45 results about "Average filter" patented technology

The invention discloses a GNSS (Global Navigation Satellite System) signal spoofing detection method, device and equipment, a storage medium and a program product. The method comprises the following steps: tracking and monitoring a plurality of GNSS satellite signals, calling a tracking loop correlator to carry out autocorrelation operation to obtain the autocorrelation power of each satellite signal, constructing the AP detection quantity of each satellite signal based on the autocorrelation power, verifying the Gaussian property of the AP detection quantity, and carrying out moving average filtering processing on the AP detection quantity conforming to Gaussian distribution to obtain the AP detection quantity of each satellite signal. The method comprises the following steps: acquiring Gaussian distribution characteristic information of a GNSS satellite signal, constructing an AP-MA detection quantity, determining a detection threshold according to the Gaussian distribution characteristic information, detecting the AP-MA detection quantity based on the detection threshold, and judging whether the GNSS satellite signal has deception jamming or not; the AP detection quantity is constructed based on the self-correlation power to carry out signal spoofing detection, so that the detection range and the detection precision are effectively improved, and the timeliness and the stability of spoofing detection are ensured under the condition of not increasing additional equipment and algorithm complexity.

The invention relates to the technical field of agricultural informatization and intelligent management, in particular to a selenium-rich bio-organic fertilizer research and development and demonstration and popularization full-process management system, which comprises a raw material digital filing module, a raw material fingerprint feature vector generation module and a selenium-rich bio-organic fertilizer research and development and demonstration and popularization module, and is characterized in that the raw material digital filing module is used for executing a moving average filtering algorithm to clean data; the market demand reverse analysis module is used for extracting, calculating and generating an environment demand constraint matrix representing market differentiation demands based on a preset element circulation rule; the research and development and production scheduling intelligent decision-making module is used for starting an evolutionary resolving model, mapping raw material fingerprints to an environment constraint matrix and outputting a dynamic production management instruction list; the production execution supervision module is used for distributing an instruction to the terminal and monitoring process execution data in real time; the full-life-cycle closed-loop optimization module is used for obtaining feedback data calculation deviation and generating correction factors to update model parameters; the method solves the contradiction between the heterogeneity of the non-standard biomass raw materials and the precise nutritional requirements of functional agricultural products.

The invention discloses a long-time filtering control method and device for analog quantity signals, and relates to the technical field of signal processing, and the method comprises the steps: constructing a circular buffer structure with the length of N as a historical data queue, and storing sampling values according to a PLC scanning period; based on the queue, calculating a filtering output value by adopting a weighted recursive average filtering algorithm; calculating a mean value and a standard deviation of the queue data in real time; comparing the current sampling value with a dynamic threshold range determined based on the mean value and the standard deviation, and if the current sampling value exceeds the dynamic threshold range, determining an abnormal value; when an abnormal value occurs, the value is discarded, the queue is not updated, and a filtering value of the last period is output as a current result, the defect that a traditional PLC filtering method is difficult to process long-time signal abnormity is effectively overcome through the circular queue with the adjustable time window and by fusing dynamic weighted filtering and a self-adaptive abnormity cleaning mechanism, and the filtering accuracy is improved. And the anti-interference capability and the reliability of the industrial control system are obviously improved.

This invention relates to the field of coordinated control of energy storage in new energy collection stations, and particularly to a coordinated control method for reactive power compensation of energy storage in new energy collection stations based on impedance reshaping. Based on the acquired three-phase AC signal, an analytical signal representation is obtained. Based on the analytical signal representation, a complex spectrum is obtained, and the impedance frequency response is determined. After processing with an exponentially weighted cumulative average filter, the actual impedance is obtained. Combined with the target impedance determined by the target impedance response model, the impedance error is determined. Based on the impedance error, a frequency-sensitive adjustment driving function is determined, and a new energy collection station adjustment residual loss function is constructed. The optimal complex impedance weighted path scheduling matrix is ​​determined. Simultaneously, a frequency segmentation objective function mechanism and a suppression function are introduced to construct a global optimization objective function, obtaining the optimal virtual impedance configuration. This solves the technical problems of resonance and stability issues, as well as reactive power response conflicts and low control efficiency caused by independent control strategies and a lack of coordinated adjustment.

