Illuminance sensor multi-point calibration method based on big data

Abstract

The invention discloses an illuminance sensor multi-point calibration method based on big data, particularly relates to the technical field of illuminance sensor calibration, and aims to screen out a faulty illuminance sensor based on the transmissivity, time null drift and zero temperature drift of the illuminance sensor, exclude subsequent calibration and effectively save the calibration cost of the illuminance sensor. The correlation coefficient set of the illuminance sensor is obtained by analyzing big data, and the illuminance sensor is calibrated based on the correlation coefficient set, so that the illuminance sensor is calibrated more scientifically, and the measurement accuracy and stability of the illuminance sensor are improved; more calibration times are set for the illuminance sensor with a large comprehensive quality coefficient, more resources can be applied to the sensor with better performance, and scientific management of the illuminance sensor is facilitated; the problems that in the prior art, the calibration cost of an illuminance sensor is high, and the illuminance sensor cannot be calibrated scientifically are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of light illumination sensor calibration, and more specifically, to a multi-point calibration method for a light illumination sensor based on big data. Background Art

[0002] As the use time increases, the performance of the electronic components inside the light sensor will gradually degrade, resulting in a decrease in the stability of the output signal; long-term exposure to certain specific environments (such as high temperature, high humidity, strong magnetic fields, etc.) may also accelerate the aging process of the components, thereby affecting the zero-point stability of the sensor; the components and materials inside the sensor will be affected by temperature changes, resulting in thermal expansion and contraction effects, leading to slight changes in the internal structure of the sensor, thereby affecting the zero-point position of the output signal.

[0003] However, it still has many disadvantages in actual use, such as the high calibration cost of the light sensor and the inability to scientifically calibrate the light sensor. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a multi-point calibration method of a light intensity sensor based on big data to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a multi-point calibration method of a light intensity sensor based on big data, comprising the following steps:

[0006] Step 1, collecting the transmittance, time zero drift average value and zero temperature drift average value of the light intensity sensor;

[0007] Step 2: When the transmittance, the time zero drift average value and the zero temperature drift average value meet the preset standard, proceed to the next step; otherwise, mark it as a faulty light sensor, do not proceed to the next step, and output the light sensor set to be calibrated;

[0008] Step 3: Based on the big data, a correlation coefficient set of the light sensor to be calibrated is established; the correlation coefficient set is obtained by: using statistical methods and machine learning algorithms to calculate the correlation coefficients of each feature dimension and the light sensor; each feature dimension and the corresponding correlation coefficient constitute the correlation coefficient set of the light sensor; the correlation coefficient set is processed by linear normalization to obtain the correlation coefficient set;

[0009] Step 4, calibrate the light sensor based on the correlation coefficient set, and output the calibrated light sensor; including: converting the nonlinear relationship into a linear relationship through data conversion variables, and outputting the characteristic dimensions of the converted variables; using a linear regression model as a calibration model, the linear regression model is expressed as y=β0+β1*x1+β2*x2+…+βs*xs; wherein y is the calibration value of the light value, x1, x2,…, xs are the characteristic dimensions of the converted variables, and β0, β1,…, βs are the parameters of the calibration model; using an optimization algorithm to determine the parameters of the calibration model until the difference between the predicted value and the standard value meets the requirements, completing the calibration of the light sensor, and outputting the calibrated light sensor.

[0010] Preferably, in step one, a transmittance measuring instrument is used to test the transmittance of the optical element of the light sensor; transmittance refers to the ratio of the transmitted light intensity to the incident light intensity, expressed as a percentage; the time zero drift and zero temperature drift of the light sensor are measured, and the time zero drift average value and the zero temperature drift average value of the light sensor are calculated; in step two, when the transmittance, time zero drift, and zero temperature drift of the light sensor all meet the preset thresholds or requirements, the light sensor is considered to be calibrable and enters the next step of the calibration process; otherwise, it is marked as a faulty light sensor and excluded from subsequent calibration.

