Needle missing prevention monitoring method and system based on pressure sensor and acceleration sensor

Abstract

The invention discloses a missing needle prevention monitoring method and system based on a pressure sensor and an acceleration sensor. The method comprises the following steps: detecting the weight reduction amount when a needle is taken out through a pressure sensor embedded in a needle taking disc, and generating a needle taking initial data set containing a needle taking timestamp and a needle taking number; and according to the needle taking timestamp range of the needle taking initial data set, detecting the weight increment when the needles are put back through a pressure sensor embedded in the needle putting disc, and generating a needle putting initial data set containing the needle putting timestamp and the needle putting number, and the like. Sensors are installed on a needle dial and the wrist of a doctor, and needle taking, needle placing and needle applying action data are collected in real time. The multi-source data fusion and time synchronization technology is utilized, needle taking and placing data and acceleration data are subjected to millisecond-level alignment, an acceleration peak value is analyzed through a mode recognition algorithm, and needle applying and needle pulling actions are accurately recognized.

Description

Technical Field

[0001] The present invention relates to the technical field of acupuncture operation monitoring, and in particular to a method and system for preventing needle leakage based on a pressure sensor and an acceleration sensor. Background Art

[0002] Needle management during acupuncture treatment presents technical challenges, primarily manifesting in the inconsistency between needle counting and actual operational status monitoring. Traditional manual counting is prone to errors, while relying solely on weight sensors makes it difficult to accurately identify the actual status of needles in use. When a doctor removes a needle from the needle tray, the pressure sensor detects a decrease in weight but cannot confirm whether the needle has actually been inserted into the patient. Similarly, when the needle tray detects an increase in weight, it cannot confirm that all removed needles have been safely retrieved. The core contradictions in this technical field are how to avoid false alarms while maintaining system sensitivity, and how to achieve accurate needle tracking without affecting the doctor's normal operation. Summary of the Invention

[0003] In light of this, the present invention aims to propose a method and system for preventing needle leakage monitoring based on pressure sensors and acceleration sensors. By installing sensors on the needle tray and the physician's wrist, real-time data on needle removal, placement, and application is collected. This system utilizes multi-source data fusion and time synchronization technology to align needle removal and placement data with acceleration data at the millisecond level. Furthermore, a pattern recognition algorithm analyzes acceleration peaks to accurately identify needle application and removal actions.

[0004] According to one aspect of the present invention, a method for monitoring needle leakage prevention based on a pressure sensor and an acceleration sensor is provided, the method comprising:

[0005] The pressure sensor embedded in the needle retrieval tray detects the weight loss of the needle when it is removed, and generates an initial needle retrieval dataset containing the needle retrieval timestamp and the number of needles removed; based on the needle retrieval timestamp range of the initial needle retrieval dataset, the pressure sensor embedded in the needle placement tray detects the weight gain of the needle when it is put back, and generates an initial needle placement dataset containing the needle placement timestamp and the number of needles placed;

[0006] According to the acupuncture timestamp range of the initial acupuncture data set, the three-axis wrist acceleration data is collected by the accelerometer worn by the doctor during acupuncture, and an initial acceleration data set containing timestamps and acceleration values is generated;

[0007] Based on the timestamps of the initial needle removal and placement data sets and the initial acceleration data set, a time alignment filtering algorithm is used to synchronize the multi-source data, generating a fused data set containing synchronized timestamps and the corresponding needle removal and placement data and acceleration data.

[0008] The needle removal timestamp and the positive Z-axis peak value in the acceleration data were analyzed by pattern recognition algorithm to determine the time and number of acupuncture. The needle placement timestamp and the negative Z-axis peak value in the acceleration data were used to determine the time and number of needle removal. A paired dataset of acupuncture insertion and removal was generated.

[0009] The insertion and removal status of the needle is determined based on the needle insertion and removal paired data set. If the insertion and removal status are inconsistent, an alarm is triggered.

[0010] In this technical solution, the fusion of pressure sensor and accelerometer data enables multi-dimensional capture of the dynamic state of needles. The pressure sensor monitors static state changes during needle removal and placement, while the accelerometer captures the dynamic motion characteristics of acupuncture and needle withdrawal. This multi-source data fusion effectively mitigates the risk of misjudgment from a single sensor, significantly improving monitoring accuracy and reliability. The pressure sensor provides a static weight change signal, while the accelerometer provides a dynamic motion characteristic signal. Together, they comprehensively reflect needle usage. A time alignment filtering algorithm achieves temporal synchronization of multi-source data, ensuring temporal consistency across the data collected by different sensors, providing a reliable foundation for subsequent pattern recognition and state assessment. The time-synchronized fused dataset accurately reflects the timing of acupuncture and needle withdrawal, avoiding misjudgments caused by time deviations. A feature-based pattern recognition algorithm extracts key features from the acceleration data (such as positive and negative Z-axis peaks) to precisely identify the specific time and number of acupuncture and needle withdrawals. It also effectively filters out irrelevant acceleration changes (such as other wrist movements of the physician), improving recognition accuracy. The algorithm further distinguishes the characteristics of acupuncture and needle withdrawal, avoiding misjudgments due to similar movements. By generating paired datasets for acupuncture and needle removal, the system monitors needle insertion and removal status in real time, ensuring consistency. If a discrepancy occurs (e.g., a needle not removed or the number of needles removed is inconsistent), the system automatically triggers an alarm, effectively preventing missed needles and significantly improving medical safety. The entire monitoring process is fully automated, requiring no human intervention. The system automatically completes data collection, time synchronization, motion recognition, and status assessment, significantly improving monitoring efficiency. This method is applicable to a variety of medical scenarios, particularly procedures involving frequent needle use (such as acupuncture and injections). Through embedded sensor design (such as a needle tray for removal and placement, and an accelerometer worn by the physician's wrist), the system can flexibly adapt to different medical equipment and procedures without placing additional burden on physicians' operational processes. It is also compatible with different needle types and usage scenarios. If a discrepancy occurs between the insertion and removal statuses, the system triggers a real-time alarm, alerting medical staff to conduct prompt inspections, significantly reducing the risk of medical errors. This anti-missed needle monitoring method utilizes multi-source data fusion, time synchronization, pattern recognition, and status assessment to achieve real-time and accurate monitoring of needle usage, effectively preventing missed needles and ensuring medical safety.

[0011] In some embodiments, a time alignment filtering algorithm is used to synchronize the multi-source data based on the timestamps of the initial needle removal and placement data sets and the initial acceleration data set, generating a fused data set containing synchronized timestamps and corresponding needle removal and placement data and acceleration data, including:

[0012] A time synchronization model was established using the Kalman filter algorithm. The input parameters included a data transmission delay compensation value to align the timestamps of the initial needle removal dataset, the initial needle placement dataset, and the initial acceleration dataset at the millisecond level.

[0013] If the deviation of the timestamps of the three sets of data exceeds the threshold, the time axis is recalibrated according to the sudden increase feature point of the acceleration data. Based on the aligned timestamps, the weight reduction value of the needle removal disk and the weight increase value of the needle placement disk are merged;

[0014] The merged weight change data and the three-axis acceleration data are matched according to the synchronized timestamps. The timestamp corresponding to the sudden increase point of Z-axis acceleration during the acupuncture stage is extracted as the benchmark reference point. The timestamp corresponding to the peak value of Y-axis reverse acceleration during the needle removal stage is extracted as the benchmark reference point. The time alignment accuracy of the multi-source data is verified secondary through the benchmark reference points to generate a fused dataset containing synchronized timestamps, weight change data and three-axis acceleration data.

[0015] In the above technical solution, the Kalman filter algorithm, as an efficient recursive filtering method, can dynamically estimate the system state and effectively handle the noise and uncertainty in multi-source data. By establishing a time synchronization model through Kalman filtering, the delay of the wireless transmission module can be compensated to achieve millisecond-level time alignment. The algorithm eliminates the time deviation caused by wireless transmission delay through dynamic estimation and noise filtering, ensures the temporal consistency of data collected by different sensors, and provides a reliable basis for subsequent pattern recognition and state judgment. When the timestamp deviation of the three sets of data exceeds the preset threshold, the system will recalibrate the time axis based on the sudden increase feature points of the acceleration data (such as the signature features of acupuncture or needle removal). This dynamic calibration mechanism can cope with sudden clock deviations or data loss problems, significantly enhancing the robustness of the system.

[0016] By extracting the timestamps of the sudden increase in Z-axis acceleration during the acupuncture phase and the peak value of the Y-axis reverse acceleration during the needle removal phase as benchmark reference points, the time alignment accuracy of multi-source data can be further verified and optimized. These benchmark reference points provide a reliable reference for time synchronization, ensuring that the time axes of different data sources are fully aligned at key events (such as acupuncture and needle removal). Based on the synchronized timestamps, the weight change data is matched with the three-axis acceleration data to generate a fused dataset containing synchronized timestamps, weight change data, and three-axis acceleration data. This fused dataset not only contains information on the removal and replacement of needles, but also combines the dynamic characteristics of the acupuncture and needle removal actions, providing richer data support for subsequent pattern recognition and state judgment. In addition, through secondary verification of the benchmark reference points, the accuracy of time alignment can be dynamically verified, potential errors in the time alignment process can be discovered and corrected, and the accuracy and consistency of the data can be further improved.

