A tea moisture determination error compensation method and system

Abstract

This invention relates to the field of tea detection technology and discloses a method and system for compensating for errors in tea moisture determination. The system includes a data acquisition and preprocessing module, an intelligent compensation and scheduling module, a transmission control and adjustment module, and a receiving verification and evaluation module. It acquires real-time moisture content data of tea samples using a moisture sensor and generates a raw moisture dataset. Then, it performs data preprocessing operations, including noise reduction, calibration, and format standardization, to generate preprocessed moisture data. Finally, it performs adaptive error compensation based on environmental parameters to generate compensated moisture data, eliminating random fluctuations in the sensor and environmental interference, ensuring the initial accuracy and environmental adaptability of the moisture determination data, and reducing initial errors.

Description

Technical Field

[0001] This invention relates to the field of tea testing technology, specifically to a method and system for compensating for errors in tea moisture determination. Background Technology

[0002] Tea leaves, commonly known as tea, generally include the leaves and buds of the tea plant. The moisture content of tea leaves is an important indicator of tea quality. Under the same storage conditions, the higher the moisture content, the faster the quality deteriorates. Therefore, tea moisture testing has become one of the essential basic testing items for tea quality control and import / export customs.

[0003] Currently, because the moisture content of tea is easily affected by environmental factors and sensor fluctuations during the process of measuring moisture content in tea samples, it is impossible to adaptively compensate for errors in real time based on environmental parameters when collecting moisture content data of tea samples through moisture measurement sensors. If there are fluctuations in temperature, humidity or air pressure, it may cause a large deviation between the measured moisture value and the true value, affecting the accuracy of moisture detection.

[0004] Therefore, a method for compensating for the error in tea moisture determination and a system for solving the above problems are proposed. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method and system for compensating for errors in tea moisture measurement, which solves the problem mentioned in the background art of large deviations between measured moisture values ​​and true values, affecting the accuracy of moisture detection.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method and system for compensating for errors in tea moisture measurement, the method comprising the following steps:

[0007] S1. Collect moisture content data of tea samples in real time using a moisture sensor and generate a raw moisture dataset.

[0008] S2. Perform data preprocessing operations based on the original moisture dataset, including noise reduction, calibration and format normalization, and generate preprocessed moisture data;

[0009] S3. Perform adaptive error compensation processing on the preprocessed moisture data based on environmental parameters to generate compensated moisture data.

[0010] S4. Using an intelligent scheduling algorithm, the compensated moisture data is arranged into a measurement sequence according to the priority of tea varieties and the importance of the measurement task, thereby generating measurement sequence data;

[0011] S5. The measured sequence data is sent to the central processing station through the data transmission module, and the data error rate is monitored in real time during the transmission process;

[0012] S6. Based on the monitoring results, perform dynamic compensation and adjustment processing, including retesting mechanism and adaptive adjustment of compensation parameters, and generate adjusted moisture data;

[0013] S7. The adjusted moisture data is verified and decoded at the receiving end to restore the accurate moisture content data.

[0014] S8. Combine the artificial intelligence model to perform a quality assessment on the accurate moisture content data and generate a moisture measurement quality report.

[0015] Preferably, the step S1 of collecting moisture content data from tea samples includes the following steps:

[0016] S11. Deploy multimodal moisture sensors, including near-infrared spectrometers and capacitive moisture meters, in tea processing production lines and storage areas to collect raw data on the moisture content of tea samples and record the sample type.

[0017] S12. Install temperature and humidity sensors and barometers in the measurement environment to collect environmental parameter data in real time, including temperature, humidity and atmospheric pressure values.

[0018] S13. Mark the collected moisture data and environmental parameters using high-precision timestamps and unique sample identifiers;

[0019] S14. Integrate the labeled data into a structured dataset to generate the original moisture dataset, which includes CSV, JSON and binary formats to support subsequent processing and analysis.

[0020] Preferably, the data preprocessing operation in S2 includes the following steps:

[0021] S21. The sliding window averaging filter algorithm is used to perform noise reduction processing on the original moisture dataset to reduce interference caused by random fluctuations in the sensor.