[0011] Preferably, based on the transmittance, the time zero drift average value and the zero temperature drift average value, the comprehensive quality coefficient of the light intensity sensor is analyzed and obtained, and corresponding measures are taken based on the comprehensive quality coefficient, including the following steps:

[0012] Step S11: When the transmittance, time zero drift and zero temperature drift of the light sensor meet the preset threshold or requirement, the light sensor is marked as the light sensor to be calibrated, the light sensor set to be calibrated is output, and the transmittance, time zero drift and zero temperature drift average values ​​of the light sensor to be calibrated are recorded;

[0013] Step S12, calculating a zero temperature drift quality index based on the time zero drift average value and the zero temperature drift average value of the light intensity sensor;

[0014] Step S13, under a standard environment, evenly set different gradient light intensities, record the error ratios of the light sensor facing different gradient light intensities, and calculate the average value of the error ratios;

[0015] Step S14, jointly analyzing the error ratio average value, transmittance, time zero drift average value and zero temperature drift average value to obtain a comprehensive quality coefficient of the light intensity sensor; taking corresponding measures based on the comprehensive quality coefficient, including:

[0016] Allocate correction resources, wherein the correction resources include: a calibration times threshold; that is, a light sensor with a large comprehensive quality coefficient has more calibration times, and the more calibration times, the better the calibration effect of the light sensor; a light sensor with a low comprehensive quality coefficient may have limited room for performance improvement due to hardware aging, damage or design defects, so a smaller number of calibration times is allocated to use more resources on light sensors with better performance;

[0017] Sort the light sensors to be calibrated from high to low according to the comprehensive quality coefficient, and give priority to calibrating the light sensors with high comprehensive quality coefficient values; or divide the set of light sensors to be calibrated into several levels, and give priority to calibrating the levels with good quality; for example, the higher the comprehensive quality coefficient, the better the quality of the light sensor, and the light sensors with good quality are calibrated first.

[0018] Preferably, the comprehensive quality coefficient is obtained in the following manner:

[0019] The error ratio average, transmittance, time zero drift average and zero temperature drift average are recorded as Er, Tm, TZDa, TZDb respectively;

[0020] By formula The zero temperature drift quality index ZQI is calculated, where TZDa T ,TZDb T Respectively represent the thresholds of time zero drift and zero temperature drift;

[0021] By formula The comprehensive quality coefficient Zp is calculated, where α1, α2, and α3 are positive real number exponents set according to the importance of each parameter on the comprehensive quality.

[0022] Preferably, before constructing the correlation coefficient set, data preprocessing is performed on the illumination sensor measurement data and the environmental characteristic parameters, including:

[0023] Data cleaning and denoising to ensure data accuracy and reliability;

[0024] Feature dimension parameter selection: Feature dimension parameters include environmental feature parameters (such as temperature, humidity, air pressure, etc.) and spatiotemporal feature parameters (such as time, geographic location, etc.);

[0025] Correlation analysis: Statistical methods (such as Pearson correlation coefficient, Spearman rank correlation coefficient) and machine learning algorithms (such as random forest, gradient boosting tree, etc.) are used to accurately calculate the correlation coefficients of each dimension and the light sensor to improve the accuracy and robustness of the correlation coefficient set;

[0026] Dynamic adjustment: The construction of the correlation coefficient set should support dynamic updating. As new data is added, the correlation coefficient is recalculated and the correlation coefficient set is adjusted to adapt to factors such as environmental changes and equipment aging.

[0027] Preferably, the multi-point calibration method for the light intensity sensor further includes:

[0028] Step 5: Based on the correlation coefficient set or the updated correlation coefficient set, a verification environment is built, the light sensor is placed in the verification environment with a known standard light value, and a calibration value corresponding to the light value is output;

[0029] Step 6: Compare the difference between the calibration value of the illuminance value of the illuminance sensor in the verification environment and the standard value, and take measures based on the relationship between the difference and the threshold.