The application provides a frequency standard measurement method, device, computer equipment and storage medium, the method comprises: frequency source output measurement signal is carried out frequency division processing, obtains low frequency pulse signal; using high-precision time interval counter to low frequency pulse signal is carried out delay measurement and time interval counting, and is output in the form of time stamp; by setting a sliding window, using sliding average algorithm to time stamp is carried out sliding average filtering processing, all time stamps are carried out sliding difference to obtain time difference, time difference is grouped according to order and is taken average in group to obtain average time difference; by increasing the length of the sliding window, the number of samples is increased to reduce the statistical error; using average time difference, calculate frequency measurement value, relative frequency deviation and allan variance. The method of the application increases the sliding window, and the frequency measurement result obtained by the algorithm of grouping and averaging the time difference has higher measurement precision and stability.

The invention discloses a robust wind speed prediction method, system and device and a medium, and the method comprises the steps: screening a wind power plant according to the relevance, and obtaining a screened wind power plant cluster data set; performing de-noising processing on the data by applying weighted average filtering to obtain preprocessed wind power plant cluster wind speed data; filling missing data values through a multi-dimensional average algorithm to obtain historical data of the wind power plant cluster; the method comprises the following steps: combining wind speed data of each wind power plant to construct a multi-dimensional sequence to obtain a wind power cluster data input set; and learning and calculating the input set through a deep extreme learning machine to obtain a final wind speed prediction result. The accuracy of the final prediction result and the efficiency of the overall prediction algorithm are effectively improved.

The present application relates to a kind of fA grade weak current signal software filtering method, electronic equipment and storage medium, filtering method includes the following steps: setting sampling parameter, calculate sampling point number preset target value;Start AD data acquisition, when sampling point number reaches preset target value, trigger data processing software interrupt, and carry out first-stage filtering processing;Call back secondary filtering processing function, carry out secondary filtering processing, obtain filtered voltage value, first-stage filtering uses arithmetic average filtering algorithm, secondary filtering uses recursive average filtering algorithm.The present application, by arithmetic average filtering algorithm superimposed recursive average filtering algorithm of two-stage software filtering algorithm can be sequentially effectively filtered out periodic and aperiodic interference in weak current signal, realize the effective extraction of system thermal noise in micro-current signal and filter out, and to the greatest extent, retain signal dispersion, make the fitting curve after filtering and actual curve further coincide, effectively improve measurement accuracy.

The invention relates to the technical field of power electronics, in particular to a diesel generator test system control method based on a moving average filter, and the method comprises the steps: collecting an output voltage instantaneous value of a diesel generator in real time, and obtaining a voltage sampling sequence; inputting the voltage sampling sequence into a pre-constructed moving average filter; performing moving average processing on the input voltage sampling sequence through a moving average filter to obtain an output signal; generating a current command signal based on an output signal of the moving average filter; and according to the current instruction signal, driving the diesel generator test system to simulate a load with a specific power factor through current closed-loop control, and carrying out load simulation test on the diesel generator. The moving average filter is directly used for generating a current instruction of the test system, and the lag angle of the current instruction is linearly adjusted by adjusting the point number of the filter through the linear phase-frequency characteristic of the filter, so that the accurate and convenient control of the simulation load power factor of the diesel generator test system is realized.

The application provides a dynamic zero point correction and gravity compensation method of a six-dimensional force sensor, which comprises the following steps: 1, determining a pose change matrix of a robot end coordinate system relative to a base coordinate system M 0 T1; 2, determining a coordinate transformation relationship between an end tool and a robot base coordinate; 3, determining force components of end tool gravity in three directions of a sensor coordinate system XYZ; 4, determining a gravity compensation value, and performing zero point correction and compensation error. The application can compensate the force of the end tool gravity on the sensor; after data processing by using an average filtering algorithm, zero point correction can be performed, so that the influence of external interference factors, operation errors and the like can be reduced; the application is suitable for gravity compensation and zero point correction under various postures and motion states of a robot, and is suitable for various types of mechanical arms, end tools and six-dimensional force sensors; and the application is suitable for zero point correction and gravity compensation in a static state and in a robot motion process.

The invention provides a method and system for improving the detection resolution of a digital closed-loop accelerometer. The method comprises the steps that a pendulous reed, a first amplifier, an integrator, a second amplifier, an ADC chip, a digital signal processor STM32, a DAC chip, a third amplifier and a torquer are sequentially connected; a median value average filtering algorithm is added in a digital signal processor STM32 to preprocess data collected by an ADC chip, a data point collected earliest in a queue is removed every time one ADC datum is collected, then the maximum value and the minimum value of N data including new data are removed, and the maximum value and the minimum value of the N data including the new data are removed; and calculating an arithmetic mean value of the N-2 data to obtain a final acquisition result. The integrator is used for integrating the system error, and the acceleration detection resolution is improved by inversely resolving the current through the voltage. According to the technical scheme, the technical problem that the resolution of a digital accelerometer in the prior art is limited by ADC and DAC devices and is low is solved.