[0022] S22. Correct sensor baseline drift using linear regression calibration technology and perform accuracy calibration by combining standard sample reference values;

[0023] S23. Convert the calibrated data into a standard JSON format to ensure compatibility with different measuring devices and systems;

[0024] S24. Perform range standardization on the converted data, normalizing all moisture content values ​​to the relative humidity range of 0-100%, and generating pre-processed moisture data.

[0025] Preferably, the adaptive error compensation process in S3 includes the following steps:

[0026] S31. Monitor changes in environmental parameters in real time, including temperature gradient, humidity fluctuation and air pressure difference, and generate environmental status indicators.

[0027] S32. Based on the environmental state indicators, dynamically select the compensation model: when the environmental fluctuation is below the threshold, use a linear compensation algorithm; when the environmental fluctuation is above the threshold, use a nonlinear compensation algorithm.

[0028] S33. Apply multiple regression analysis to model the error of moisture data and calculate the compensation coefficient of environmental factors on the measured values.

[0029] S34. The compensation parameters are adjusted through a real-time feedback mechanism to generate compensated moisture data, the accuracy of which is adaptive to environmental conditions.

[0030] Preferably, determining the sequence arrangement in step S4 includes the following steps:

[0031] S41. Define tea variety priority parameters, including high-value varieties, conventional varieties, and experimental varieties, and assign different weights to them;

[0032] S42. Use convolutional neural networks to analyze moisture data trends, identify key events: abnormal moisture fluctuations, measure failure risks, and assign importance scores.

[0033] S43. Combining priority parameters and importance scores, a weighted round-robin algorithm is used to arrange the order of moisture determination;

[0034] S44. Allocate more processing resources to high-priority determination tasks to generate determination sequence data.

[0035] Preferably, the real-time monitoring of the data error rate in S5 includes the following steps:

[0036] S51. Divide the sequence data into fixed-size data packets, and add a sequence number, timestamp, and CRC checksum to each data packet;

[0037] S52. Data transmission is performed via the LoRa wireless communication protocol, and the ACK confirmation mechanism is enabled to monitor the data packet reception status.

[0038] S53. Calculate the data error rate, compare the difference between the sent and received data based on the feedback from the receiving end, and trigger an alarm when the error rate exceeds 3%.

[0039] S54. Record transmission timestamps, environmental indicators, and device status to generate transmission monitoring logs.

[0040] Preferably, the dynamic compensation adjustment process in S6 includes the following steps:

[0041] S61. When the data error rate exceeds the threshold, a selective retest request is automatically initiated, and only the error data points are retested.

[0042] S62. Adjust the parameter redundancy of the compensation algorithm and dynamically adjust the filter coefficient according to the environmental noise level;

[0043] S63. Use the Kalman filter algorithm to smooth the retest data and reduce random errors;

[0044] S64. Verify the consistency and stability of the adjusted data and generate adjusted moisture data.

[0045] Preferably, the verification and decoding process in S7 includes the following steps:

[0046] S71. At the receiving end, the data packets are reassembled in order, and the original measurement sequence is recovered based on the sequence number;

[0047] S72. Apply the inverse compensation algorithm to decode the moisture data and restore it to the physical moisture content value;

[0048] S73. Perform data smoothing and outlier removal to output accurate moisture content data.

[0049] Preferably, generating a moisture content quality report in step S8 includes the following steps:

[0050] S81. Use a pre-trained deep learning model LSTM network to evaluate the quality of moisture data, including standard deviation, confidence interval and bias.

[0051] S82. Compare the differences in moisture data before and after compensation, and calculate the absolute error and relative error scores;

[0052] S83. Generate a moisture content determination quality report, including the quality grade, error analysis results, and improvement suggestions.

[0053] Preferably, the system includes a data acquisition and preprocessing module, an intelligent compensation and scheduling module, a transmission control and adjustment module, and a receiving verification and evaluation module;

[0054] The data acquisition and preprocessing module acquires moisture data and environmental parameters of tea samples through a multi-sensor acquisition unit, optimizes the data through a noise elimination and calibration unit, and outputs the preprocessed moisture data through a format standardization unit.

[0055] The intelligent compensation and scheduling module receives the preprocessed moisture data, obtains real-time parameters through the environmental state sensing unit, performs error processing using the adaptive compensation unit, and outputs the measurement sequence data through the priority scheduling unit.