[0030] Preferably, in step 5, the building of the verification environment includes:

[0031] The number of gradients of each feature dimension is set based on the correlation coefficient set, fluctuates up and down with the standard environment as the center, and the upper and lower limits of each feature dimension of the light sensor are used as constraints to output a number of gradient data of each feature dimension;

[0032] Based on a number of gradient data of each feature dimension, a number of groups of verification environment feature parameters are obtained by arrangement and combination, the verification environment feature parameters are output, and the verification environment is built based on the verification environment feature parameters.

[0033] Preferably, the calibration quality index JQI is used to represent the gap, and it is determined whether the calibration quality index exceeds the corresponding calibration quality index threshold; if so, it indicates that the calibration is successful; otherwise, it is determined whether the number of calibrations reaches the calibration number threshold, if not, recalibrate, and obtain a set of correlation coefficients based on new data analysis, recorded as updating the set of correlation coefficients, until the calibration number threshold is reached; if the calibration number threshold is reached but still does not meet the requirements, it is marked as a faulty light illumination sensor, and the calibration is terminated;

[0034] By formula Among them, yi is the standard value of the illuminance value, y is the calibration value of the illuminance value, n represents the number of verifications, and i represents the sequential number of verifications.

[0035] Preferably, the multi-point calibration method for the illuminance sensor further comprises a calibration quality management step, wherein the calibration quality management step is used to jointly analyze and obtain an average value of the calibration quality index, a calibration consumption time fluctuation coefficient and a calibration success rate, obtain a calibration operation reliability coefficient, and take measures based on the calibration operation reliability coefficient;

[0036] The average value of the calibration quality index and the fluctuation coefficient of the calibration time σ T and calibration success rate Jc, denoted as JQIa, σT , Jc;

[0037] By formula The calibration operation reliability coefficient RJ is calculated, where ∈ is a very small positive number to prevent the denominator from being zero; ∈ is adjusted according to actual conditions; if RJ is high, it means that the calibration operation reliability is good and only routine maintenance and monitoring are required; if RJ is low, it means that there are problems with the calibration operation and the calibration process should be optimized, more accurate calibration equipment should be replaced, and operators should be trained.

[0038] Technical effects and advantages of the present invention:

[0039] (1) The multi-point calibration method for a light sensor provided by the present invention can screen out faulty light sensors based on the transmittance, time zero drift, and zero temperature drift of the light sensor, and exclude them from subsequent calibration, thereby effectively saving the calibration cost of the light sensor; by analyzing big data, a set of correlation coefficients of the light sensor is obtained; the light sensor is calibrated based on the set of correlation coefficients, and the light sensor is calibrated more scientifically, thereby improving the measurement accuracy and stability of the light sensor, and ensuring that it can provide reliable data in different environments and different lighting conditions; more calibration times are set for light sensors with a large comprehensive quality coefficient, so that more resources can be used for light sensors with better performance, which is conducive to the scientific management of light sensors; the problem of high calibration cost and inability to scientifically calibrate light sensors in the prior art is solved.

[0040] (2) The multi-point calibration method for the light sensor provided by the present invention builds a verification environment based on a set of correlation coefficients of the light sensor to be calibrated; it can effectively ensure that the abnormal light sensor is effectively calibrated, and solve the problem that the calibration quality verification in the prior art is not scientific enough. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a flow chart of the multi-point calibration method of the light intensity sensor of the present invention.

[0042] Figure 2 The present invention is a flow chart of a multi-point calibration method for an illuminance sensor based on a comprehensive quality coefficient.

[0043] Figure 3 This is a flow chart of calibration quality verification of the present invention. DETAILED DESCRIPTION

[0044] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0045] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0046] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present application, its application, or uses.