The invention discloses a zero point measuring device of a Coriolis flowmeter and a zero point error compensation method. The device comprises a constant-temperature water tank, a heat preservation tank, a heater, an iron support, electric control ball valves, a Coriolis flow meter, a pipeline supporting frame, a variable-diameter clamping sleeve and an upper computer, and the zero point measurement and error compensation principle is that the upper computer issues a command to control opening and closing of the upstream electric control ball valve and the downstream electric control ball valve in a time-sharing mode, and after it is ensured that pipeline fluid is in a static state, the pipeline fluid is subjected to zero point measurement and error compensation. At the moment, the instantaneous flow collected by the flow meter is a zero value, N zero values are continuously collected according to the set sampling frequency f, the zero values are arranged in an ascending order through a rapid sorting algorithm, then zero correction is conducted through a median value average filtering algorithm, and therefore the compensated zero value Q0 of the Coriolis flow meter is obtained; and zero error compensation is carried out on the instantaneous flow Q1 and the accumulated flow Qm in the T time period. According to the method, the zero point of the Coriolis flowmeter is measured and the zero point error is corrected in a pure software mode, manual participation is not needed, and the zero point measurement efficiency and the flow measurement precision are improved.

A training method of an artificial intelligence model for optimizing operation variables of a circulating fluidized bed boiler is disclosed. The method may comprise the steps of: using an exponential moving average (EMA) filter and a time series decomposition method to preprocess sensing data of a circulating fluidized bed boiler; using the preprocessed sensing data so as to train a reinforcement learning algorithm; and using the learned reinforcement learning algorithm so as to output a recommended operation variable of the circulating fluidized bed boiler. The time series decomposition method can include a seasonal and trend decomposition using loess (STL) decomposition method. The step of preprocessing the sensing data can include a step of using the EMA filter to smooth the sensing data during a soot-blowing time of the circulating fluidized bed boiler.

The invention relates to the technical field of green building power supply management, and discloses a green building management system based on intelligent power supply, which comprises a data acquisition module, a preprocessing module, a correlation analysis module and a power supply management module, according to the invention, filtering cleaning, abnormity correction and normalization processing are carried out on environment and power utilization data through a moving average filtering algorithm, a 3 sigma principle and the like, so that the data accuracy and reliability are improved; a correlation coefficient model is used for analyzing the correlation between power consumption and the environment, and multi-level power supply priorities are divided according to correlation coefficients and environment threshold values, so that on-demand power supply is realized; when the power supply amount is dynamically adjusted, basic power consumption is calculated in combination with historical data and a real-time environment, distribution is performed according to priorities, and optimization is constrained according to the total power supply capacity, so that excessive or insufficient power supply is avoided, the energy efficiency is improved, and the energy consumption is reduced; the energy waste can be reduced, the operation cost and carbon emission are reduced, the operation of equipment in an extreme environment is ensured, and an efficient scheme is provided for intelligence and low carbon of green buildings.

The invention belongs to the technical field of face living body detection, and provides a face living body detection method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the average filtering and residual processing of an original image, and obtaining an accumulated noise feature and an average filtering image; determining residual features of the original image according to the average filtering image, performing spatial conversion on the original image to obtain I channel features and Q channel features, and fusing the accumulated noise features, the residual features of the original image, the I channel features and the Q channel features to obtain fused features; performing classification training on the fusion features by adopting MLP-Mixer to obtain a face living body detection model; and the target face image is detected through the face in-vivo detection model to obtain a face in-vivo detection result, and through the technical scheme of the invention, the accuracy of face detection is improved, and the detection cost is reduced.

The invention provides an industrial equipment fault prediction method and device, relates to the technical field of fault analysis, and solves the technical problem of insufficient fault prediction accuracy in the prior art. The method specifically comprises the following steps: acquiring equipment data; the equipment data comprises first data and second data; performing noise filtering on the third data through a moving average filtering algorithm; performing feature extraction on the third data after noise filtering to obtain equipment feature data; calculating a correlation coefficient of the equipment feature data and the equipment fault tag, and selecting and removing non-correlation features based on a threshold value; principal component analysis is carried out on the equipment feature data after the non-relevant features are removed, and low-dimensional equipment feature data are obtained; and inputting the low-dimensional equipment feature data into the dual-channel model for fault classification and fault prediction. The method is used for industrial equipment fault prediction.