[0056] The transmission control and adjustment module receives the measurement sequence data, performs a transmission operation through the wireless data transmission unit, collects data error information using the error monitoring unit, and outputs adjusted moisture data through the dynamic adjustment unit.

[0057] The receiving verification and evaluation module receives the adjusted moisture data, restores the accurate moisture content through the data verification and decoding unit, generates a moisture measurement quality report through the quality evaluation unit, and outputs the evaluation results.

[0058] Compared with the prior art, the present invention provides a method and system for compensating for errors in tea moisture measurement, which has the following beneficial effects:

[0059] 1. In this invention, moisture content data of tea samples is collected in real time by a moisture measuring sensor to generate a raw moisture dataset. Then, data preprocessing operations are performed, including noise reduction, calibration and format standardization, to generate preprocessed moisture data. Adaptive error compensation processing is then performed based on environmental parameters to generate compensated moisture data, eliminating random fluctuations of the sensor and environmental interference, ensuring the initial accuracy and environmental adaptability of the moisture measurement data, and reducing initial errors.

[0060] 2. In this invention, an intelligent scheduling algorithm is used to arrange the compensated moisture data into a measurement sequence according to the priority of tea varieties and the importance of the measurement task, generate measurement sequence data, and send the data through a data transmission module and monitor the data error rate in real time, thereby optimizing the measurement task order and resource allocation, detecting transmission anomalies in real time, and improving measurement efficiency and data transmission reliability.

[0061] 3. In this invention, dynamic compensation and adjustment processing is performed based on monitoring results, including a retesting mechanism and adaptive adjustment of compensation parameters, to generate adjusted moisture data. Then, verification and decoding processing is performed at the receiving end to restore accurate moisture content data. In addition, quality assessment is performed in conjunction with an artificial intelligence model to generate a moisture measurement quality report, which ultimately ensures the accuracy and consistency of the moisture data and provides quality feedback and a basis for improvement. Attached Figure Description

[0062] Figure 1 This is a flowchart of a method for compensating for errors in tea moisture determination according to the present invention;

[0063] Figure 2 This is a schematic diagram of the error compensation system for tea moisture measurement according to the present invention. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] Specific embodiment: A method and system for compensating for errors in tea moisture determination, the method comprising the following steps:

[0066] S1. Collect moisture content data of tea samples in real time using a moisture sensor and generate a raw moisture dataset.

[0067] S2. Perform data preprocessing operations based on the original moisture dataset, including noise reduction, calibration and format normalization, and generate preprocessed moisture data;

[0068] S3. Based on environmental parameters, perform adaptive error compensation processing on the pre-processed moisture data to generate compensated moisture data.

[0069] S4. An intelligent scheduling algorithm is used to arrange the measurement sequence of the compensated moisture data according to the priority of tea varieties and the importance of the measurement task, and the measurement sequence data is generated.

[0070] S5. The measured sequence data is sent to the central processing station through the data transmission module, and the data error rate is monitored in real time during the transmission process;

[0071] S6. Based on the monitoring results, perform dynamic compensation and adjustment processing, including retesting mechanism and adaptive adjustment of compensation parameters, and generate adjusted moisture data;

[0072] S7. At the receiving end, the adjusted moisture data is verified and decoded to restore the accurate moisture content data.

[0073] S8. Combine artificial intelligence models to conduct quality assessments on accurate moisture content data and generate a moisture measurement quality report.

[0074] The moisture content data of tea samples collected in S1 includes the following steps:

[0075] S11. Deploy multimodal moisture sensors, including near-infrared spectrometers and capacitive moisture meters, in tea processing production lines and storage areas to collect raw data on the moisture content of tea samples and record the sample type.

[0076] S12. Install temperature and humidity sensors and barometers in the measurement environment to collect environmental parameter data in real time, including temperature, humidity and atmospheric pressure values.

[0077] S13. Mark the collected moisture data and environmental parameters using high-precision timestamps and unique sample identifiers;

[0078] S14. Integrate the labeled data into a structured dataset to generate the original moisture dataset, which includes CSV, JSON and binary formats to support subsequent processing and analysis.

[0079] Data preprocessing operations in S2 include the following steps:

[0080] S21. The sliding window averaging filter algorithm is used to perform noise reduction processing on the original moisture dataset to reduce interference caused by random fluctuations in the sensor.