[0047] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0048] Background technology: Time zero drift refers to the situation where the output zero point of the light intensity sensor drifts over time, that is, when there is no light input (i.e., the ideal zero point state), the output signal of the sensor may also shift over time; zero temperature drift refers to the situation where the output zero point of the light intensity sensor drifts with temperature changes, that is, when the ambient temperature changes, the output signal of the sensor will also shift at the zero point position. Time zero drift will cause the accuracy of the sensor's measurement results to gradually decrease during long-term use, thereby affecting the overall performance of the system; zero temperature drift will cause deviations in the sensor's measurement results at different temperatures, affecting the measurement accuracy and reliability of the system. Abnormalities in time zero drift and zero temperature drift are often caused by sensor design defects, problems in the manufacturing process, or irreversible damage caused by long-term use, and it is difficult to restore performance through calibration.

[0049] Example 1, see Figure 1 The present invention provides a flowchart of a multi-point calibration method for an illuminance sensor. Figure 1 A multi-point calibration method for an illuminance sensor based on big data is shown, characterized in that it includes:

[0050] Step 1, collecting the transmittance, time zero drift average value and zero temperature drift average value of the light intensity sensor;

[0051] Step 2: When the transmittance, the time zero drift average value and the zero temperature drift average value meet the preset standard, proceed to the next step; otherwise, mark it as a faulty light sensor, do not proceed to the next step, and output the light sensor set to be calibrated;

[0052] It should be explained in the embodiments of the present invention that a transmittance measuring instrument (such as a spectrophotometer) is used to test the transmittance of the optical element of the light sensor; transmittance refers to the ratio of the intensity of transmitted light to the intensity of incident light, expressed as a percentage, and high transmittance means high transparency; the time zero drift and zero temperature drift of the light sensor are measured, and the time zero drift average value and the zero temperature drift average value of the light sensor are calculated; it is explained that when the transmittance, time zero drift, and zero temperature drift of the light sensor meet the preset thresholds or requirements, the light sensor is considered to be calibrable and enters the next step of the calibration process; otherwise, it is marked as a faulty light sensor and excluded from subsequent calibration;

[0053] Step 3: Based on the big data, a correlation coefficient set of the light sensor to be calibrated is established; the correlation coefficient set is obtained by: using statistical methods and machine learning algorithms to calculate the correlation coefficients of each feature dimension and the light sensor; each feature dimension and the corresponding correlation coefficient constitute the correlation coefficient set of the light sensor; the correlation coefficient set is processed by linear normalization to obtain the correlation coefficient set;

[0054] Explanation: The statistical method used may be Pearson correlation coefficient or Spearman rank correlation coefficient; the machine learning algorithm may be decision tree, random forest, gradient boosting machine, neural network; the feature dimensions may include environmental feature parameters and spatiotemporal feature parameters; linear normalization is used to convert the value of the correlation coefficient in the correlation coefficient set to the range of 0 to 1.

[0055] Step 4: calibrate the light sensor based on the correlation coefficient set and output the calibrated light sensor; including: converting the nonlinear relationship into a linear relationship through data conversion variables (such as logarithmic conversion, square root conversion, polynomial conversion, etc.), and outputting the characteristic dimensions of the converted variables; using a linear regression model as a calibration model, the linear regression model is expressed as y=β0+β1*x1+β2*x2+…+βs*xs; wherein y is the calibration value of the light value, x1, x2,…, xs are the characteristic dimensions of the converted variables (such as temperature, humidity, etc.), and β0, β1,…, βs are the parameters of the calibration model; using an optimization algorithm to determine the parameters of the calibration model (such as the β parameter in the linear regression model) until the difference between the predicted value and the standard value meets the requirements, completing the calibration of the light sensor, and outputting the calibrated light sensor; the standard value refers to the standard value of the light, that is, the true value.

[0056] Explanation: The correlation coefficient can be linear or nonlinear. If there is a nonlinear relationship between the characteristic dimension parameter and the measured value, a polynomial regression model or a neural network model is selected as the calibration model. The optimization algorithm includes the least squares method and the gradient descent method:

[0057] Least squares method: Use the linear regression model as the calibration model, select the least squares method as the optimization algorithm, and find the optimal parameters by minimizing the residual sum of squares; Gradient descent method: For more complex models, such as neural network models, the gradient descent method (or its variants, such as stochastic gradient descent, batch gradient descent, Adam, etc.) minimizes the loss function by iteratively adjusting the model parameters; The loss function in the embodiment of the present invention refers to the difference between the predicted value and the standard value.

