by adjusting the position of the optical element.

[0107] Adjustment is performed based on the deviation value of the abnormal working parameter. This targeted adjustment method can directly act on the abnormal parameter to adjust it to the normal range, which helps to quickly restore the normal working state of the correlation spectrum control unit, thereby reducing or eliminating the baseline drift phenomenon. By correcting the abnormal working parameters, the working state of the correlation spectrum control unit can be made more stable, thereby improving the working stability of the entire micro-nano grating device. A stable working state helps to improve the accuracy of spectral characteristic measurement and ensure the performance of micro-nano grating devices in applications such as optical sensing.

[0108] In addition, the method further comprises the following steps:

[0109] Obtaining the corresponding fault characteristic data when various types of faults occur in the target micro-nano grating device through the big data network; and obtaining the historical working characteristic data of the target micro-nano grating device within a preset time period before various types of faults occur;

[0110] The maximum likelihood estimation method is introduced to estimate the historical working characteristic data of the target micro-nano grating device in a preset time period before various types of faults occur, and the historical working characteristic data in a preset time period before various types of faults occur are estimated based on the maximum likelihood estimation method, so as to obtain the transition probability of the target micro-nano grating device from a normal state to a fault state under various historical working characteristic data conditions;

[0111] A transition probability matrix is ​​constructed according to the transition probability of the target micro-nano grating device from a normal state to a fault state under various historical working characteristic data conditions;

[0112] Constructing a support vector machine model, and importing the transition probability matrix into the support vector machine model for coding learning, until the model parameters are learned to meet the preset requirements;

[0113] When the abnormal working parameters are adjusted, it is continuously determined whether the target micro-nano grating device still has a baseline drift phenomenon. If so, the real-time working characteristic data of the target micro-nano grating device at the current time node is obtained;

[0114] Importing the real-time working characteristic data of the target micro-nano grating device at the current time node into the support vector machine model for evaluation, and outputting the transition probability of the target micro-nano grating device from a normal working state to a fault state at the current time node and several future time nodes;

[0115] If the transition probability of the target micro-nano grating device from the normal working state to the fault state at the current time node is greater than the preset transition probability value, the current time node is output as the fault warning time node;

[0116] If the transition probability of the target micro-nano grating device from a normal working state to a fault state at a certain future time node is greater than a preset transition probability value, the future time node is output as a fault warning time node.

[0117] It should be noted that first, two aspects of data are obtained through the big data network: one is the corresponding fault feature data when various types of faults occur in the target micro-nano grating device. These data can reflect the performance characteristics of the device under different fault types, such as specific spectral changes, electrical parameter anomalies, etc.; the second is to obtain the historical working feature data of the target micro-nano grating device in the preset time period before various types of faults occur. These data record various characteristics of the device in the normal working state before the fault is about to occur, such as working temperature, working voltage, spectral characteristics, etc. The maximum likelihood estimation method is introduced to estimate the historical working feature data of the target micro-nano grating device in the preset time period before various types of faults occur. Here, the historical working feature data before the fault is taken as input to estimate the transition probability of the device from the normal state to the fault state under these data conditions. The transition probability reflects the probability of the device changing from the normal state to the fault state under specific historical working feature data conditions. It is a key indicator for quantifying the possibility of device failure. It is obtained by estimating a large amount of historical data and can comprehensively consider the impact of multiple working characteristics on the occurrence of faults. The transition probability matrix is ​​an integrated representation of the relationship between historical working feature data and fault transition probability. It provides a data basis for the support vector machine model, enabling the model to learn the probability pattern of faults under different working features. The transition probability matrix is ​​imported into the support vector machine model for encoding learning. During the learning process, the model adjusts its own parameters according to the data in the matrix to find the best classification or regression rules. The preset requirements include that the performance indicators such as the accuracy and recall rate of the model meet certain standards, or that the loss function value of the model is reduced to a certain level. Through such a learning process, the support vector machine model can learn the complex relationship between historical working feature data and fault transition probability.