Disclosed are a genetic algorithm optimization-based fuzzy Proportional-Integral-Derivative (PID) control method, apparatus, and device. The method includes: S1, acquiring a preset standard electrode plate quality, a measured electrode plate quality, and a system default ray intensity, acquiring an ambient temperature parameter, retrieving a monitoring index, and monitoring the stability of a surface density measurement system according to the monitoring index; S2, setting a data sampling interval according to a preset system parameter of the surface density measurement system; S3, according to the data sampling interval, continuously sampling N pieces of data using a median average filtering algorithm, removing extreme values, calculating an arithmetic mean value of the N-2 pieces of data, and eliminating a sampling value deviation of the arithmetic mean value to obtain surface density sampling data; S4, performing fuzzification processing on the surface density sampling data, determining a fuzzy rule, and performing optimization in real time according to the fuzzy rule and by means of fuzzy reasoning, to obtain applicable PID parameters; and S5, according to the applicable PID parameters, controlling a servo motor of the surface density measurement system using a genetic algorithm. Therefore, the technical problems of difficulty in parameter tuning and low robustness of closed-loop control are solved.

The invention relates to the technical field of intelligent adjustment control, in particular to a refrigeration device operation control method based on the temperature change trend, which comprises the following steps: acquiring the temperature of an evaporation heat exchange part and a condensation heat exchange part in a refrigeration cycle through a temperature sensor, and calculating the temperature difference; generating a refrigeration cycle temperature difference change trend curve through moving average filtering, extracting ascending and descending slopes, obtaining a refrigeration cycle load demand grade through a fuzzy algorithm, judging a threshold value, triggering an adjusting logic, generating a refrigeration unit operation adjusting instruction, and adjusting the rotation speed of a refrigeration compressor, the opening degree of a refrigeration throttling mechanism and the rotation speed of a fan in a grading manner. According to the method, the temperature difference between the evaporation end and the condensation end is monitored, moving average smooth calculation is carried out, heat exchange subtle changes are dynamically captured, loads are judged according to slope changes, refrigeration requirement recognition and trend prediction are achieved, progressive adjustment is triggered when the system deviates from a load interval, and the stability and energy efficiency of the system are improved.

This invention provides a system and method for real-time attitude monitoring during the floating loading and unloading process of a semi-submersible barge. The system includes a data acquisition unit, a data processing unit, and an attitude calculation unit. The data acquisition unit simultaneously collects the vertical distance from each monitoring point to the corresponding water surface using radar level gauges installed at the four corners of the semi-submersible barge's main deck, generating a raw height data sequence. The data processing unit performs two-stage cascade filtering on the raw data: first, a Savitzky-Golay filter is used for local polynomial least squares smoothing fitting, and then a moving average filter is used for secondary smoothing to output a stable height value. The attitude calculation unit calculates the real-time roll and pitch angles of the semi-submersible barge based on the stable height value. This invention effectively suppresses wave interference, improves the stability and accuracy of attitude monitoring, and has low hardware costs.

The application provides a tunnel health monitoring method based on distributed optical fiber sensing, and relates to the technical field of tunnel engineering health monitoring.The application adopts an analytic hierarchy process to perform risk assessment on a tunnel, sets a differentiated optical fiber layout scheme according to an assessment result, and lays distributed optical fiber sensors on an inner surface of tunnel lining; obtains original strain data through Brillouin optical time domain analysis, and performs pretreatment by using a sliding average filtering algorithm; builds a finite element model based on actual physical data of the tunnel, inverses strain data into full-field node displacement, and obtains complete strain tensors of finite element units; performs unit-level strain state evaluation, time sequence trend risk evaluation and spatial clustering risk evaluation, and obtains a comprehensive risk value; sets a dynamic evaluation threshold interval, and completes accurate regional health evaluation of the tunnel.The application realizes accurate monitoring and risk evaluation of the health state of the tunnel, and can be widely applied to the safety monitoring field of highway and railway tunnels.