[0058] In the embodiment of the present invention, it is further designed that, refer to Figure 2 The flowchart of the multi-point calibration method of the illuminance sensor based on the comprehensive quality coefficient is as follows: based on the transmittance, the time zero drift average value and the zero temperature drift average value, the comprehensive quality coefficient of the illuminance sensor is obtained by analysis, and corresponding measures are taken based on the comprehensive quality coefficient, including the following steps:

[0059] Step S11: When the transmittance, time zero drift and zero temperature drift of the light sensor meet the preset threshold or requirement, the light sensor is marked as the light sensor to be calibrated, the light sensor set to be calibrated is output, and the transmittance, time zero drift and zero temperature drift average values ​​of the light sensor to be calibrated are recorded;

[0060] Step S12, calculating a zero temperature drift quality index based on the time zero drift average value and the zero temperature drift average value of the light intensity sensor;

[0061] Step S13, under a standard environment, evenly set different gradient light intensities, record the error ratios of the light sensor facing different gradient light intensities, and calculate the average value of the error ratios;

[0062] Step S14, jointly analyzing the error ratio average value, transmittance, time zero drift average value and zero temperature drift average value to obtain a comprehensive quality coefficient of the light intensity sensor; taking corresponding measures based on the comprehensive quality coefficient, including:

[0063] Allocate correction resources, wherein the correction resources include: a calibration times threshold; that is, a light sensor with a large comprehensive quality coefficient has more calibration times, and the more calibration times, the better the calibration effect of the light sensor; a light sensor with a low comprehensive quality coefficient may have limited room for performance improvement due to hardware aging, damage or design defects, so a smaller number of calibration times is allocated to use more resources on light sensors with better performance;

[0064] Sort the light sensors to be calibrated from high to low according to the comprehensive quality coefficient, and give priority to calibrating the light sensors with high comprehensive quality coefficient values; or divide the set of light sensors to be calibrated into several levels, and give priority to calibrating the levels with good quality; for example, the higher the comprehensive quality coefficient, the better the quality of the light sensor, and the light sensors with good quality are calibrated first.

[0065] In the embodiment of the present invention, it is further designed that the comprehensive quality coefficient is obtained in the following manner:

[0066] The error ratio average, transmittance, time zero drift average and zero temperature drift average are recorded as Er, Tm, TZDa, TZDb respectively;

[0067] By formula The zero temperature drift quality index ZQI is calculated, where TZDa T ,TZDb T Respectively represent the thresholds of time zero drift and zero temperature drift;

[0068] By formula The comprehensive quality coefficient Zp is calculated, where α1, α2, and α3 are positive real number exponents set according to the importance of each parameter on the comprehensive quality.

[0069] It should be explained in the embodiment of the present invention that before constructing the correlation coefficient set, data preprocessing is performed on the illumination sensor measurement data and the environmental characteristic parameters, including:

[0070] Data cleaning and denoising to ensure data accuracy and reliability;

[0071] Feature dimension parameter selection: Feature dimension parameters include environmental feature parameters (such as temperature, humidity, air pressure, etc.) and spatiotemporal feature parameters (such as time, geographic location, etc.);

[0072] Correlation analysis: Statistical methods (such as Pearson correlation coefficient, Spearman rank correlation coefficient) and machine learning algorithms (such as random forest, gradient boosting tree, etc.) are used to accurately calculate the correlation coefficients of each dimension and the light sensor to improve the accuracy and robustness of the correlation coefficient set;

[0073] Dynamic adjustment: The construction of the correlation coefficient set should support dynamic updating. As new data is added, the correlation coefficient is recalculated and the correlation coefficient set is adjusted to adapt to factors such as environmental changes and equipment aging.