[0017] In some embodiments, based on the timestamps of the initial needle removal dataset and the initial needle placement dataset and the initial acceleration dataset, a time alignment filtering algorithm is used to synchronize the multi-source data to generate a fused dataset containing synchronized timestamps and corresponding needle removal and placement data and acceleration data, which then includes:

[0018] The data from the pressure sensor and accelerometer are processed using a normalization method. The errors of the pressure sensor and accelerometer are calibrated using a pre-established calibration curve. The standard deviation of the pressure sensor and accelerometer data is calculated, and data points with deviations exceeding 2 times the standard deviation are eliminated.

[0019] The calibrated pressure sensor and acceleration sensor data are fused and the Kalman filter algorithm is used to smooth the fused data to generate a calibrated fusion data set.

[0020] In the above technical solution, normalization is used to preprocess the pressure sensor and acceleration sensor data, eliminating dimensional differences between sensors and making the data comparable. Normalization maps multi-dimensional data to a unified scale, providing standardized input for subsequent data fusion and analysis. Calibrating sensor system errors using pre-calibrated calibration curves significantly improves measurement data accuracy. By establishing a correspondence between sensor output and actual physical quantities, the calibration curve eliminates inherent sensor bias and ensures the authenticity of the measurement data.

[0021] By calculating the standard deviation of pressure sensor and accelerometer data and removing anomalous data points with deviations exceeding two times the standard deviation, we can effectively filter out noise and improve data reliability. Standard deviation calculation quantifies the degree of data dispersion, and the anomalous data removal mechanism reduces the impact of random errors on subsequent analysis, ensuring data quality. Fusion of calibrated pressure sensor and accelerometer data allows for the combined use of static weight change data and dynamic acceleration data to provide more comprehensive information on needle usage and enhance system monitoring capabilities.

[0022] Using a Kalman filter algorithm to smooth the fused data further reduces data noise and improves data smoothness and stability. By dynamically estimating system status and updating data in real time, the Kalman filter makes the data more stable and reliable, providing high-quality data support for subsequent pattern recognition and state assessment. The entire data calibration and fusion process is fully automated, requiring no human intervention. The system automatically completes data processing and calibration, significantly improving monitoring efficiency.

[0023] In some embodiments, a pattern recognition algorithm is used to analyze the needle removal timestamp and the positive Z-axis peak value in the acceleration data to determine the time and number of acupuncture applications, and the needle removal timestamp and the negative Z-axis peak value in the acceleration data to determine the time and number of needle removals, thereby generating a paired acupuncture application and removal dataset, including:

[0024] A Kalman filter algorithm is used to synchronize the needle removal timestamp with the Z-axis positive peak data. The calibrated Z-axis positive peak data is analyzed using a pattern recognition algorithm to identify the characteristic points of the acupuncture action. The needle removal timestamp is matched according to the characteristic point time window to determine the acupuncture time and number of needles, and the weight increase data recorded by the needle placement plate pressure sensor is obtained.

[0025] Extract needle placement timestamps and weight gain values. Use the Kalman filter algorithm to synchronize the needle placement timestamps with the Z-axis negative peak data. Analyze the calibrated Z-axis negative peak data using a pattern recognition algorithm to identify needle removal action feature points. Match the needle placement timestamps against the feature point time window to determine the needle removal time and number.

[0026] The number of needle applications and needle removals were compared to determine whether there was a quantitative difference. A paired data set containing the time of acupuncture, time of needle removal, number of needle applications, and number of needle removals was generated.

[0027] In the above technical solution, the Kalman filter algorithm dynamically estimates the system state and compensates for timestamp deviation in real time, ensuring precise alignment of the needle removal and placement timestamps with the acceleration data time axis. This algorithm effectively eliminates time deviations caused by factors such as wireless transmission delays, ensures timestamp accuracy, and provides a reliable time reference for subsequent action recognition. The pattern recognition algorithm accurately identifies the characteristic points of acupuncture and needle removal actions based on calibrated Z-axis positive and negative peak data. By distinguishing the positive Z-axis peak of acupuncture and the negative Z-axis peak of needle removal, the algorithm accurately identifies the acupuncture and needle removal actions, ensuring the accuracy of action recognition.

[0028] The feature point time window matching mechanism precisely matches the times of needle insertion and removal with the timestamps of needle removal and placement. By matching the feature point time windows, the system accurately records the time and number of needle insertion and removal, ensuring that the action time is consistent with the actual operation time. The weight increase data recorded by the needle placement plate pressure sensor further verifies the accuracy of the needle removal action, providing additional verification information to ensure that the needle removal action actually occurred and avoid misidentification errors.

[0029] By comparing the number of needles inserted and removed, the system monitors needle usage in real time, promptly identifying discrepancies and triggering alarms. This ensures consistency between insertion and removal, effectively preventing missed needles. The entire pattern recognition and time synchronization calibration process is fully automated, with the system automatically completing action recognition, time matching, and quantity comparison, significantly improving monitoring efficiency.

[0030] In some embodiments, the insertion and removal states of the needle are determined based on the acupuncture and removal paired dataset, and if the insertion and removal states are inconsistent, an alarm is triggered, including:

[0031] Based on the acupuncture and needle removal paired dataset, the time sequence correspondence between the positive peak value of acupuncture Z axis and the needle removal timestamp, as well as the negative peak value of needle removal Z axis and the needle placement timestamp, is extracted to generate an action time sequence association dataset.

[0032] According to the action timing association data set, the timing deviations between the positive peak of acupuncture and the timestamp of needle removal, as well as the negative peak of needle removal and the timestamp of needle placement, the difference between the number of needles applied and the number of needles removed, the average time difference between each needle application and needle removal, the absolute difference between the weight reduction of needle removal and the weight increase of needle placement, and the similarity between the acceleration peak curve and the predefined acupuncture and needle removal pattern are calculated respectively. Based on the above parameters, the insertion and removal status of the needle are judged. If the insertion and removal status are inconsistent, an alarm is triggered.

[0033] In this technical solution, a time-series association dataset is generated by extracting the temporal correspondence between the positive Z-axis peak during acupuncture application and the needle removal timestamp, and the negative Z-axis peak during needle removal and the needle placement timestamp. This step correlates the data collected by different sensors on the time axis, providing a basis for subsequent state judgment. The time-series association dataset clearly reflects the time sequence of acupuncture application and removal, ensuring that each action is supported by corresponding timestamps and sensor data.

[0034] By calculating the timing deviation between the positive peak of acupuncture and the timestamp of needle removal, and the negative peak of needle removal and the timestamp of needle placement, we can quantify the time difference between acupuncture and needle removal, and between needle removal and placement. This timing deviation reflects the accuracy of the action execution; excessive deviations may indicate improper operation or equipment delays. Further calculating the difference between the number of needles applied and removed allows for intuitive identification of any needles that have not been removed or a discrepancy between the number of needles removed. A zero difference indicates that the number of needles applied and removed matches; a non-zero difference triggers an alarm, alerting medical staff to conduct a prompt inspection.

[0035] Calculating the average time difference between each needle insertion and removal assesses the smoothness and consistency of the entire acupuncture and removal process. This average time difference reflects the stability of the procedure; excessively large or fluctuating averages may indicate delays or interruptions during the procedure. The consistency of weight change is verified by calculating the absolute difference between the decrease in needle weight upon removal and the increase in needle weight upon insertion. This absolute difference in weight change provides physical verification, ensuring that the actual needle removal and replacement are consistent with the sensor data.

[0036] The similarity between the peak acceleration curve and the predefined acupuncture and removal patterns is further calculated to verify the standardization of acupuncture and removal movements from the perspective of motion characteristics. This similarity calculation improves judgment accuracy and ensures that operations conform to standard patterns. Based on the comprehensive judgment of these multiple parameters, the system can accurately identify the insertion and removal status of the needle, triggering an alarm if the status is inconsistent. This multi-dimensional judgment mechanism significantly improves the reliability and accuracy of monitoring, effectively reduces misjudgments, and ensures timely warning of abnormal situations.

[0037] In some embodiments, based on the acupuncture and needle removal paired dataset, the temporal correspondence between the positive peak value of the acupuncture Z axis and the needle removal timestamp, and the negative peak value of the needle removal Z axis and the needle placement timestamp is extracted to generate an action timing association dataset, including:

[0038] Establish a dataset of acupuncture events, including weight loss value, number of needles, and Z-axis positive peak time and filtered timestamp;

[0039] Construct a needle removal event dataset, including weight increase, needle number, Z-axis negative peak time, and calibration timestamp;

[0040] Based on the time series characteristics of the acupuncture and needle removal event dataset, an action temporal correlation matrix is generated.

[0041] In the above technical solution, by establishing acupuncture event data sets and needle removal event data sets, key information during the acupuncture and needle removal processes can be systematically recorded, providing a basis for subsequent time series analysis. The acupuncture event data set records the weight reduction value, the number of needles, the positive peak moment of the Z axis, and the filtered timestamp, fully presenting the time series and related parameters of the acupuncture action; the needle removal event data set records the weight increase value, the number of needles, the negative peak moment of the Z axis, and the calibrated timestamp, ensuring that the needle removal action record accurately corresponds to the actual operation time. Based on the time series characteristics of the acupuncture and needle removal event data sets, an action time series correlation matrix is generated. This matrix intuitively displays the time correspondence between acupuncture and needle removal actions, provides a clear timeline view, and facilitates the identification of potential time deviations and operational anomalies. By multi-dimensionally correlating weight change data, acceleration data, and timestamps, the consistency of acupuncture and needle removal actions can be verified from different angles, reducing the misjudgment that may be caused by a single data source.