[0081] First, the size of the sliding window is set. The window size can be adjusted according to the sampling frequency and noise level. Usually, the sliding window size is an odd value to avoid bias. For each time point, the filtered output value is obtained by calculating the arithmetic mean of the original data points within the window. The formula is as follows:

[0082] ;

[0083] in This represents the moisture data value at time t after filtering. This represents the sampled value of the original moisture data at time ti. Indicates the size of the sliding window;

[0084] During implementation, a sliding window is started from the beginning of the data sequence, and the average value is calculated point by point. When the window exceeds the data boundary, mirror filling and truncation are used to avoid edge effects. Finally, the smoothed moisture data sequence is output for subsequent calibration, and the filtering parameters are recorded for auditing.

[0085] S22. Correct sensor baseline drift using linear regression calibration technology and perform accuracy calibration by combining standard sample reference values;

[0086] First, a set of standard tea samples was collected, including reference values ​​for moisture content and corresponding raw sensor output values, to form a calibration dataset. The sample size should be at least 30 samples to ensure representativeness. Then, a linear regression model was applied for fitting the data, with the following formula:

[0087] ;

[0088] in This indicates the calibrated moisture value. This represents the sensor's raw output value. Indicates the slope. Indicates the intercept;

[0089] After fitting, the model is applied to real-time data, and a linear transformation is performed on each new sampled value to output a calibration value. At the same time, the calibration dataset is updated regularly to adapt to sensor drift, ensure long-term accuracy, and the model is only put into use after verifying that the model value is greater than 0.95.

[0090] S23. Convert the calibrated data into a standard JSON format to ensure compatibility with different measuring devices and systems;

[0091] S24. Perform range standardization on the converted data, normalize all moisture content values ​​to the relative humidity range of 0-100%, and generate pre-processed moisture data.

[0092] First, determine the maximum and minimum values ​​of the original moisture data by traversing the dataset and using real-time statistical methods; then, apply a linear transformation formula to each data point to map it to the target range [0, 100]:

[0093] ;

[0094] in This represents the standardized moisture value. This represents the original moisture value. This represents the minimum value in the dataset. This represents the maximum value in the dataset;

[0095] During implementation, to avoid the influence of extreme values, the maximum and minimum values ​​were calculated using historical data statistics. After processing, all data fell between 0% and 100%, facilitating subsequent compensation and evaluation.

[0096] The adaptive error compensation process in S3 includes the following steps:

[0097] S31. Monitor changes in environmental parameters in real time, including temperature gradient, humidity fluctuation and air pressure difference, and generate environmental status indicators.

[0098] S32. Based on environmental status indicators, dynamically select the compensation model: when environmental fluctuations are below the threshold, use a linear compensation algorithm; when environmental fluctuations are above the threshold, use a nonlinear compensation algorithm.

[0099] Linear compensation algorithm: suitable for stable environments, the formula is:

[0100] ;

[0101] in This indicates the moisture content after compensation. This indicates preprocessed data. and These are linear coefficients.

[0102] During implementation, environmental fluctuation indicators are monitored in real time, and a linear model is applied when they fall below the threshold.

[0103] Nonlinear compensation algorithm: suitable for environments with drastic changes, employing a polynomial model.

[0104] ;

[0105] in , , For nonlinear compensation coefficients, in the formula , , The relationship between environmental factors and moisture error is fitted by nonlinear regression; when environmental fluctuations exceed the threshold, the model is switched to nonlinear model to calculate the compensation value.

[0106] S33. Apply multiple regression analysis to model the error of moisture data and calculate the compensation coefficient of environmental factors on the measured values.

[0107] First, collect historical data, including moisture measurements and multiple environmental variables, covering different seasons; then establish a multiple linear regression model:

[0108] ;

[0109] in This indicates the moisture content measured. Indicates the intercept. , , Indicates the compensation coefficient. , , Represents the independent variable of the environment. Indicates the error term;

[0110] During implementation, the scikit-learn software library for Python was used to fit the model, calculate the coefficients and verify their significance. Then, the coefficients were applied to real-time data to adjust the moisture measurement values ​​according to the current environmental values. The model was periodically refitted to adapt to changes and a compensation coefficient report was output.

[0111] S34. Adjust the compensation parameters through a real-time feedback mechanism to generate compensated moisture data, the accuracy of which is adaptive to environmental conditions.