The invention relates to the technical field of industrial measurement and sensing, in particular to a high-precision conductivity measurement system and method based on multi-layer sensors and temperature and pressure cooperative compensation, and adopts a differentiated path of'weak signal layer + strong signal layer 'to realize full-layer signal precise adaptation; a temperature and pressure interaction item and a quadratic item are introduced to construct a coupling model, and a dual-core architecture of main controller algorithm customization and secondary controller compensation execution is matched, so that the problems of temperature and pressure coupling drift and real-time performance are solved. Meanwhile, a 0.01% 1PPM precision impedance conversion module and dynamic gain conditioning are adopted, so that high-precision impedance matching between the sensor and a rear-end circuit is realized, and signal distortion is reduced; 2-6 signals of a coverage layer are polled through an analog switch period, an interlayer correction coefficient is introduced to adapt to interlayer differences, and the measurement consistency of the whole layer is improved; and finally, the device has the functions of fault self-diagnosis, sliding average filtering window adjustment, polling mode self-definition and the like, and the flexibility and reliability of the device are enhanced.

The application discloses a phase-locked loop performance optimization method and system based on a sliding average filter, and comprises the following steps: obtaining a relative reference phase change amount in real time through a phase-locked loop based on a sliding average filter; traversing each time, and predicting a power grid fundamental frequency positive sequence phase jump amount at a current traversal time according to the relative reference phase change amount at the current traversal time and the relative reference phase change amount at a first time before the current traversal time; in a time period after a response delay of the filter in a fault and a time period after the filter is removed from the fault and the system restores to a steady state, analyzing a phase compensation amount at each time according to the relative reference phase change amount and the power grid fundamental frequency positive sequence phase jump amount at each time, and compensating the phase of an output signal of the filter by using all the phase compensation amounts. The application considers the response delay of the filter in the case that the fault causes the power grid phase jump, so as to alleviate the problems of phase-locked loop disconnection and slow phase-locked speed.

The present disclosure relates to a signal processing method and device for vector control, electronic equipment and storage medium. The method comprises the following steps: obtaining an analog current signal; performing analog-digital conversion on the analog current signal to obtain a digital signal; passing the digital signal through an average filter to obtain an input signal; performing filter processing on the input signal based on a target filter corresponding to a target transfer function to obtain an output signal; and using the output signal for vector control. The embodiments of the present disclosure can improve control accuracy and stability on the basis of improving current sampling accuracy, and avoid the risk of motor out of control and other problems.

This invention discloses a method and device for interactive calligraphy and painting appreciation based on eye-tracking, relating to the field of virtual reality technology. The method includes: on a backend server, using an improved pigeon flocking optimization algorithm to optimize parameters, configuring and deploying the parameters of the moving average filter and interest evaluator for calligraphy and painting appreciation interaction; on an eye-tracking digital terminal, collecting real-time eye-tracking interaction data; mapping the real-time raw gaze point sequence to the texture coordinate system of the calligraphy and painting digital model, and filtering it using a moving average filter to obtain a real-time smooth gaze point sequence; based on the real-time smooth gaze point sequence and the real-time pupil diameter sequence, using an interest evaluator to assess interest and determine the interaction lock area and its cumulative duration; generating an interaction trigger command based on the cumulative duration, sending it to the backend server, and executing calligraphy and painting appreciation auxiliary behavior. This invention achieves a highly immersive, anti-interference, low-latency, and highly interactive calligraphy and painting appreciation experience.

The utility model discloses a sectional position correction circuit of a multi-magnetic-pole-pair motor, and relates to the technical field of magnetic field detection. Comprising a driver and an execution module, the driver is electrically connected with a power supply, a motor and a communication module, the communication module is electrically connected with a PC, and the driver performs information interaction with a magnetic encoder in the motor through an SEI TX port, an SEI RX port and an SEI VCC port; the communication module comprises a USB (Universal Serial Bus) connector and an RS485 communicator, the driver is electrically connected with the USB connector through the RS485 communicator, and the USB connector is electrically connected with a PC (Personal Computer). According to the utility model, the real-time position of the motor rotor is obtained and is combined with an extreme value removal average filtering algorithm, and each magnetic pole pair is corrected in a segmented manner, so that the problem of accumulative errors within a 360-degree mechanical angle caused by the traditional calibration of a single magnetic pole pair is solved, and the global position detection precision and the anti-interference capability are remarkably improved.