[0042] In some embodiments, based on the action timing association data set, the timing deviations between the positive peak of acupuncture and the timestamp of needle removal, and the negative peak of needle removal and the timestamp of needle placement, the difference between the number of needles applied and the number of needles removed, the average time difference between each needle application and needle removal, the absolute difference between the weight reduction of needle removal and the weight increase of needle placement, and the similarity between the acceleration peak curve and the predefined acupuncture and needle removal pattern are calculated, including:

[0043] Based on the action timing correlation data set, the timing deviation between the positive peak of acupuncture and the timestamp of needle removal, as well as the timing deviation between the negative peak of needle removal and the timestamp of needle placement, is calculated. If the deviation exceeds 50 milliseconds, a timing anomaly data set containing the abnormal timestamp and the deviation amount is generated.

[0044] Based on the time series anomaly dataset and the paired acupuncture and removal dataset, the difference between the number of acupuncture needles applied and the number of needles removed, as well as the mean time difference between each acupuncture needle application and removal, are calculated. If the difference is non-zero or the mean exceeds 2 seconds, a state anomaly dataset containing both quantity and time anomalies is generated.

[0045] Based on the abnormal state data set, combined with the initial needle removal data set and the initial needle placement data set, the absolute difference between the needle removal weight reduction and the needle placement weight increase is calculated. If the difference exceeds 0.5 grams, an alarm trigger data set containing weight anomalies is generated;

[0046] Based on the alarm trigger dataset and the acupuncture and removal paired dataset, the dynamic time warping algorithm is used to compare the similarity between the acceleration peak curve and the predefined acupuncture and removal pattern. If the similarity is lower than 0.8, a verification failure dataset containing the pattern mismatch timestamp is generated.

[0047] In this technical solution, the timing deviations between the positive peak of acupuncture insertion and the timestamp of needle removal, and the negative peak of needle removal and the timestamp of needle placement, are calculated to quantify the timing differences in action execution. If the deviation exceeds 50 milliseconds, a timing anomaly dataset containing the abnormal timestamp and the deviation is generated. This timing deviation reflects the accuracy of action execution, allowing for timely detection of timing anomalies and ensuring the consistency of timing between acupuncture insertion and removal, and between needle removal and placement.

[0048] Based on the time series anomaly dataset and the paired acupuncture and removal dataset, the difference between the number of needle applications and the number of needle removals, as well as the mean time difference between each application and removal, are calculated. If the number difference is non-zero or the mean time difference exceeds 2 seconds, a state anomaly dataset containing both number and time anomalies is generated. The number difference and mean time difference directly reflect the consistency and smoothness of acupuncture and removal actions, allowing for the timely detection of potential operational anomalies.

[0049] The data set containing the abnormal status, initial needle removal, and initial needle placement is combined to calculate the absolute difference between the needle removal weight loss and the needle placement weight gain. If the difference exceeds 0.5 grams, an alarm triggering data set containing the weight anomaly is generated. The absolute difference in weight change provides physical verification, ensuring that the actual needle removal and placement are consistent with the sensor data, further improving monitoring reliability.

[0050] A dynamic time warping algorithm was used to compare the similarity of the acceleration peak curve with predefined acupuncture and removal patterns. If the similarity was less than 0.8, a validation failure dataset containing the timestamp of the pattern mismatch was generated. The dynamic time warping algorithm can handle nonlinear changes in time series, improve the accuracy of similarity calculations, and ensure the standardization of acupuncture and removal movements.

[0051] Through comprehensive judgment based on multiple parameters, including time series deviation, quantity difference, time difference mean, absolute weight change difference, and similarity, the reliability and accuracy of monitoring are significantly improved. The multi-dimensional anomaly judgment mechanism effectively reduces misjudgments, ensuring that alarms are triggered only when anomalies exist, improving system practicality.

[0052] According to another aspect of the present invention, a needle leakage prevention monitoring system based on a pressure sensor and an acceleration sensor is provided, based on the above method; the system comprises:

[0053] The weight change monitoring module is used to detect the weight loss of the needle when it is removed through the pressure sensor embedded in the needle removal tray, and generate an initial needle removal data set containing the needle removal timestamp and the number of needles removed; based on the needle removal timestamp range of the initial needle removal data set, the pressure sensor embedded in the needle placement tray detects the weight gain of the needle when it is returned, and generates an initial needle placement data set containing the needle placement timestamp and the number of needles placed;

[0054] The acceleration speed acquisition module is used to collect wrist triaxial acceleration data through the acceleration sensor worn by the doctor's wrist during acupuncture according to the acupuncture timestamp range of the initial acupuncture data set, and generate an initial acceleration data set containing timestamps and acceleration values;

[0055] The data fusion module is used to synchronize the multi-source data using a time alignment filtering algorithm based on the timestamps of the initial needle removal and placement data sets and the initial acceleration data set, and generate a fused data set containing the synchronized timestamps and the corresponding needle removal and placement data and acceleration data;

[0056] The pairing module is used to analyze the needle removal timestamp and the positive Z-axis peak value in the acceleration data through a pattern recognition algorithm to determine the time and number of acupuncture, and to determine the time and number of needle removal by analyzing the needle placement timestamp and the negative Z-axis peak value in the acceleration data, thereby generating a paired dataset for acupuncture;

[0057] The alarm and feedback module is used to determine the insertion and removal status of the needle according to the needle insertion and removal paired data set, and trigger an alarm if the insertion and removal status are inconsistent.

[0058] In the above technical solution, in order to better use the above method, this application proposes an anti-leakage needle monitoring system based on a pressure sensor and an acceleration sensor. Each module corresponds to each step of the above method. The specific principles have been described above and will not be repeated here.

[0059] According to another aspect of the present invention, there is provided a device for monitoring needle leakage prevention based on a pressure sensor and an acceleration sensor, comprising:

[0060] at least one processor and a memory communicatively coupled to the at least one processor;

[0061] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the above method.

[0062] In the above technical solution, in order to better run and process the method, the above method is stored in a memory and a processor is used to execute the stored method. It should be noted that the principle and effect of each step have been described above and will not be further explained here.

[0063] According to another aspect of the present invention, a computer-readable storage medium is provided, storing a computer program, wherein the computer program implements the above method when executed by a processor.

[0064] In the above technical solution, in order to better run and use the method, the above method is stored in a computer-readable storage medium and implemented by a processor. It should be noted that the principle and effect of each step have been described above and will not be further explained here. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0066] Figure 1 This is a flow chart of an embodiment of a method for monitoring needle leakage prevention based on a pressure sensor and an acceleration sensor according to the present invention;

[0067] Figure 2 It is a structural diagram of an embodiment of an anti-leakage needle monitoring system based on a pressure sensor and an acceleration sensor of the present invention. DETAILED DESCRIPTION

[0068] The present invention will be described in further detail below with reference to the accompanying drawings and examples. It is particularly noted that the following examples are intended only to illustrate the present invention and are not intended to limit the scope of the present invention. Similarly, the following examples are only some embodiments of the present invention and are not intended to be exhaustive. All other embodiments obtained by those of ordinary skill in the art without creative effort are intended to fall within the scope of protection of the present invention.

[0069] This invention provides a method and system for preventing needle leakage monitoring based on pressure sensors and acceleration sensors. By installing sensors on the needle tray and the physician's wrist, the system collects real-time data on needle removal, placement, and application. Utilizing multi-source data fusion and time synchronization, the system aligns needle removal and placement data with acceleration data at the millisecond level. Furthermore, a pattern recognition algorithm analyzes acceleration peaks to accurately identify needle application and removal actions.

[0070] Example 1

[0071] In this embodiment, the overall solution framework is as follows:

[0072] Step S1: Monitoring the weight change during needle removal and placement

[0073] Needle Removal Tray Design: The tray incorporates a high-precision HX711 pressure sensor module to monitor needle weight changes during acupuncture. As the physician removes a needle from the tray, the pressure sensor detects the decrease in weight and records the duration of each operation, the amount of weight removed, and the number of needles removed. This precise record of weight changes during removal generates a timestamp and operational data for subsequent verification and analysis.

[0074] Needle tray design: The tray also incorporates a high-precision pressure sensor to monitor the weight gain during needle removal. This sensor senses the weight gain during needle removal and records data for each needle removed, including timestamp, weight gain, and number of needles returned. This ensures that every needle removal and return action is accurately recorded to prevent missed needles.

[0075] Step S2: Real-time recording of pressure data and data analysis

[0076] Real-time data collection: When the doctor removes a needle from the needle tray, the pressure sensor inside the tray detects the decrease in needle weight, and the system generates a record (including timestamp, weight loss, number of needles removed, etc.). This record is transmitted to the data processing system in real time via the wireless transmission module. When the doctor removes the needle and places it on the needle tray, the pressure sensor inside the tray records the increase in needle weight, and the system generates a corresponding record (including timestamp, weight gain, number of needles inserted, etc.).

[0077] Data Storage and Analysis: All collected pressure change data is transmitted to the data processing module, processed and stored using data analysis algorithms. The data analysis module verifies the needling and removal data and records the accuracy of each operation.