[0112] First, compare the compensated moisture value with the reference standard value in real time and calculate the absolute error:

[0113] ;

[0114] in For absolute error, To compensate for the moisture content, For reference standard values;

[0115] Then, use a PID controller to adjust the parameters:

[0116] ;

[0117] in This indicates the updated compensation parameters. Indicates the current parameter value. Indicates absolute error. , , Indicates the PID gain coefficient;

[0118] The feedback loop executes once every N sampling periods to ensure real-time performance. Simultaneously, a parameter change threshold is set to prevent over-adjustment; if the parameter exceeds the limit, an alarm is triggered.

[0119] Determining the sequence arrangement in S4 includes the following steps:

[0120] S41. Define tea variety priority parameters, including high-value varieties, conventional varieties, and experimental varieties, and assign different weights to them;

[0121] S42. Use convolutional neural networks to analyze moisture data trends, identify key events: abnormal moisture fluctuations, measure failure risks, and assign importance scores.

[0122] First, design the CNN architecture: the input layer receives the water content data sequence, followed by convolutional layers: using one-dimensional convolutional kernels, size 3, stride 1, activation function ReLU; pooling layers: max pooling, size 2; fully connected layers: output importance scores; during training, use a labeled dataset, the loss function is cross-entropy, the optimizer is Adam, and the learning rate is 0.001.

[0123] S43. Combining priority parameters and importance scores, a weighted round-robin algorithm is used to arrange the order of moisture determination;

[0124] First, calculate the overall weight for each measurement task using the following formula:

[0125] ;

[0126] in Indicates the overall weight. This indicates the task priority parameter. This indicates the importance score of the CNN output. , Represents the weight coefficient and ;

[0127] Then, a weighted round-robin scheduling method is adopted: the task queue is initialized and sorted in descending order of weight; the task with the highest weight is selected for execution each time, and its weight is reduced after execution to avoid starvation; the implementation includes real-time updating of the queue and allocation of transmission resources; the execution time of each task is configurable to ensure fairness.

[0128] S44. Allocate more processing resources to high-priority determination tasks to generate determination sequence data.

[0129] Real-time monitoring of data error rate in S5 includes the following steps:

[0130] S51. Divide the sequence data into fixed-size data packets, and add a sequence number, timestamp, and CRC checksum to each data packet;

[0131] S52. Data transmission is performed via the LoRa wireless communication protocol, and the ACK confirmation mechanism is enabled to monitor the data packet reception status.

[0132] First, the sequence data is divided into fixed-size data packets, and each packet is assigned a sequence number, timestamp, and CRC checksum. The sender sends the data packets via UDP protocol, and the receiver returns an ACK signal upon successful reception. ACK confirmation mechanism implementation details: The sender maintains a timer; if no ACK is received within the timeout period, the packet is retransmitted. Selective ACK handling of multiple lost packets is supported. The mechanism is implemented at the application layer to handle duplicate ACKs and out-of-order packets, improving robustness.

[0133] S53. Calculate the data error rate, compare the difference between the sent and received data based on the feedback from the receiving end, and trigger an alarm when the error rate exceeds 3%.

[0134] Error rate ;

[0135] in Indicates an error data packet. This represents the total number of data packets. During implementation, for transmission errors, the number of lost packets is counted based on the ACK mechanism. For measurement errors, the real-time value is compared with the calibration reference value, and the number of samples with absolute errors exceeding the threshold is calculated. The sliding window error rate for each sampling period is calculated and monitored in real time. If the error rate exceeds the threshold, an alarm is triggered and the log is recorded.

[0136] S54. Record transmission timestamps, environmental indicators, and device status to generate transmission monitoring logs.

[0137] The dynamic compensation adjustment process in S6 includes the following steps:

[0138] S61. When the data error rate exceeds the threshold, a selective retest request is automatically initiated, and only the error data points are retested.

[0139] S62. Adjust the parameter redundancy of the compensation algorithm and dynamically adjust the filter coefficient according to the environmental noise level;

[0140] First, parameter redundancy is defined as the number of compensation models running in parallel; adjustment rules: increase redundancy when channel noise is high, decrease it when noise is low; dynamically adjust filter coefficients: for sliding window filtering, the coefficients are dynamically adjusted according to the data variance, using the following formula:

[0141] ;

[0142] in Represents the filter coefficients. This represents the estimated variance of the data; during implementation, adjustments are performed every 5 seconds to ensure a balance between smoothness and response, and the adjustment history is recorded.