The invention discloses an error code prediction model construction method for LVDS link dynamic characteristics, and belongs to the field of chip test algorithms. Inputting the fault type experiment data with different characteristics into the dynamic model, and extracting and processing the data characteristics by the dynamic model; dividing the time sequence data into an input group and an output group by adopting a sliding window method; the bit error rate is placed in an interval of [0, 1], and the influence of numerical range difference on model training is eliminated; a moving average filter is adopted to carry out de-noising processing on error code noise of experimental data, and the weight of a real-time updating model is designed. In order to verify the actual effect of the dynamic error code prediction model and algorithm, various changing scenes are simulated, and prediction precision and adaptive capacity tests are carried out on the model. Experiments show that the prediction model improves the accuracy of error code prediction, and has obvious advantages compared with a static model.

Methods and systems for compensating for movement in a fluid supply source of a fluid management system. An illustrative fluid management system may comprise at least one load cell for measuring fluid supply bag weight, an analog-to-digital converter (ADC) interface, a processor, and a memory. The memory may store instructions that, when executed by the processor, cause the processor to: receive digital values from the ADC interface, apply a moving average filter over a predetermined time period, determine a deviation between a current sample value and a calculated mean of the filtered digital values, compare the determined deviation to an acceptable deviation threshold, and process the filtered digital values to calculate the fluid supply bag weight when the determined deviation is less than the acceptable deviation threshold.

The application provides a FIR low-pass digital filtering method for reducing resource consumption, which comprises the following steps: determining the filtering condition required when the filtering effects of the average filtering method and the FIR filtering method are consistent by comparing the filtering form of the average filtering method with the filtering form of the FIR filtering method; determining the filtering times according to the design parameters of the filtering effect; obtaining the filtering operation point number ratio according to the required filtering condition and the filtering times; filtering the to-be-filtered signal according to the filtering times by using the filtering operation point number ratio. The application uses the average filtering with twice optimal operation point number ratio continuously, can effectively inhibit the influence of various interference signals in the data acquisition process, improves the filtering effect, greatly reduces the resource consumption, can perform high-performance calculation under the condition of consuming extremely low calculation resource, and is convenient for designers to filter.

The application discloses a kind of non-contact human respiratory rate measurement methods based on video and frequency modulation continuous wave radar information fusion, comprising:1 respectively processes video and radar bimodal data to obtain pixel motion trajectory and distance angle chart time sequence;2 video measurement part uses spectral subtraction and principal component analysis technique to carry out pretreatment, radar measurement part uses static clutter removal and average filtering technique to carry out pretreatment, two kinds of modal all use empirical mode decomposition technique to carry out single modal measurement;3 feature level fusion, after bimodal signal using multivariate singular spectrum analysis extraction shared respiratory signal is pretreated;4 decision level fusion, according to the signal-to-noise ratio weighted frequency domain estimation value summation of feature level fusion result and single modal measurement result obtains final measurement result.The application can measure respiration under a variety of spontaneous motion scene, measurement result is compared with single modal method and has significant improvement, can effectively expand the application range of video and radar fusion measurement.

The invention discloses an FPGA-based weighted moving average filter and a storage method thereof. The FPGA-based weighted moving average filter comprises an ad acquisition control module, a data calibration module, a data caching module, a moving average filtering operation module, a weighted operation module and a dual-port RAM storage module which are connected in sequence, the device further comprises a ZYNQ main control module and a clock generation module, the ZYNQ main control module is sequentially connected with the clock generation module and the data caching module, and the clock generation module is connected with the ad acquisition control module. A weighted moving average filtering mode is adopted, weighting is set, so that an acquired data curve is smooth, data fluctuation interference is effectively restrained, newest changes of signals can be better reflected, the sensitivity of a system is improved, and meanwhile the sensitivity to noise can be possibly improved.

The invention belongs to the technical field of satellite analysis, and particularly relates to a time sequence characteristic analysis method and system based on GNSS orbit determination multi-source anomaly, and the method comprises the following steps: S1, carrying out the cleaning and arrangement of GNSS real-time orbit determination data, calculating the residual error between an observation value and a model prediction value, and carrying out the moving average filtering processing; s2, implementing anomaly detection by adopting a sliding window Z-score method, and identifying potential abnormal event points from the residual sequence; s3, for each abnormal event point, extracting a multi-dimensional time sequence feature from a specific time window, and carrying out standardization processing on the multi-dimensional time sequence feature; s4, organizing the extracted features into feature vectors, and storing the feature vectors together with corresponding type labels and metadata into a database; s5, performing similarity calculation on the feature vector of the current newly detected abnormal event point and a sample in the feature library by adopting a nearest neighbor classification method; and S6, outputting a classification result of the current newly detected abnormal event point, and performing incremental updating on the feature library.