[0078] Step S3: Synchronizing the accelerometer with hand movements

[0079] Accelerometer Design: A triaxial MPU6050 accelerometer sensor module, mounted on the doctor's wrist, monitors hand movements during acupuncture insertion and removal. During acupuncture, the accelerometer records changes in vertical acceleration. When the needle contacts the skin, a significant increase in acceleration occurs, confirming successful insertion. During needle removal, the accelerometer records upward acceleration, with greater acceleration in the reverse direction, helping to determine whether the needle has been fully removed.

[0080] Real-time action detection: By combining acceleration data with pressure sensor data, the system will make real-time judgments on each acupuncture and needle removal action, and ensure the complete operation of each needle.

[0081] Step S4: Comprehensive analysis of pressure and acceleration sensor data

[0082] Data Fusion and Analysis: The data fusion module of this invention integrates data from pressure sensors and accelerometers. Using Kalman filtering technology, it precisely synchronizes weight changes and hand movements during acupuncture insertion and removal. The system compares each needle insertion and removal data to determine if any needles were missed. Multiple comparisons are performed on each needle's insertion and removal time, weight change, and acceleration changes.

[0083] Missed needle judgment mechanism:

[0084] If the system detects that the weight data for the needle insertion and removal process do not match, or the time and acceleration data for the needle insertion and removal do not meet expectations (for example, the weight does not completely decrease or increase), it is considered that there is a risk of missed needles.

[0085] Step S5: Anti-needle leakage alarm mechanism

[0086] Alarm Mechanism: If the system determines that the needle insertion and removal weights do not match, or if the accelerometer detects incomplete insertion or removal (incomplete needle insertion), the system will trigger a missed needle alarm. This alarm alerts the physician through audible alarms, vibrations, and visual reminders (LED display).

[0087] Alarm conditions: Abnormal needle removal disk data: If the recorded weight reduction does not match the expected value (for example, the weight reduction is not equal to the number of needles pulled out), a needle missing alarm is issued.

[0088] Abnormal needle plate data: If the recorded weight increase does not match the expected value (for example, the increased weight does not match the number of needles), a needle missing alarm is issued.

[0089] Abnormal acceleration data: If the acupuncture or needle removal action recorded by the acceleration sensor is inconsistent with the normal standard, the system will consider the action incomplete and trigger an alarm.

[0090] The present invention can accurately record the insertion and removal process of each needle by real-time monitoring of pressure changes and acceleration changes during acupuncture and needle removal, ensuring that no needle leakage occurs. By integrating pressure sensors and acceleration sensors, the present invention can determine in real time whether there is a risk of needle leakage during acupuncture and needle removal, and promptly issue an alarm to remind the doctor to make adjustments. This system effectively prevents the occurrence of needle leakage, improves the accuracy and safety of acupuncture treatment, and reduces the risk to patients. Each acupuncture and needle removal operation will be automatically recorded, providing a complete operation log, which is convenient for subsequent tracing and ensures the standardization of treatment.

[0091] Specifically:

[0092] See also Figure 1 , a method for monitoring needle leakage prevention based on a pressure sensor and an acceleration sensor, the method comprising:

[0093] S1. Detect the weight loss of the needles when they are removed using a pressure sensor embedded in the needle retrieval tray, generating an initial needle retrieval dataset containing a needle retrieval timestamp and the number of needles removed. Detect the weight gain of the needles when they are returned using a pressure sensor embedded in the needle placement tray, generating an initial needle placement dataset containing a needle placement timestamp and the number of needles placed, based on the needle retrieval timestamp range of the initial needle retrieval dataset.

[0094] For example, the needle retrieval tray is embedded with a high-precision pressure sensor HX711 sensor module to monitor the weight change when the needle is removed. Based on the monitored weight change, the weight reduction is calculated. An initial needle retrieval data set containing the needle retrieval timestamp and the number of needles removed is generated. Based on the initial needle retrieval data set, the matching relationship between the number of needles removed and the weight reduction is checked. If the number of needles removed matches the weight reduction, the needle retrieval operation is recorded as legal. If the number of needles removed does not match the weight reduction, the needle retrieval operation is marked as abnormal. The needle retrieval operation results are correlated with the subsequent needle removal operation data for analysis.

[0095] Specifically, a pressure sensor embedded in the needle retrieval tray monitors needle weight changes in real time. When a weight loss of 0.5 grams is detected, the system automatically records the current timestamp and marks it as a needle removal event. For example, if a weight loss of 0.5 grams is detected at 10:15:23, the system generates a record containing the timestamp "2023-10-05 10:15:23" and the number of needles removed, "1." The system then transmits this data to the data processing system via a wireless transmission module in JSON format. The JSON packet contains information such as the timestamp, the number of needles removed, and the sensor ID. Upon receiving the data, the data processing system first cleans the data to remove outliers, such as records with weight losses less than 0.3 grams or greater than 1 gram. The system then uses a sliding window algorithm to analyze needle removal events and calculate the frequency of needle removals within every 5 minutes. For example, if three needle removal events are detected between 10:15 and 10:20, the system generates a frequency of "0.6 times / minute." Finally, the system stores the analysis results in a database and generates visual reports for managers to review needle removal trends and anomalies. The entire process requires no human intervention, enabling automated collection, transmission, and analysis of needle removal data.

[0096] Exemplarily, a pressure sensor embedded in the needle placement disk is used to detect the weight increase when the needle is put back. Based on the weight increase, an initial needle placement data set containing the needle placement timestamp and the number of needles placed is generated. The initial needle placement data set is transmitted to the data processing system through the wireless transmission module. The data processing system receives the initial needle placement data set and matches the timestamp with the initial needle removal data set. If the needle placement timestamp is within the needle removal timestamp range, it is determined whether the number of needles placed is consistent with the number of needles removed. If the number of needles placed is consistent with the number of needles removed, the operation data is recorded as a normal operation. If the number of needles placed is inconsistent with the number of needles removed, the operation is marked as an abnormal operation. The normal operation data and the abnormal operation data are respectively stored in different modules of the data processing system. Based on the stored data, an operation log is generated for subsequent data tracing and analysis.

[0097] Specifically, the needle removal timestamp range for the initial needle removal dataset is set to 2023-10-01 08:00:00 to 2023-10-01 18:00:00. The system monitors weight changes in real time using a pressure sensor embedded in the needle tray. The sensor sampling frequency is 10Hz and the accuracy is ±0.1g. When the detected weight increase exceeds the threshold of 0.5g (the average weight of a single needle is 0.3g), a needle removal event is recorded.

[0098] For example, at 09:15:23, the system detects an increase in weight from an initial value of 500.0g to 500.9g. After eliminating environmental vibration interference using a Kalman filter, the system determines the number of needles placed as three (Δ weight 0.9g ÷ 0.3g / needle) and records the placement timestamp as 2023-10-01 09:15:23.450. The data processing system aggregates consecutive needle placement events using a sliding window algorithm with a 5-second window size. If a second weight increase of 0.6g is detected at 09:15:25, it is merged into a single event and the number of needles placed is updated to 5. All data is transmitted in JSON format, containing the fields {timestamp:"2023-10-01T09:15:23.450Z", weight:500.9, count:5}, with a transmission period of 1 minute. CRC-16 checksums are used to ensure data integrity. Upon receipt, the server uses a timestamp alignment algorithm to associate the placement data with the needle removal dataset.

[0099] S2. Based on the acupuncture timestamp range of the initial acupuncture data set, the doctor collects wrist triaxial acceleration data using an accelerometer worn on the doctor's wrist during acupuncture, generating an initial acceleration data set containing timestamps and acceleration values.

[0100] For example, the MPU6050 sampling frequency is configured to be 100Hz, the range is set to ±8g, and the vertical axis data priority acquisition mode is enabled. Data preprocessing: A sliding window filtering algorithm is used to eliminate hand tremor noise, and the window width is set to 10 sampling points. Feature extraction: The acceleration vector modulus is calculated, and a first-order difference operation is performed on the vertical acceleration data. Action recognition: If the vertical direction difference value exceeds the preset threshold of 0.5g / 10ms, it is determined that the acupuncture action is triggered. Action verification: The pressure sensor synchronization data is obtained. If the time difference between the sudden increase in acupuncture acceleration and the recorded time of the decrease in the weight of the needle removal plate is less than 200ms, it is marked as a valid action. Needle removal detection: When the angle between the wrist acceleration direction and the gravity direction exceeds 120 degrees and the acceleration modulus is greater than 1.2g, the needle removal action judgment is triggered. Specifically, when obtaining the acceleration data of the doctor's hand movement, the hand movement data during acupuncture and needle removal is first collected by the acceleration sensor installed on the doctor's wrist. The sensor records triaxial acceleration data at a sampling frequency of 100Hz, ensuring that subtle changes in hand movements are captured. During acupuncture, when the needle tip contacts the skin, the hand experiences a sudden increase in vertical acceleration, typically manifested as an increase in Z-axis acceleration from 0.5g to 2.5g within 0.1 second. Using a threshold algorithm, the system identifies the start of acupuncture when the Z-axis acceleration exceeds 1.5g. During needle removal, the hand experiences a reverse acceleration change, typically manifested as a decrease in Z-axis acceleration from 2.0g to -1.0g within 0.2 seconds. By analyzing the slope of the acceleration curve, the system identifies the start of needle removal when the slope changes from positive to negative and the absolute value exceeds a preset threshold. To further improve accuracy, the system uses a Kalman filter algorithm to smooth the collected acceleration data and remove noise. By analyzing this processed data in real time, the system can accurately identify key moments during acupuncture and needle removal, providing reliable data support for subsequent acupuncture operation analysis.