[0118] After adjusting the abnormal working parameters, if the target micro-nano grating device still has baseline drift, the real-time working characteristic data of the device at the current time node is obtained. Then these real-time working characteristic data are imported into the support vector machine model for evaluation. The support vector machine model will output the transition probability of the target micro-nano grating device from the normal working state to the fault state at the current time node and several future time nodes according to the input real-time working characteristic data. If the transition probability at the current time node is greater than the preset transition probability value, the current time node is output as the fault warning time node; if the transition probability at a certain future time node is greater than the preset transition probability value, the future time node is output as the fault warning time node. This step realizes the real-time prediction and warning of device failures, and can detect possible failures in advance so as to take corresponding measures. By judging the fault transition probability at the current and future time nodes, real-time fault warning is realized, which makes it possible to take measures in advance to avoid the occurrence of failures or reduce the losses caused by failures, thereby improving the reliability and stability of the micro-nano grating device.

[0119] In addition, the method further comprises the following steps:

[0120] Acquire structural engineering drawing information of the target micro-nano grating device, construct a three-dimensional structure diagram of the target micro-nano grating device according to the structural engineering drawing information; and acquire preset working parameter range information of each component unit in the target micro-nano grating device;

[0121] Establishing a mapping relationship between each component unit and the preset working parameter range information, and importing the preset working parameter range information of each component unit into the three-dimensional structure diagram according to the mapping relationship between each component unit and the preset working parameter range information;

[0122] Set the parameters of dynamic simulation, including time step, simulation duration and boundary conditions; obtain the dynamic simulation model diagram of the target micro-nano grating device;

[0123] If the transition probability of the target micro-nano grating device from the normal working state to the fault state at the current time node is greater than the preset transition probability value, or if the transition probability of the target micro-nano grating device from the normal working state to the fault state at a future time node is greater than the preset transition probability value; then the fault type of the target micro-nano grating device transitioning from the normal working state to the fault state is obtained;

[0124] According to the fault type of the target micro-nano grating device transitioning from a normal working state to a fault state, several maintenance plans are retrieved in the big data network;

[0125] Importing each maintenance plan into the dynamic simulation model diagram to perform simulated maintenance, and obtaining predicted working characteristic data of the target micro-nano grating device after the simulated maintenance of each maintenance plan;

[0126] Calculate the Pearson correlation coefficient between the predicted working characteristic data and the preset working characteristic data of the target micro-nano grating device after the simulated maintenance of each maintenance plan;

[0127] A maintenance plan corresponding to the maximum Pearson correlation coefficient is obtained, and the maintenance plan corresponding to the maximum Pearson correlation coefficient is output as a recommended maintenance plan.