[0074] Summary: Based on the transmittance, time zero drift and zero temperature drift of the illuminance sensor, the faulty illuminance sensor can be screened out and excluded from the subsequent calibration; it can effectively save the calibration cost of the illuminance sensor; set the calibration number threshold based on the comprehensive quality coefficient; obtain the correlation coefficient set of the illuminance sensor by analyzing big data; calibrate the illuminance sensor based on the correlation coefficient set, calibrate the illuminance sensor more scientifically, improve the measurement accuracy and stability of the illuminance sensor, and ensure that it can provide reliable data in different environments and different lighting conditions; set more calibration times for illuminance sensors with large comprehensive quality coefficients, so that more resources can be used on illuminance sensors with better performance, which is conducive to the scientific management of illuminance sensors.

[0075] Example 2, see Figure 3 The calibration quality verification flow chart of the embodiment of the present invention is different from embodiment 1 in that the multi-point calibration method of the light intensity sensor of the present invention further includes:

[0076] Step 5: Based on the correlation coefficient set or the updated correlation coefficient set, a verification environment is built, the light sensor is placed in the verification environment with a known standard light value, and a calibration value corresponding to the light value is output;

[0077] Step 6: Compare the difference between the calibration value of the illuminance value of the illuminance sensor in the verification environment and the standard value, and take measures based on the relationship between the difference and the corresponding threshold value.

[0078] It should be explained in the embodiment of the present invention that the establishment of the verification environment includes:

[0079] The number of gradients of each feature dimension is set based on the correlation coefficient set, fluctuates up and down with the standard environment as the center, and the upper and lower limits of each feature dimension of the light sensor are used as constraints to output a number of gradient data of each feature dimension;

[0080] When setting the number of gradients for each feature dimension, the gradient density should be allocated based on the size of the correlation coefficient, that is, more gradient points should be set for dimensions with high correlation to examine their impact in more detail; at the same time, the feasibility and cost-effectiveness of actual operations should be considered;

[0081] Based on a number of gradient data of each feature dimension, a number of groups of verification environment feature parameters are obtained by arrangement and combination, the verification environment feature parameters are output, and the verification environment is built based on the verification environment feature parameters.

[0082] What needs to be explained in the embodiments of the present invention is that, during the verification process, high-precision environmental control equipment (such as a constant temperature and humidity chamber, a light simulator, etc.) is used to accurately simulate the changes in light intensity under different environmental conditions, ensure the accuracy and repeatability of the verification environment, realize automatic control and data collection of the verification environment, reduce human errors, and improve verification efficiency and accuracy.

[0083] In the embodiment of the present invention, it is further designed that the calibration quality index JQI is used to represent the gap, and it is determined whether the calibration quality index exceeds the corresponding calibration quality index threshold; if so, it indicates that the calibration is successful; otherwise, it is determined whether the number of calibrations reaches the calibration number threshold, if not, recalibrate, and obtain a set of correlation coefficients based on new data analysis, which is recorded as updating the set of correlation coefficients, until the calibration number threshold is reached; if the calibration number threshold is reached but still does not meet the requirements, it is marked as a faulty light illumination sensor and the calibration is terminated;

[0084] By formula Among them, yi is the standard value of the illuminance value, y is the calibration value of the illuminance value, n represents the number of verifications, and i represents the sequential number of verifications.

[0085] In the embodiment of the present invention, it is further designed that the multi-point calibration method of the light intensity sensor also includes a calibration quality management step, the calibration quality management step is used to jointly analyze and obtain the calibration quality index average value, the calibration consumption time fluctuation coefficient and the calibration success rate, obtain the calibration operation reliability coefficient, and take measures based on the calibration operation reliability coefficient;

[0086] The average value of the calibration quality index and the fluctuation coefficient of the calibration time σ T and calibration success rate Jc, denoted as JQIa, σ T , Jc;

[0087] By formula The calibration operation reliability coefficient RJ is calculated, where ∈ is a very small positive number to prevent the denominator from being zero; ∈ is adjusted according to actual conditions; if RJ is high, it means that the calibration operation reliability is good and only routine maintenance and monitoring are required; if RJ is low, it means that there are problems with the calibration operation and the calibration process should be optimized, more accurate calibration equipment should be replaced, and operators should be trained.