[0101] S3. Based on the timestamps of the initial needle removal and placement data sets and the initial acceleration data set, a time alignment filtering algorithm is used to synchronize the multi-source data to generate a fused data set containing synchronized timestamps and the corresponding needle removal and placement data and acceleration data.

[0102] In this embodiment, S3, based on the timestamps of the initial needle removal and placement data sets and the initial acceleration data set, a time alignment filtering algorithm is used to synchronize the multi-source data to generate a fused data set containing synchronized timestamps and corresponding needle removal and placement data and acceleration data, including:

[0103] S31. Using a Kalman filter algorithm to establish a time synchronization model, with the input parameters including a data transmission delay compensation value, the timestamps of the needle removal initial dataset, the needle placement initial dataset, and the acceleration initial dataset are aligned at the millisecond level.

[0104] S32. If the time stamp deviation of the three sets of data exceeds a threshold, recalibrate the time axis according to the sudden increase feature point of the acceleration data. Combine the weight reduction value of the needle removal disk and the weight increase value of the needle placement disk based on the aligned time stamps.

[0105] S33. Match the merged weight change data with the three-axis acceleration data according to the synchronized timestamps, extract the timestamp corresponding to the sudden increase point of Z-axis acceleration during the acupuncture phase as the benchmark reference point, and extract the timestamp corresponding to the peak value of Y-axis reverse acceleration during the needle removal phase as the benchmark reference point. Use the benchmark reference points to perform a secondary verification of the time alignment accuracy of the multi-source data to generate a fused data set containing synchronized timestamps, weight change data, and three-axis acceleration data.

[0106] Exemplarily, a Kalman filter algorithm is used to establish a time synchronization model, and the input parameters include the delay compensation value of the wireless transmission module. The timestamps of the needle removal initial data set, the needle placement initial data set and the acceleration initial data set are aligned at the millisecond level. If the timestamp deviation of the three sets of data exceeds the threshold, the time axis is recalibrated according to the sudden increase feature point of the acceleration data. Based on the aligned timestamps, the weight reduction value of the needle removal disk and the weight increase value of the needle placement disk are merged. The merged weight change data are matched with the three-axis acceleration data according to the synchronized timestamps. The timestamp corresponding to the Z-axis acceleration sudden increase point in the needle application stage is extracted as the benchmark reference point. The timestamp corresponding to the Y-axis reverse acceleration peak in the needle removal stage is extracted as the benchmark reference point. The time alignment accuracy of the multi-source data is verified twice through the benchmark reference point to generate a fused data set containing synchronized timestamps, weight change data and three-axis acceleration data. Specifically, a timestamp sequence, such as [1630000000123, 1630000000125, 1630000000127], was first extracted from the initial needle removal dataset. Simultaneously, the corresponding timestamps [1630000000124, 1630000000126, 1630000000128] were obtained from the initial needle placement dataset. A systematic offset of 1-2 milliseconds was observed between the two sequences. A dynamic time warping algorithm was used to calculate the optimal matching path, setting a sliding window of 5 milliseconds. Sequence alignment was achieved by minimizing the cumulative distance function, resulting in a corrected time offset of 1.5 milliseconds. For the acceleration data, the original sequence [0.12, 0.15, 0.18] g was acquired at a sampling frequency of 1000 Hz. Cubic spline interpolation was used to generate interpolated data of 0.147 g at the synchronized time point 1630000000125. A Kalman filter model was established with state variables including position, velocity, and acceleration. The process noise covariance was set to 0.01, and the observation noise covariance was set to 0.05. After 10 iterations, the optimal estimate was output. The resulting fused dataset included the aligned timestamp 1630000000125, the corresponding needle removal force of 5.6 N, the needle placement displacement of 3.2 mm, and a filtered acceleration of 0.146 g. The data synchronization error was kept within ±0.3 milliseconds.

[0107] In this embodiment, S3, based on the timestamps of the initial needle removal dataset and the initial needle placement dataset and the initial acceleration dataset, a time alignment filtering algorithm is used to synchronize the multi-source data to generate a fused dataset containing synchronized timestamps and corresponding needle removal and placement data and acceleration data; and then the following steps are also included:

[0108] S34, processing the data of the pressure sensor and the acceleration sensor using a normalization method, calibrating the errors of the pressure sensor and the acceleration sensor using a pre-established calibration curve; calculating the standard deviation of the data of the pressure sensor and the acceleration sensor, and eliminating data points with deviations exceeding 2 times the standard deviation;

[0109] S32: Fusing the calibrated pressure sensor and acceleration sensor data, and smoothing the fused data using a Kalman filter algorithm to generate a calibrated fused data set.

[0110] Exemplarily, a normalization method is used to process the data of the pressure sensor and the acceleration sensor, unifying the dimensions. The errors of the pressure sensor and the acceleration sensor are calibrated using a pre-established calibration curve. The standard deviation of the pressure sensor and the acceleration sensor data is calculated. It is determined whether the sensor data deviation exceeds 2 times the standard deviation. The data points with deviations exceeding 2 times the standard deviation are eliminated. The calibrated pressure sensor and acceleration sensor data are fused. The fused data is smoothed using the Kalman filter algorithm to generate a calibrated fused data set. The calibrated fused data set is transmitted to the data analysis module for subsequent processing. Specifically, in the fused data set, the raw data of the pressure sensor and the acceleration sensor are first unified in dimension using a normalization method. For example, the pressure sensor has a range of 0-10MPa, and the output value is 5V corresponding to 10MPa, while the acceleration sensor has a range of ±5g, and the output value is 2.5V corresponding to 5g. Min-max normalization is used to convert both to the range of 0-1. Then, the error is calibrated using a pre-established calibration curve. The pressure sensor uses the quadratic polynomial y=0.98x 2 Nonlinear compensation was performed using a y = +1.02x - 0.05 calibration, while the accelerometer used a linear compensation of y = 1.05x - 0.1. After calibration, outliers were removed using the 3σ criterion. The calculated pressure data had a mean of 0.52 and a standard deviation of 0.12, with data points exceeding the range of 0.52 ± 2 × 0.12 removed. The acceleration data had a mean of 0.48 and a standard deviation of 0.15, with data exceeding 0.48 ± 0.3 removed. The resulting dataset contained pressure eigenvalues of 0.51, 0.53, and 0.49, and acceleration eigenvalues of 0.47, 0.50, and 0.45. These features were timestamp-aligned to form a spatiotemporally synchronized fused dataset. Data processing employed a sliding window algorithm with a window size of 128 samples and a 50% overlap to ensure data continuity.

[0111] S4. Analyze the needle removal timestamp and the positive Z-axis peak value in the acceleration data using a pattern recognition algorithm to determine the time and number of acupuncture applications. Analyze the needle placement timestamp and the negative Z-axis peak value in the acceleration data to determine the time and number of needle removals, thereby generating a paired dataset for acupuncture application and removal.

[0112] In this embodiment, S4, using a pattern recognition algorithm to analyze the needle removal timestamp and the positive Z-axis peak value in the acceleration data to determine the time and number of acupuncture, and using the needle placement timestamp and the negative Z-axis peak value in the acceleration data to determine the time and number of needle removal, thereby generating a paired acupuncture data set; including:

[0113] S41. Use a Kalman filter algorithm to perform time synchronization calibration between the needle removal timestamp and the Z-axis positive peak data. Analyze the calibrated Z-axis positive peak data using a pattern recognition algorithm to identify characteristic points of acupuncture. Match the needle removal timestamp according to the characteristic point time window to determine the acupuncture time and number of needles. Obtain the weight increase data recorded by the needle placement plate pressure sensor.

[0114] S42. Extract the needle placement timestamp and weight gain value. Use the Kalman filter algorithm to perform time synchronization calibration on the needle placement timestamp and the Z-axis negative peak data. Analyze the calibrated Z-axis negative peak data using a pattern recognition algorithm to identify the needle removal action feature points. Match the needle placement timestamp with the feature point time window to determine the needle removal time and number of needles removed.

[0115] S43. Compare the number of needle applications and the number of needle removals to determine whether there is a quantity difference and generate a paired data set including the time of needle application, the time of needle removal, the number of needle applications, and the number of needle removals.