[0128] It should be noted that the structural engineering drawings contain key structural information such as the geometric shape, size, relative position, etc. of each component unit of the micro-nano grating device. The preset working parameter range information of each component unit in the target micro-nano grating device is obtained. These parameter ranges cover the possible parameter value ranges of each component unit under normal working conditions, such as electrical parameters (voltage, current range, etc.), optical parameters (refractive index, reflectivity range, etc.). Then, a mapping relationship between each component unit and the preset working parameter range information is established. This mapping relationship enables each component unit to be associated with the corresponding working parameter range. Through this mapping relationship, the preset working parameter range information of each component unit is imported into the three-dimensional structure diagram, so that the three-dimensional structure diagram contains not only structural information but also working parameter information, which lays the foundation for building a dynamic simulation model. If the transition probability of the target micro-nano grating device from the normal working state to the fault state at the current time node or a future time node is greater than the preset transition probability value, this indicates that the device may be about to fail or has already failed. At this time, it is necessary to obtain the fault type of the target micro-nano grating device from the normal working state to the fault state. The big data network contains a large amount of micro-nano grating device fault repair cases, experience data, and related technical documents. Each repair plan is imported into the dynamic simulation model diagram for simulated repair. During the simulation process, the micro-nano grating device in the dynamic simulation model diagram is adjusted accordingly according to the repair plan (such as changing the working parameters of a component unit, repairing a damaged structure, etc.), and then the predicted working characteristic data of the target micro-nano grating device after the simulated repair of each repair plan is obtained. These predicted working characteristic data reflect the possible working state of the device after adopting different repair plans. The Pearson correlation coefficient between the predicted working characteristic data and the preset working characteristic data of the target micro-nano grating device after the simulated repair of each repair plan is calculated. The Pearson correlation coefficient is a statistical indicator for measuring the linear correlation between two variables. It is used here to evaluate the similarity between the predicted working characteristic data and the preset working characteristic data. Judging the occurrence of faults based on the transition probability can warn of possible faults of micro-nano grating devices in advance, provide a basis for timely maintenance measures, improve the scientific nature of maintenance decisions, and make it more likely to select the most suitable maintenance plan, improve the maintenance efficiency and performance of micro-nano grating devices after repair.

[0129] The present invention also discloses a data analysis system for micro-nano grating devices for optical sensing, such as Figure 3 As shown, the micro-nano grating device data analysis system includes a memory 20 and a processor 60. The memory 20 stores a micro-nano grating device data analysis method program. When the micro-nano grating device data analysis method program is executed by the processor 60, any one of the steps of the micro-nano grating device data analysis method is implemented.

[0130] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for analyzing data of micro-nano grating devices for optical sensing, characterized in that: The following steps are involved: Acquire the actual sensing conditions of the target micro-nano grating device at several preset time nodes, and determine various preset spectral characteristic data sets of the target micro-nano grating device within a preset time period based on the actual sensing conditions and in combination with the spectral characteristic query database; Collecting actual spectral characteristic data of the target micro-nano grating device at several preset time nodes; Screening the actual spectral characteristic data to obtain various actual spectral characteristic data sets within a preset time period; Determine whether there is a baseline drift phenomenon in each actual spectral characteristic data of the target micro-nano grating device according to the preset spectral characteristic data set and the actual spectral characteristic data set within the preset time period; If one or more actual spectral characteristic data of the target micro-nano grating device have a baseline drift phenomenon, the actual spectral characteristic data with the baseline drift phenomenon is defined as baseline drift spectral characteristic data; a correlation spectral control unit query database is constructed, and a correlation spectral control unit of the baseline drift spectral characteristic data is determined according to the correlation spectral control unit query database; Acquire various real-time working parameters of the spectrum control unit of the correlation spectrum characteristic data of the baseline drift spectrum characteristic data, perform parameter analysis and processing on various real-time working parameters of the spectrum control unit of the correlation spectrum characteristic data of the baseline drift spectrum characteristic data, acquire abnormal working parameters, and perform adjustment processing on the abnormal working parameters.

2. The method for analyzing data of a micro-nano grating device for optical sensing according to claim 1, characterized in that: The actual sensing conditions of the target micro-nano grating device at several preset time nodes are obtained, and the preset spectral characteristic data sets of the target micro-nano grating device within a preset time period are determined based on the actual sensing conditions and in combination with the spectral characteristic query database, specifically: Obtain various preset spectral characteristic data of the target micro-nano grating device under various preset sensing conditions through the big data network; Construct a blank database, divide the blank database into several sub-databases, and import various preset spectral characteristic data of the target micro-nano grating device under various preset sensing conditions into corresponding sub-databases for storage, so as to obtain a spectral characteristic query database; Acquire the actual sensing conditions of the target micro-nano grating device at a number of preset time nodes, and import the actual sensing conditions of each preset time node into the spectral characteristic query database; Calculate the mutual information value between the actual sensing condition at each preset time node and the preset sensing condition in each sub-storage repository; and sort the mutual information values ​​between the actual sensing condition at each preset time node and the preset sensing condition in each sub-storage repository to obtain the sorting result; Obtain a maximum mutual information value according to the sorting result, obtain a sub-repository corresponding to the maximum mutual information value, and extract various preset spectral characteristic data of the target micro-nano grating device when it works at a corresponding preset time node from the sub-repository corresponding to the maximum mutual information value; And so on, until the preset spectral characteristic data of the target micro-nano grating device when it works at each preset time node are obtained, and the preset spectral characteristic data sets within the preset time period are generated according to the preset spectral characteristic data of the target micro-nano grating device when it works at each preset time node.