[0143] S63. Use the Kalman filter algorithm to smooth the retest data and reduce random errors;

[0144] The Kalman filter algorithm consists of two phases: prediction and update.

[0145] Prediction steps:

[0146] ;

[0147] ;

[0148] in Let represent the prior state estimate at time t. This represents the posterior state estimate at time t-1. Represents the state transition matrix. Represents the control input matrix. Indicates control input, This indicates a priori estimation of the covariance. This indicates the posterior estimate of the covariance. Represents the process noise covariance;

[0149] Update steps:

[0150] ;

[0151] ;

[0152] ;

[0153] in Indicates Kalman gain, This represents the updated state estimate. This represents the updated state covariance matrix. Represents the observation matrix. Represents the observation noise covariance. Represents the observed value at time t. Represents the identity matrix;

[0154] S64. Verify the consistency and stability of the adjusted data and generate adjusted moisture data.

[0155] The verification and decoding process in S7 includes the following steps:

[0156] S71. At the receiving end, the data packets are reassembled in order, and the original measurement sequence is recovered based on the sequence number;

[0157] S72. Apply the inverse compensation algorithm to decode the moisture data and restore it to the physical moisture content value;

[0158] Linear inverse compensation algorithm formula:

[0159] ;

[0160] in This represents the value after compensation. Represents the original value. and The parameters are for compensation; the nonlinear logic is the same as above.

[0161] S73. Perform data smoothing and outlier removal to output accurate moisture content data;

[0162] Exponential smoothing is used for smoothing:

[0163] ;

[0164] in This represents the smoothed moisture value at time t. This represents the original moisture value at time t. Represents the smoothing factor. This represents the smoothed value at time t-1;

[0165] Outlier removal:

[0166] ;

[0167] in Standard score, For individual observations, This is the overall average. The population standard deviation; If the value is greater than 3, it will be removed and replaced with adjacent values.

[0168] The process of generating a moisture content quality report in S8 includes the following steps:

[0169] S81. Use a pre-trained deep learning model LSTM network to evaluate the quality of moisture data, including standard deviation, confidence interval and bias.

[0170] A pre-trained LSTM model is used, with a water content sequence of length 50 as input and a PSNR (Power Score) output as the model architecture. The model consists of an LSTM layer with 64 units and a fully connected layer. During implementation, the model weights are loaded, real-time data is input, and the output score is displayed. The model is then fine-tuned periodically with new data.

[0171] S82. Compare the differences in moisture data before and after compensation, and then calculate the absolute error and relative error scores;

[0172] Absolute error:

[0173] ;

[0174] in Indicates absolute error. Indicates the measured value. Indicates the reference truth value;

[0175] Relative error:

[0176] = ;

[0177] in This represents the relative error value;

[0178] S83. Generate a moisture content determination quality report, including the quality grade, error analysis results, and improvement suggestions.

[0179] The system includes a data acquisition and preprocessing module, an intelligent compensation and scheduling module, a transmission control and adjustment module, and a receiving verification and evaluation module.

[0180] The data acquisition and preprocessing module acquires moisture data and environmental parameters of tea samples through a multi-sensor acquisition unit, optimizes the data through a noise elimination and calibration unit, and outputs the preprocessed moisture data through a format standardization unit.

[0181] The intelligent compensation and scheduling module receives pre-processed moisture data, obtains real-time parameters through the environmental state sensing unit, performs error processing using the adaptive compensation unit, and outputs measurement sequence data through the priority scheduling unit.

[0182] The transmission control and adjustment module receives the measured sequence data, performs the transmission operation through the wireless data transmission unit, collects data error information using the error monitoring unit, and outputs the adjusted moisture data through the dynamic adjustment unit.

[0183] The receiving verification and evaluation module receives the adjusted moisture data, restores the accurate moisture content through the data verification and decoding unit, generates a moisture measurement quality report through the quality evaluation unit, and outputs the evaluation results.