[0116] Exemplarily, a Kalman filter algorithm is used to perform time synchronization calibration between the needle removal timestamp and the Z-axis positive peak data. The calibrated Z-axis positive peak data is analyzed by a pattern recognition algorithm to identify the characteristic points of the acupuncture action, match the needle removal timestamp according to the characteristic point time window, and determine the acupuncture time and the number of needles. The weight increase data recorded by the needle placement disk pressure sensor is obtained to extract the needle placement timestamp and the weight increase value. The Kalman filter algorithm is used to perform time synchronization calibration between the needle placement timestamp and the Z-axis negative peak data. The calibrated Z-axis negative peak data is analyzed by a pattern recognition algorithm to identify the characteristic points of the needle removal action, match the needle placement timestamp according to the characteristic point time window, and determine the needle removal time and the number of needles removed. The number of acupuncture and the number of needles removed are compared to determine whether there is a quantity difference. A paired data set containing the acupuncture time, needle removal time, the number of acupuncture and the number of needles removed is generated. Specifically, the accelerometer data in the calibrated fusion dataset was first calibrated, and a sliding window peak detection algorithm was used to identify the extreme points in the Z-axis direction. The window size was set to 50 sampling points (corresponding to 0.5 seconds at a 100Hz sampling rate), and the positive peak threshold was set to 1.5g and the negative peak threshold was set to -1.2g. When the Z-axis acceleration was detected to exceed the positive threshold for three consecutive sampling points, it was recorded as an acupuncture event, and the time of the first sampling point exceeding the threshold was taken as the timestamp. For example, if the Z-axis value at 12:35:23.456 was 1.52g, 1.58g, and 1.61g, it was recorded as the acupuncture time. Similarly, when three consecutive sampling points were detected to be below the negative threshold, it was recorded as a needle removal event. For example, if -1.25g, -1.31g, and -1.28g appeared at 12:36:45.789, it was recorded as the needle removal time. A dynamic time warping (DTW) algorithm was used to pair acupuncture application and removal events, with a maximum time interval constraint of 300 seconds. A pairing was established if the time difference between the two events was within the constraint and the acceleration waveform similarity exceeded 0.85. For example, if the time difference between the acupuncture application event at 12:35:23.456 and the needle removal event at 12:36:45.789 was 82 seconds, and the waveform similarity was 0.87, the pairing was successful. The resulting paired dataset contained fields such as the original timestamp, Z-axis peak amplitude, and duration. For example, a complete data entry such as "Acupuncture application time 12:35:23.456 | Needle removal time 12:36:45.789 | Z-axis peak +1.61g / -1.31g | Duration 82 seconds" was recorded. The timestamps were smoothed using a Kalman filter to eliminate ±5 milliseconds of sensor jitter error.

[0117] S5. Determine the insertion and removal status of the needle according to the needle insertion and removal paired data set. If the insertion and removal status are inconsistent, trigger an alarm.

[0118] In this embodiment, S5, judging the insertion and removal status of the needle according to the acupuncture and removal paired data set, and triggering an alarm if the insertion and removal status are inconsistent, includes:

[0119] S51. Extract the time sequence correspondence between the positive peak value of the acupuncture application and the needle removal timestamp, and the negative peak value of the needle removal and the needle placement timestamp, based on the acupuncture application and needle removal paired dataset, to generate an action time sequence association dataset;

[0120] In this embodiment, S51, based on the acupuncture and needle removal paired data set, extract the time sequence correspondence between the positive peak value of the acupuncture Z axis and the needle removal timestamp, and the negative peak value of the needle removal Z axis and the needle placement timestamp, to generate an action time sequence association data set, including:

[0121] S511, establishing an acupuncture event data set, including weight loss value, needle number, Z-axis positive peak time and filtered timestamp;

[0122] S512, constructing a needle removal event data set, including weight increase value, needle number, Z-axis negative peak time, and calibration timestamp;

[0123] S513: Generate an action temporal correlation matrix based on the time series characteristics of the needle insertion and removal event data set.

[0124] From the paired acupuncture and needle removal dataset, we extracted the correspondence between the positive Z-axis peaks of acupuncture application and the needle removal timestamps. By setting the threshold for the positive Z-axis peaks to 0.5, we used a peak detection algorithm to identify peak points during acupuncture application and record their corresponding timestamps. For example, a positive Z-axis peak of 0.6 was detected during acupuncture application, corresponding to a timestamp of 2023-10-01 10:00:00. Next, we extracted the correspondence between the negative Z-axis peaks of needle removal and the timestamps of needle placement. By setting the threshold for the negative Z-axis peaks to -0.5, we also used a peak detection algorithm to identify peak points during acupuncture removal and record their corresponding timestamps. For example, a negative Z-axis peak of -0.7 was detected during acupuncture removal, corresponding to a timestamp of 2023-10-01 10:05:00. We then matched the temporal correspondences between the positive Z-axis peaks of acupuncture application and the needle removal timestamps, and between the negative Z-axis peaks of needle removal and the needle placement timestamps, to generate a temporal association dataset. By calculating the difference in timestamps, we can determine the time interval between acupuncture and removal, for example, a 5-minute interval. Finally, we use the resulting action temporal association dataset for further analysis, such as action duration and frequency, to provide data support for subsequent action optimization. For example, analysis revealed an average time interval of 5 minutes between acupuncture and removal, with a standard deviation of 1 minute, providing a reference for action standardization.

[0125] S52. Based on the action timing association data set, calculate the timing deviation between the positive peak of acupuncture and the timestamp of needle removal, as well as the negative peak of needle removal and the timestamp of needle placement, the difference between the number of needles applied and the number of needles removed, the average time difference between each needle application and needle removal, the absolute difference between the weight reduction of needle removal and the weight increase of needle placement, and the similarity between the acceleration peak curve and the predefined acupuncture and needle removal pattern. Based on the above parameters, judge the insertion and removal status of the needle. If the insertion and removal status are inconsistent, trigger an alarm.

[0126] In this embodiment, S52, based on the action timing association data set, respectively calculates the timing deviation between the positive peak of acupuncture and the timestamp of needle removal, and the negative peak of needle removal and the timestamp of needle placement, the difference between the number of needles applied and the number of needles removed, the average time difference between each needle application and needle removal, the absolute difference between the weight reduction of needle removal and the weight increase of needle placement, and the similarity between the acceleration peak curve and the predefined acupuncture and needle removal pattern, including:

[0127] S521. Calculate the timing deviation between the positive peak of acupuncture and the timestamp of needle removal, and between the negative peak of needle removal and the timestamp of needle placement, based on the action timing association data set. If the deviation exceeds 50 milliseconds, generate a timing anomaly data set containing the abnormal timestamp and the deviation amount.

[0128] For example, the acupuncture timing deviation ΔT1 = |T1-T2| is calculated, and the same algorithm is used to process the T3 and T4 data to calculate the needle removal timing deviation ΔT2 = |T3-T4|. If ΔT1 or ΔT2 exceeds the 50 millisecond threshold, the corresponding timestamp is marked as an abnormal data point, and the abnormal data point is associated with the deviation ΔT1 / ΔT2 to generate a structured timing anomaly data set. The data of the acupuncture positive peak and needle removal timestamp are extracted from the action timing association data set. Assume that the acupuncture positive peak occurs at the timestamp 123456789 milliseconds and the needle removal timestamp is 123456850 milliseconds. By calculating the timing deviation between the two, that is, subtracting the acupuncture positive peak timestamp from the needle removal timestamp, the deviation is 61 milliseconds. Since the deviation exceeds 50 milliseconds, the system will generate an abnormal timestamp record containing the timestamp 123456789 milliseconds and the deviation 61 milliseconds. Next, extract the timestamps for the needle removal and placement timestamps from the dataset. Assume that the negative peak for needle removal occurs at timestamp 123457000 milliseconds, and the timestamp for placement is 123457060 milliseconds. Calculate the timing deviation between the two: subtract the timestamp for the needle placement from the timestamp for the negative peak for needle removal, resulting in a deviation of 60 milliseconds. Similarly, since the deviation exceeds 50 milliseconds, the system generates another abnormal timestamp record with a timestamp of 123457000 milliseconds and a deviation of 60 milliseconds. Finally, the system integrates these abnormal timestamps and deviations into a single time series anomaly dataset for further analysis and processing.

[0129] S522. Based on the time series anomaly dataset and the paired acupuncture and removal dataset, calculate the difference between the number of acupuncture needles applied and the number of needles removed, as well as the mean time difference between each acupuncture needle application and removal. If the difference is non-zero or the mean exceeds 2 seconds, generate a state anomaly dataset containing both quantity and time anomalies.

[0130] For example, the time difference between each acupuncture and needle removal is calculated based on the acupuncture timestamp and the needle removal timestamp. The difference in the number of acupuncture needles and the number of needles removed is calculated. If the difference in the number of acupuncture needles and the number of needles removed is non-zero, a quantity anomaly record is generated. If the average time difference between acupuncture needles and the number of needles removed exceeds 2 seconds, a time anomaly record is generated. The quantity anomaly records and the time anomaly records are merged to generate a state anomaly data set. The timestamps of acupuncture and needle removal are extracted from the time series anomaly data set. Assuming that the acupuncture timestamp is [10:00:00, 10:00:05, 10:00:10] and the needle removal timestamp is [10:00:03, 10:00:08, 10:00:15]. By calculating the difference between the number of acupuncture needles and the number of needles removed, it is found that the number of acupuncture needles is 3, the number of needles removed is 3, and the difference is 0. Next, the time difference between each acupuncture and needle removal is calculated, which is 3 seconds, 3 seconds, and 5 seconds respectively, with an average time difference of 3.67 seconds. Because the mean time difference exceeds 2 seconds, a status anomaly dataset is generated, recording the anomaly information as "Abnormal mean time difference: 3.67 seconds." Simultaneously, the paired information for acupuncture and removal is extracted from the acupuncture and removal paired dataset. Assuming the number of acupuncture insertions is 4 and the number of removals is 3, the difference is 1. Because the difference is non-zero, a status anomaly dataset is generated, recording the anomaly information as "Abnormal difference between the number of acupuncture insertions and removals: 1." Through this analysis, combined with the time series anomaly dataset and the acupuncture and removal paired dataset, anomalies in the difference in the number of acupuncture insertions and removals and the mean time difference can be effectively identified. The corresponding status anomaly dataset is then generated, providing data support for subsequent anomaly handling.