3. The method for analyzing data of a micro-nano grating device for optical sensing according to claim 1, characterized in that: The actual spectral characteristic data is screened to obtain various actual spectral characteristic data sets within a preset time period, specifically: Acquire characteristic data corresponding to various spectral noise data and characteristic data corresponding to various spectral characteristic data through a big data network; and acquire the number of data types of the spectral noise data and the number of data types of the spectral characteristic data; Constructing a binary tree model, and dividing a corresponding number of branches in the binary tree model according to the number of data types of the spectral noise data; According to the number of data types of the spectral characteristic data, a corresponding number of two-category branches are cut out in the binary tree model; Mapping the characteristic data corresponding to each type of spectral noise data to a type of branch corresponding to the binary tree model; and mapping the characteristic data corresponding to each type of spectral characteristic data to a type of branch corresponding to the binary tree model; Obtain the collected actual spectral characteristic data, and calculate the cosine similarity between each actual spectral characteristic data and each characteristic data in the binary tree model; Allocate each actual spectral characteristic data to the first or second branch with the largest cosine similarity; After completing the initial distribution of each actual spectral characteristic data, the contour coefficient of the actual spectral characteristic data in each first-class branch and second-class branch is calculated; Compare the contour coefficients of the actual spectral characteristic data in each of the first-class branches and the second-class branches with a preset coefficient threshold; If the silhouette coefficients of the actual spectral characteristic data in each of the first-class branches and the second-class branches are greater than the preset coefficient threshold, the initial allocation result is output as the final allocation result; If there is a situation where the profile coefficient of the actual spectral characteristic data in at least one of the first-class branches or the second-class branches is not greater than the preset coefficient threshold, the actual spectral characteristic data of the first-class branches or the second-class branches whose profile coefficient is not greater than the preset coefficient threshold are redistributed until the profile coefficients of the actual spectral characteristic data in each of the first-class branches and the second-class branches are greater than the preset coefficient threshold, and the last distribution result is output as the final distribution result; After the final allocation result is obtained, each of the first-class branches and the second-class branches of the binary tree model is pruned to obtain an actual spectral characteristic data set and a spectral noise data set.

4. The method for analyzing data of a micro-nano grating device for optical sensing according to claim 1, characterized in that: According to the preset spectral characteristic data set and the actual spectral characteristic data set within the preset time period, it is judged whether there is a baseline drift phenomenon in the actual spectral characteristic data of the target micro-nano grating device, specifically: Sorting each actual spectral characteristic data set based on the acquisition timestamp to obtain an actual spectral characteristic data set based on the time series; constructing a curve diagram of each actual spectral characteristic data change according to the actual spectral characteristic data set based on the time series; Acquire various preset spectral characteristic data sets of the target micro-nano grating device within a preset time period, and construct various preset spectral characteristic data change curves according to the various preset spectral characteristic data sets of the target micro-nano grating device within the preset time period; Performing a coincidence analysis on each actual spectral characteristic data change curve graph and the corresponding preset spectral characteristic data change curve graph to obtain the degree of coincidence between each actual spectral characteristic data change curve graph and the corresponding preset spectral characteristic data change curve graph; If the degree of overlap between a certain actual spectral characteristic data change curve and a corresponding preset spectral characteristic data change curve is not greater than a preset overlap threshold, it indicates that the actual spectral characteristic data in the micro-nano grating device has a baseline drift phenomenon; If the degree of overlap between a certain actual spectral characteristic data change curve and the corresponding preset spectral characteristic data change curve is greater than the preset overlap threshold, it means that there is no baseline drift phenomenon in the actual spectral characteristic data in the micro-nano grating device.