[0184] The method and system operation steps are as follows:

[0185] First, moisture content data of tea samples is collected in real time using a moisture sensor to generate a raw moisture dataset, while environmental parameter data is recorded simultaneously. Next, data preprocessing is performed on the raw moisture dataset, including noise reduction using a sliding window averaging filter algorithm, correction of sensor baseline drift and accuracy calibration using linear regression calibration technology, and finally, format and range standardization to generate preprocessed moisture data.

[0186] Subsequently, adaptive error compensation processing is performed on the preprocessed moisture data based on real-time monitored environmental parameters. By dynamically selecting linear and nonlinear compensation algorithms, multiple regression analysis is applied to calculate the compensation coefficients of environmental factors, and a real-time feedback mechanism is used to adjust the compensation parameters to generate compensated moisture data.

[0187] Then, an intelligent scheduling algorithm is used to arrange the compensated moisture data into measurement sequences according to the priority of tea varieties and the importance of the measurement task, generating measurement sequence data. The measurement sequence data is sent to the central processing station via a data transmission module, and the data error rate is monitored in real time during transmission using an ACK confirmation mechanism. Based on the monitoring results, dynamic compensation and adjustment processing is performed, including triggering a retest mechanism, dynamically adjusting compensation algorithm parameters, applying a Kalman filter algorithm for smoothing, and verifying data consistency and stability to generate adjusted moisture data. At the receiving end, the adjusted moisture data is verified and decoded, and accurate moisture content data is restored through an inverse compensation algorithm, data smoothing, and outlier removal.

[0188] Finally, an artificial intelligence model is used to assess the quality of accurate moisture content data and generate a moisture content measurement quality report. The system includes a data acquisition and preprocessing module, an intelligent compensation and scheduling module, a transmission control and adjustment module, and a receiving verification and evaluation module, which work together to implement the above-mentioned methodology.

[0189] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0190] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for compensating for errors in tea moisture determination, characterized in that: The method includes the following steps: S1. Collect moisture content data of tea samples in real time using a moisture sensor and generate a raw moisture dataset. S2. Perform data preprocessing operations based on the original moisture dataset, including noise reduction, calibration and format normalization, and generate preprocessed moisture data; S3. Perform adaptive error compensation processing on the preprocessed moisture data based on environmental parameters to generate compensated moisture data. The adaptive error compensation process in S3 includes the following steps: S31. Monitor changes in environmental parameters in real time, including temperature gradient, humidity fluctuation and air pressure difference, and generate environmental status indicators. S32. Based on the environmental state indicators, dynamically select the compensation model: when the environmental fluctuation is below the threshold, use a linear compensation algorithm; when the environmental fluctuation is above the threshold, use a nonlinear compensation algorithm. S33. Apply multiple regression analysis to model the error of moisture data and calculate the compensation coefficient of environmental factors on the measured values. S34. Adjust the compensation parameters through a real-time feedback mechanism to generate compensated moisture data, the accuracy of which is adaptive to environmental conditions. S4. Using an intelligent scheduling algorithm, the compensated moisture data is arranged into a measurement sequence according to the priority of tea varieties and the importance of the measurement task, thereby generating measurement sequence data; S5. The measured sequence data is sent to the central processing station through the data transmission module, and the data error rate is monitored in real time during the transmission process; S6. Based on the monitoring results, perform dynamic compensation and adjustment processing, including retesting mechanism and adaptive adjustment of compensation parameters, and generate adjusted moisture data; The dynamic compensation adjustment process in S6 includes the following steps: S61. When the data error rate exceeds the threshold, a selective retest request is automatically initiated, and only the error data points are retested. S62. Adjust the parameter redundancy of the compensation algorithm and dynamically adjust the filter coefficient according to the environmental noise level; S63. Use the Kalman filter algorithm to smooth the retest data and reduce random errors; S64. Verify the consistency and stability of the adjusted data, and generate the adjusted moisture data; S7. The adjusted moisture data is verified and decoded at the receiving end to restore the accurate moisture content data. S8. Combine the artificial intelligence model to perform a quality assessment on the accurate moisture content data and generate a moisture measurement quality report.