[0088] Summary: In the prior art, the method of verifying the calibration quality is often not scientific enough, and the verification is based on the user's subjective consciousness and fixed processes; resulting in the inability to ensure that the abnormal light illumination sensor is effectively calibrated; the embodiment of the present invention provides a verification method after calibration, and builds a verification environment based on a set of correlation coefficients of the light illumination sensor to be calibrated; it can effectively ensure that the abnormal light illumination sensor is effectively calibrated.

[0089] Finally: 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 protection scope of the present invention.

Claims

1. A multi-point calibration method for an illuminance sensor based on big data, characterized in that: include: Step 1, collecting the transmittance, time zero drift average value and zero temperature drift average value of the light intensity sensor; Step 2: When the transmittance, the time zero drift average value and the zero temperature drift average value meet the preset standard, proceed to the next step; otherwise, mark it as a faulty light sensor, do not proceed to the next step, and output the light sensor set to be calibrated; Step 3: Establish a correlation coefficient set of the light sensor to be calibrated based on big data; the correlation coefficient set is obtained by: using statistical methods and machine learning algorithms to calculate the correlation coefficients of each feature dimension and the light sensor; each feature dimension and the corresponding correlation coefficient constitute the correlation coefficient set of the light sensor; and the correlation coefficient set is processed by linear normalization to obtain the correlation coefficient set; Step 4, calibrate the light sensor based on the correlation coefficient set, and output the calibrated light sensor; including: converting the nonlinear relationship into a linear relationship through data conversion variables, and outputting the characteristic dimensions of the converted variables; using a linear regression model as a calibration model, the linear regression model is expressed as y=β0+β1*x1+β2*x2+…+βs*xs; wherein y is the calibration value of the light value, x1, x2,…, xs are the characteristic dimensions of the converted variables, and β0, β1,…, βs are the parameters of the calibration model; using an optimization algorithm to determine the parameters of the calibration model until the difference between the predicted value and the standard value meets the requirements, completing the calibration of the light sensor, and outputting the calibrated light sensor.

2. The multi-point calibration method of an illuminance sensor based on big data according to claim 1, characterized in that: Use a transmittance measuring instrument to test the transmittance of the optical element of the light sensor; transmittance refers to the ratio of the transmitted light intensity to the incident light intensity, expressed as a percentage; measure the time zero drift and zero temperature drift of the light sensor; calculate the average time zero drift and zero temperature drift of the light sensor; when the transmittance, time zero drift, and zero temperature drift of the light sensor meet the preset thresholds or requirements, the light sensor is considered calibrable and enters the next calibration process; otherwise, it is marked as a faulty light sensor and excluded from subsequent calibration.

3. The multi-point calibration method of an illuminance sensor based on big data according to claim 1, characterized in that: Based on the transmittance, the time zero drift average value and the zero temperature drift average value, the comprehensive quality coefficient of the light intensity sensor is analyzed and obtained, and corresponding measures are taken based on the comprehensive quality coefficient, including the following steps: Step S11: When the transmittance, time zero drift and zero temperature drift of the light sensor meet the preset threshold or requirement, the light sensor is marked as the light sensor to be calibrated, the light sensor set to be calibrated is output, and the transmittance, time zero drift and zero temperature drift average values ​​of the light sensor to be calibrated are recorded; Step S12, calculating a zero temperature drift quality index based on the time zero drift average value and the zero temperature drift average value of the light intensity sensor; Step S13, under a standard environment, evenly set different gradient light intensities, record the error ratios of the light sensor facing different gradient light intensities, and calculate the average value of the error ratios; Step S14, jointly analyzing the error ratio average value, transmittance, time zero drift average value and zero temperature drift average value to obtain a comprehensive quality coefficient of the light intensity sensor; Take corresponding measures based on the comprehensive quality coefficient, including: Allocate correction resources, the correction resources include: a calibration times threshold Tha; that is, light sensors with a large comprehensive quality coefficient have more calibration times; sort the light sensors to be calibrated from high to low according to the comprehensive quality coefficient, and give priority to calibrating light sensors with high comprehensive quality coefficient values.