5. The method for analyzing data of a micro-nano grating device for optical sensing according to claim 1, characterized in that: Construct a correlation spectrum control unit query database, and determine the correlation spectrum control unit of the baseline drift spectrum characteristic data according to the correlation spectrum control unit query database, specifically: Obtaining a product structure specification of a target micro-nano grating device, determining each spectral control unit of the target micro-nano grating device according to the product structure specification, and obtaining control function characteristic information of each spectral control unit; According to the control function characteristic information of each spectral control unit, correlation analysis is performed on each spectral control unit and each spectral characteristic data, and spectral control units having correlation with each spectral characteristic data are obtained to obtain correlation spectral control units of each spectral characteristic data; Constructing a correlation spectrum regulation unit query database according to the correlation spectrum regulation unit of each spectrum characteristic data; The baseline drift spectral characteristic data in the target micro-nano grating device is obtained, and the baseline drift spectral characteristic data in the target micro-nano grating device is imported into the correlation spectrum control unit query database for query matching to obtain the correlation spectrum control unit of the baseline drift spectral characteristic data.

6. The method for analyzing data of a micro-nano grating device for optical sensing according to claim 1, characterized in that: Acquire various real-time working parameters of the correlation spectrum control unit of the baseline drift spectrum characteristic data, perform parameter analysis and processing on various real-time working parameters of the correlation spectrum control unit of the baseline drift spectrum characteristic data, acquire abnormal working parameters, and perform adjustment processing on the abnormal working parameters, specifically: Acquire various real-time working parameters of the spectrum control unit of the correlation spectrum characteristic data of the baseline drift, and calculate the difference between each real-time working parameter and the corresponding preset working parameter to obtain the parameter deviation value of each real-time working parameter; Preset the deviation value threshold of each real-time working parameter, and compare the parameter deviation value of each real-time working parameter with the corresponding deviation value threshold; If the parameter deviation value of a certain real-time working parameter is greater than the corresponding deviation value threshold, the real-time working parameter is defined as an abnormal working parameter; If the parameter deviation value of a certain real-time working parameter is not greater than the corresponding deviation value threshold, the real-time working parameter is defined as a normal working parameter; A parameter deviation value defined as an abnormal operating parameter is obtained, and the abnormal operating parameter is adjusted based on the corresponding parameter deviation value to adjust the corresponding real-time operating parameter to within a normal range.

7. The method for analyzing data of a micro-nano grating device for optical sensing according to claim 1, characterized in that: Sensing conditions include light intensity, dust concentration, temperature, humidity, and air pressure.

8. The method for analyzing data of a micro-nano grating device for optical sensing according to claim 1, characterized in that: The spectral characteristic data includes reflection spectrum data, transmission spectrum data, absorption spectrum data, Raman spectrum data, ellipsometry spectrum data and dispersion spectrum data.

9. The method for analyzing data of a micro-nano grating device for optical sensing according to claim 1, characterized in that: The real-time working parameters of the correlation spectrum control unit include the wavelength of the light source, the intensity of the light source, the polarization state of the light source, the incident angle, the bias voltage and the current.

10. A micro-nano grating device data analysis system for optical sensing, characterized in that: The micro-nano grating device data analysis system includes a memory and a processor, wherein the memory stores a micro-nano grating device data analysis method program. When the micro-nano grating device data analysis method program is executed by the processor, the micro-nano grating device data analysis method steps as described in any one of claims 1 to 9 are implemented.

Images