[0131] S523. Calculate the absolute difference between the needle removal weight reduction and the needle placement weight increase based on the abnormal state data set, combined with the initial needle removal data set and the initial needle placement data set. If the difference exceeds 0.5 grams, generate an alarm trigger data set containing the weight abnormality.

[0132] For example, the corresponding needle removal weight reduction and needle placement weight increase are matched according to the needle removal timestamp and needle placement timestamp. The absolute difference between the needle removal weight reduction and the needle placement weight increase is calculated. If the difference exceeds 0.5 grams, an alarm trigger data set containing weight anomalies is generated. The alarm trigger data set is converted into an audio-visual prompt or vibration signal through the alarm and feedback module. The weight data of needle removal and needle placement are extracted from the state abnormality data set. Assume that the weight in the initial needle removal data set is 10.2 grams and the weight in the initial needle placement data set is 9.8 grams. By calculating the needle removal weight reduction, that is, the initial needle removal weight minus the weight after needle removal, assuming that the weight after needle removal is 9.7 grams, the needle removal weight reduction is 0.5 grams. Similarly, the needle placement weight increase is calculated, that is, the weight after needle placement minus the initial needle placement weight. Assuming that the weight after needle placement is 10.3 grams, the needle placement weight increase is 0.5 grams. Next, the absolute difference between the needle removal weight reduction and the needle placement weight increase is calculated, that is, |0.5 grams - 0.5 grams| = 0 grams. Since the difference does not exceed 0.5 grams, no alarm trigger data set is generated. If it is assumed that the weight after needle removal is 9.5 grams, the weight reduction after needle removal is 0.7 grams, and the weight after needle placement is 10.5 grams, the weight increase after needle placement is 0.7 grams. At this time, the absolute difference is |0.7 grams - 0.7 grams| = 0 grams, and no alarm trigger data set is generated. However, if the weight after needle removal is 9.3 grams, the weight reduction after needle removal is 0.9 grams, and the weight after needle placement is 10.1 grams, the weight increase after needle placement is 0.3 grams. At this time, the absolute difference is |0.9 grams - 0.3 grams| = 0.6 grams, which exceeds 0.5 grams. The system will generate an alarm trigger data set containing weight anomalies, record the weight changes of needle removal and needle placement, and trigger the corresponding alarm mechanism.

[0133] S524. Based on the alarm trigger data set and the acupuncture and removal paired data set, a dynamic time warping algorithm is used to compare the similarity between the acceleration peak curve and the predefined acupuncture and removal pattern. If the similarity is lower than 0.8, a verification failure data set containing a pattern mismatch timestamp is generated.

[0134] Exemplarily, an acceleration sensor module is used to collect hand motion data during acupuncture and needle removal, and obtain the peak acceleration curves of acupuncture and needle removal. Based on the weight change data of the needle removal plate and the needle placement plate, a paired dataset for acupuncture and needle removal is generated. The peak acceleration curves of acupuncture and needle removal are extracted through the acceleration sensor data. A dynamic time warping algorithm is used to compare the similarity of the peak acceleration curve with the predefined acupuncture and needle removal pattern. If the similarity is less than 0.8, a verification failure dataset containing a pattern mismatch timestamp is generated. Based on the acupuncture and needle removal paired dataset, the timestamps of acupuncture and needle removal are matched. Through the verification failure dataset, the time points of pattern mismatch during acupuncture and needle removal are identified. A data processing module is used to store the timestamps of pattern mismatch and the corresponding acupuncture and needle removal data. Based on the stored data, a detailed report on pattern mismatch during acupuncture and needle removal is generated. The peak acceleration curve is extracted from the alarm trigger dataset. For example, a 5-second acceleration data segment is collected with a sampling frequency of 100 Hz and a total of 500 data points. The peak value is detected through a sliding window, and the threshold is set to 2.5 m / s. 2 , identifying three significant peak points with timestamps of 1.2 seconds, 2.8 seconds, and 4.1 seconds. A predefined pattern template was then loaded from the acupuncture and needle removal paired dataset. This template contained typical acceleration characteristic curves for the acupuncture phase (lasting 0.6 seconds) and the needle removal phase (lasting 0.4 seconds), with a peak interval of 1.0±0.2 seconds. The dynamic time warping algorithm was used to match the two curves, setting the warping window width to 50 sampling points, and calculating the similarity score of the two curves. The algorithm first constructed a 500×500 cumulative cost matrix and found the optimal alignment path through a local constrained path search. The final similarity score was 0.72, which was lower than the set threshold of 0.8. The system automatically recorded the two peak points at 1.2 seconds and 4.1 seconds. The offset from the template exceeded the allowed range of ±0.15 seconds, and these two timestamps and their corresponding acceleration amplitudes (3.1m / s 2 and 2.9m / s 2 ) is stored in the failed validation dataset, with the expected peak times marked as 1.05 seconds and 3.95 seconds. The raw data is normalized using Z-score to eliminate dimensionality effects, and the template data is resampled to 100Hz using cubic spline interpolation to ensure a consistent sampling rate. If the similarity falls below 0.8 for five consecutive times, the system triggers an anomaly warning.

[0135] This embodiment scheme collects needle removal, needle placement and acupuncture action data in real time by installing sensors on the needle tray and the doctor's wrist. This scheme uses multi-source data fusion and time synchronization technology to align needle removal and placement data with acceleration data at the millisecond level, and analyzes the acceleration peak through pattern recognition algorithm to accurately identify acupuncture and needle removal actions. This embodiment scheme also establishes an action timing association model, and realizes intelligent detection of abnormal operations by comparing the timing relationship between needle removal and placement and the acceleration peak, as well as the number and time differences of acupuncture and needle removal. When abnormal weight, mismatch of action mode, etc. are found, the present invention can trigger an alarm in time to effectively prevent medical risks such as needle loss and improve the safety and standardization of acupuncture treatment.

[0136] Example 2

[0137] See also Figure 2 , a needle leakage prevention monitoring system based on a pressure sensor and an acceleration sensor, based on the method described in one of the embodiments; the system comprises:

[0138] The weight change monitoring module is used to detect the weight loss of the needle when it is removed through the pressure sensor embedded in the needle removal tray, and generate an initial needle removal data set containing the needle removal timestamp and the number of needles removed; based on the needle removal timestamp range of the initial needle removal data set, the pressure sensor embedded in the needle placement tray detects the weight gain of the needle when it is returned, and generates an initial needle placement data set containing the needle placement timestamp and the number of needles placed;

[0139] The acceleration speed acquisition module is used to collect wrist triaxial acceleration data through the acceleration sensor worn by the doctor's wrist during acupuncture according to the acupuncture timestamp range of the initial acupuncture data set, and generate an initial acceleration data set containing timestamps and acceleration values;

[0140] The data fusion module is used to synchronize the multi-source data using a time alignment filtering algorithm based on the timestamps of the initial needle removal and placement data sets and the initial acceleration data set, and generate a fused data set containing the synchronized timestamps and the corresponding needle removal and placement data and acceleration data;

[0141] The pairing module is used to analyze the needle removal timestamp and the positive Z-axis peak value in the acceleration data through a pattern recognition algorithm to determine the time and number of acupuncture, and to determine the time and number of needle removal by analyzing the needle placement timestamp and the negative Z-axis peak value in the acceleration data, thereby generating a paired dataset for acupuncture;

[0142] The alarm and feedback module is used to determine the insertion and removal status of the needle according to the needle insertion and removal paired data set, and trigger an alarm if the insertion and removal status are inconsistent.

[0143] In the above technical solution, in order to better use the method described in one of the embodiments, the present application proposes an anti-leakage needle monitoring system based on a pressure sensor and an acceleration sensor. Each module corresponds to each step of the above method. The specific principles have been described above and will not be repeated here.

[0144] Embodiment 3

[0145] A device for monitoring needle leakage prevention based on a pressure sensor and an acceleration sensor, comprising:

[0146] at least one processor and a memory communicatively coupled to the at least one processor;

[0147] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in one of the embodiments.

[0148] In the above technical solution, in order to better run and process the method described in one of the embodiments, the above method is stored in a memory and executed by a processor. It should be noted that the principle and effect of each step have been described above and will not be further explained here.

[0149] Example 4

[0150] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the method described in one of the embodiments.

[0151] In the above technical solution, in order to better run and use the method described in one of the embodiments, the above method is stored in a computer-readable storage medium and implemented using a processor. It should be noted that the principles and effects of each step have been described above and will not be further explained here.

[0152] The above descriptions are only some embodiments of the present invention and do not limit the scope of protection of the present invention. Any equivalent device or equivalent process transformation made by using the contents of the description and drawings of the present invention, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for monitoring needle leakage based on a pressure sensor and an acceleration sensor, characterized in that: The method comprises: The pressure sensor embedded in the needle retrieval tray detects the weight loss of the needle when it is removed, and generates an initial needle retrieval dataset containing the needle retrieval timestamp and the number of needles removed; based on the needle retrieval timestamp range of the initial needle retrieval dataset, the pressure sensor embedded in the needle placement tray detects the weight gain of the needle when it is put back, and generates an initial needle placement dataset containing the needle placement timestamp and the number of needles placed; According to the acupuncture timestamp range of the initial acupuncture data set, the three-axis wrist acceleration data is collected by the accelerometer worn by the doctor during acupuncture, and an initial acceleration data set containing timestamps and acceleration values is generated; Based on the timestamps of the initial needle removal and placement data sets and the initial acceleration data set, a time alignment filtering algorithm is used to synchronize the multi-source data, generating a fused data set containing synchronized timestamps and the corresponding needle removal and placement data and acceleration data. The needle removal timestamp and the positive Z-axis peak value in the acceleration data were analyzed by pattern recognition algorithm to determine the time and number of acupuncture. The needle placement timestamp and the negative Z-axis peak value in the acceleration data were used to determine the time and number of needle removal. A paired dataset of acupuncture insertion and removal was generated. The insertion and removal status of the needle is determined based on the needle insertion and removal paired data set. If the insertion and removal status are inconsistent, an alarm is triggered.