2. The method for compensating for errors in tea moisture determination according to claim 1, characterized in that: The step S1 involves collecting moisture content data from tea samples, including the following steps: S11. Deploy multimodal moisture sensors, including near-infrared spectrometers and capacitive moisture meters, in tea processing production lines and storage areas to collect raw data on the moisture content of tea samples and record the sample type. S12. Install temperature and humidity sensors and barometers in the measurement environment to collect environmental parameter data in real time, including temperature, humidity and atmospheric pressure values. S13. Mark the collected moisture data and environmental parameters using high-precision timestamps and unique sample identifiers; S14. Integrate the labeled data into a structured dataset to generate the original moisture dataset, which includes CSV, JSON and binary formats to support subsequent processing and analysis.

3. The method for compensating for errors in tea moisture determination according to claim 1, characterized in that: The data preprocessing operation in S2 includes the following steps: S21. The sliding window averaging filter algorithm is used to perform noise reduction processing on the original moisture dataset to reduce interference caused by random fluctuations in the sensor. S22. Correct sensor baseline drift using linear regression calibration technology and perform accuracy calibration by combining standard sample reference values; S23. Convert the calibrated data into a standard JSON format to ensure compatibility with different measuring devices and systems; S24. Perform range standardization on the converted data, normalizing all moisture content values ​​to the relative humidity range of 0-100%, and generating pre-processed moisture data.

4. The method for compensating for errors in tea moisture determination according to claim 1, characterized in that: The sequence arrangement determination in S4 includes the following steps: S41. Define tea variety priority parameters, including high-value varieties, conventional varieties, and experimental varieties, and assign different weights to them; S42. Use convolutional neural networks to analyze moisture data trends, identify key events: abnormal moisture fluctuations, measure failure risks, and assign importance scores. S43. Combining priority parameters and importance scores, a weighted round-robin algorithm is used to arrange the order of moisture determination; S44. Allocate more processing resources to high-priority determination tasks to generate determination sequence data.

5. The method for compensating for errors in tea moisture determination according to claim 1, characterized in that: The real-time monitoring of data error rate in S5 includes the following steps: S51. Divide the sequence data into fixed-size data packets, and add a sequence number, timestamp, and CRC checksum to each data packet; S52. Data transmission is performed via the LoRa wireless communication protocol, and the ACK confirmation mechanism is enabled to monitor the data packet reception status. S53. Calculate the data error rate, compare the difference between the sent and received data based on the feedback from the receiving end, and trigger an alarm when the error rate exceeds 3%. S54. Record transmission timestamps, environmental indicators, and device status to generate transmission monitoring logs.

6. The method for compensating for errors in tea moisture determination according to claim 1, characterized in that: The verification and decoding process in S7 includes the following steps: S71. At the receiving end, the data packets are reassembled in order, and the original measurement sequence is recovered based on the sequence number; S72. Apply the inverse compensation algorithm to decode the moisture data and restore it to the physical moisture content value; S73. Perform data smoothing and outlier removal to output accurate moisture content data.

7. The method for compensating for errors in tea moisture determination according to claim 1, characterized in that: The step S8 involves generating a moisture content quality report, including the following steps: S81. Use a pre-trained deep learning model LSTM network to evaluate the quality of moisture data, including standard deviation, confidence interval and bias. S82. Compare the differences in moisture data before and after compensation, and calculate the absolute error and relative error scores; S83. Generate a moisture content determination quality report, including the quality grade, error analysis results, and improvement suggestions.

8. A tea moisture content measurement error compensation system, used to implement the tea moisture content measurement error compensation method according to any one of claims 1-7, characterized in that: The system includes a data acquisition and preprocessing module, an intelligent compensation and scheduling module, a transmission control and adjustment module, and a receiving verification and evaluation module. The data acquisition and preprocessing module acquires moisture data and environmental parameters of tea samples through a multi-sensor acquisition unit, optimizes the data through a noise elimination and calibration unit, and outputs the preprocessed moisture data through a format standardization unit. The intelligent compensation and scheduling module receives the preprocessed moisture data, obtains real-time parameters through the environmental state sensing unit, performs error processing using the adaptive compensation unit, and outputs the measurement sequence data through the priority scheduling unit. The transmission control and adjustment module receives the measurement sequence data, performs a transmission operation through the wireless data transmission unit, collects data error information using the error monitoring unit, and outputs adjusted moisture data through the dynamic adjustment unit. The receiving verification and evaluation module receives the adjusted moisture data, restores the accurate moisture content through the data verification and decoding unit, generates a moisture measurement quality report through the quality evaluation unit, and outputs the evaluation results.

Images