4. The multi-point calibration method of an illuminance sensor based on big data according to claim 3 is characterized in that: The comprehensive quality coefficient is obtained as follows: The error ratio average, transmittance, time zero drift average and zero temperature drift average are recorded as Er, Tm, TZDa, TZDb respectively; By formula The zero temperature drift quality index ZQI is calculated, where TZDa T ,TZDb T Respectively represent the thresholds of time zero drift and zero temperature drift; By formula The comprehensive quality coefficient Zp is calculated, where α1, α2, and α3 are positive real number exponents set according to the importance of each parameter on the comprehensive quality.

5. The multi-point calibration method of an illuminance sensor based on big data according to claim 1, characterized in that: The multi-point calibration method of the light intensity sensor also includes: Step 5: Based on the correlation coefficient set or the updated correlation coefficient set, a verification environment is built, the light sensor is placed in the verification environment with a known standard light value, and a calibration value corresponding to the light value is output; Step 6: Compare the difference between the calibration value of the illuminance value of the illuminance sensor in the verification environment and the standard value, and take measures based on the relationship between the difference and the threshold.

6. The multi-point calibration method of an illuminance sensor based on big data according to claim 5, characterized in that: The verification environment construction includes: The number of gradients of each feature dimension is set based on the correlation coefficient set, fluctuates up and down with the standard environment as the center, and the upper and lower limits of each feature dimension of the light sensor are used as constraints to output a number of gradient data of each feature dimension; Based on a number of gradient data of each feature dimension, a number of groups of verification environment feature parameters are obtained by arrangement and combination, the verification environment feature parameters are output, and the verification environment is built based on the verification environment feature parameters.

7. The multi-point calibration method of an illuminance sensor based on big data according to claim 6, characterized in that: The calibration quality index JQI is used to represent the gap, and it is determined whether the calibration quality index exceeds the corresponding calibration quality index threshold; if so, it indicates that the calibration is successful; otherwise, it is determined whether the calibration times have reached the calibration times threshold. If it has not reached the calibration times threshold, recalibrate, and obtain a set of correlation coefficients based on new data analysis, which is recorded as the updated correlation coefficient set, until the calibration times threshold is reached; if the calibration times threshold is reached but still does not meet the requirements, it is marked as a faulty light sensor and the calibration is terminated; By formula Among them, yi is the standard value of the illuminance value, y is the calibration value of the illuminance value, n represents the number of verifications, and i represents the sequential number of verifications.

8. The multi-point calibration method of an illuminance sensor based on big data according to claim 6, characterized in that: The multi-point calibration method for the illuminance sensor further comprises a calibration quality management step, wherein the calibration quality management step is used to jointly analyze and obtain a calibration quality index average value, a calibration consumption time fluctuation coefficient and a calibration success rate, obtain a calibration operation reliability coefficient, and take measures based on the calibration operation reliability coefficient; The average value of the calibration quality index and the fluctuation coefficient of the calibration time σ T and calibration success rate Jc, denoted as JQIa, σ T , Jc; By formula The calibration operation reliability coefficient RJ is calculated, where ∈ is a small positive number to prevent the denominator from being zero; ∈ is adjusted according to the actual situation; If RJ is high, it indicates that the calibration operation is reliable and only routine maintenance and monitoring are required. If RJ is low, it indicates that there are problems with the calibration operation and the calibration process needs to be optimized, more accurate calibration equipment needs to be replaced, and operators need to be trained.

Images