2. The anti-leakage needle monitoring method based on a pressure sensor and an acceleration sensor according to claim 1, characterized in that: Based on the timestamps of the initial needle removal and placement datasets, as well as the initial acceleration dataset, a time alignment filtering algorithm is used to synchronize the multi-source data. This generates a fused dataset containing synchronized timestamps and the corresponding needle removal and placement data and acceleration data, including: A time synchronization model was established using the Kalman filter algorithm. The input parameters included a data transmission delay compensation value to align the timestamps of the initial needle removal dataset, the initial needle placement dataset, and the initial acceleration dataset at the millisecond level. If the deviation of the timestamps of the three sets of data exceeds the threshold, the time axis is recalibrated according to the sudden increase feature point of the acceleration data. Based on the aligned timestamps, the weight reduction value of the needle removal disk and the weight increase value of the needle placement disk are merged; The merged weight change data and the three-axis acceleration data are matched according to the synchronized timestamps. The timestamp corresponding to the sudden increase point of Z-axis acceleration during the acupuncture stage is extracted as the benchmark reference point. The timestamp corresponding to the peak value of Y-axis reverse acceleration during the needle removal stage is extracted as the benchmark reference point. The time alignment accuracy of the multi-source data is verified secondary through the benchmark reference points to generate a fused dataset containing synchronized timestamps, weight change data and three-axis acceleration data.

3. The anti-leakage needle monitoring method based on a pressure sensor and an acceleration sensor according to claim 1, characterized in that: Based on the timestamps of the initial needle removal and placement data sets and the initial acceleration data set, a time alignment filtering algorithm is used to synchronize the multi-source data to generate a fused data set containing synchronized timestamps and the corresponding needle removal and placement data and acceleration data. This data set also includes: The data from the pressure sensor and accelerometer are processed using a normalization method. The errors of the pressure sensor and accelerometer are calibrated using a pre-established calibration curve. The standard deviation of the pressure sensor and accelerometer data is calculated, and data points with deviations exceeding 2 times the standard deviation are eliminated. The calibrated pressure sensor and acceleration sensor data are fused and the Kalman filter algorithm is used to smooth the fused data to generate a calibrated fusion data set.

4. The anti-leakage needle monitoring method based on a pressure sensor and an acceleration sensor according to claim 1, characterized in that: The pattern recognition algorithm analyzes the needle removal timestamp and the positive Z-axis peak value in the acceleration data to determine the time and number of acupuncture. The needle placement timestamp and the negative Z-axis peak value in the acceleration data are used to determine the time and number of needle removal. This generates a paired dataset of acupuncture insertion and removal, including: A Kalman filter algorithm is used to synchronize the needle removal timestamp with the Z-axis positive peak data. The calibrated Z-axis positive peak data is analyzed using a pattern recognition algorithm to identify the characteristic points of the acupuncture action. The needle removal timestamp is matched according to the characteristic point time window to determine the acupuncture time and number of needles, and the weight increase data recorded by the needle placement plate pressure sensor is obtained. Extract needle placement timestamps and weight gain values. Use the Kalman filter algorithm to synchronize the needle placement timestamps with the Z-axis negative peak data. Analyze the calibrated Z-axis negative peak data using a pattern recognition algorithm to identify needle removal action feature points. Match the needle placement timestamps against the feature point time window to determine the needle removal time and number. The number of needle applications and needle removals were compared to determine whether there was a quantitative difference. A paired data set containing the time of acupuncture, time of needle removal, number of needle applications, and number of needle removals was generated.

5. The anti-leakage needle monitoring method based on a pressure sensor and an acceleration sensor according to claim 1, characterized in that: The insertion and removal status of the needle is determined based on the needle insertion and removal paired data set. If the insertion and removal status are inconsistent, an alarm is triggered, including: Based on the paired dataset of acupuncture and needle removal, the time sequence correspondence between the positive peak value of acupuncture Z axis and the timestamp of needle removal, as well as the negative peak value of needle removal Z axis and the timestamp of needle placement, is extracted to generate an action time sequence association dataset. According to the action timing association data set, the timing deviations between the positive peak of acupuncture and the timestamp of needle removal, as well as the negative peak of needle removal and the timestamp of needle placement, the difference between the number of needles applied and the number of needles removed, the average time difference between each needle application and needle removal, the absolute difference between the weight reduction of needle removal and the weight increase of needle placement, and the similarity between the acceleration peak curve and the predefined acupuncture and needle removal pattern are calculated respectively. Based on the above parameters, the insertion and removal status of the needle are judged. If the insertion and removal status are inconsistent, an alarm is triggered.

6. The method for monitoring needle leakage prevention based on a pressure sensor and an acceleration sensor according to claim 5, characterized in that: Based on the acupuncture and needle removal paired dataset, the time sequence correspondence between the positive peak value of the acupuncture Z axis and the needle removal timestamp, as well as the negative peak value of the needle removal Z axis and the needle placement timestamp, is extracted to generate an action time sequence association dataset, including: Establish a dataset of acupuncture events, including weight loss value, number of needles, and Z-axis positive peak time and filtered timestamp; Construct a needle removal event dataset, including weight increase, needle number, Z-axis negative peak time, and calibration timestamp; Based on the time series characteristics of the acupuncture and needle removal event dataset, an action temporal correlation matrix is generated.

7. The method for monitoring needle leakage prevention based on a pressure sensor and an acceleration sensor according to claim 5, wherein: Based on the action timing association data set, the timing deviations between the positive peak of acupuncture and the timestamp of needle removal, as well as the negative peak of needle removal and the timestamp of needle placement, the difference between the number of needles applied and the number of needles removed, the average time difference between each needle application and needle removal, the absolute difference between the weight reduction of needle removal and the weight increase of needle placement, and the similarity between the acceleration peak curve and the predefined acupuncture and needle removal pattern are calculated, including: Based on the action timing correlation data set, the timing deviation between the positive peak of acupuncture and the timestamp of needle removal, as well as the timing deviation between the negative peak of needle removal and the timestamp of needle placement, is calculated. If the deviation exceeds 50 milliseconds, a timing anomaly data set containing the abnormal timestamp and the deviation amount is generated. Based on the time series anomaly dataset and the paired acupuncture and removal dataset, the difference between the number of acupuncture needles applied and the number of needles removed, as well as the mean time difference between each acupuncture needle application and removal, are calculated. If the difference is non-zero or the mean exceeds 2 seconds, a state anomaly dataset containing both quantity and time anomalies is generated. Based on the abnormal state data set, combined with the initial needle removal data set and the initial needle placement data set, the absolute difference between the needle removal weight reduction and the needle placement weight increase is calculated. If the difference exceeds 0.5 grams, an alarm trigger data set containing weight anomalies is generated; Based on the alarm trigger dataset and the acupuncture and removal paired dataset, the dynamic time warping algorithm is used to compare the similarity between the acceleration peak curve and the predefined acupuncture and removal pattern. If the similarity is lower than 0.8, a verification failure dataset containing the pattern mismatch timestamp is generated.

8. A needle leakage prevention monitoring system based on a pressure sensor and an acceleration sensor, characterized in that: Based on the method according to any one of claims 1 to 7; the system comprises: The weight change monitoring module is used to detect the weight loss of the needle when it is removed through the pressure sensor embedded in the needle removal tray, and generate an initial needle removal data set containing the needle removal timestamp and the number of needles removed; based on the needle removal timestamp range of the initial needle removal data set, the pressure sensor embedded in the needle placement tray detects the weight gain of the needle when it is returned, and generates an initial needle placement data set containing the needle placement timestamp and the number of needles placed; The acceleration speed acquisition module is used to collect wrist triaxial acceleration data through the acceleration sensor worn by the doctor's wrist during acupuncture according to the acupuncture timestamp range of the initial acupuncture data set, and generate an initial acceleration data set containing timestamps and acceleration values; The data fusion module is used to synchronize the multi-source data using a time alignment filtering algorithm based on the timestamps of the initial needle removal and placement data sets and the initial acceleration data set, and generate a fused data set containing the synchronized timestamps and the corresponding needle removal and placement data and acceleration data; The pairing module is used to analyze the needle removal timestamp and the positive Z-axis peak value in the acceleration data through a pattern recognition algorithm to determine the time and number of acupuncture, and to determine the time and number of needle removal by analyzing the needle placement timestamp and the negative Z-axis peak value in the acceleration data, thereby generating a paired dataset for acupuncture; The alarm and feedback module is used to determine the insertion and removal status of the needle according to the needle insertion and removal paired data set, and trigger an alarm if the insertion and removal status are inconsistent.

9. A needle leakage prevention monitoring device based on a pressure sensor and an acceleration sensor, characterized in that: include: at least one processor and a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images