CN-121987159-A - Brain intensive care data abnormality identification method and system

CN121987159ACN 121987159 ACN121987159 ACN 121987159ACN-121987159-A

Abstract

The invention relates to the technical field of medical data processing, in particular to a brain intensive care data abnormality identification method and system, comprising the following steps: the temperature sensor is used for collecting analog voltage signals, a Kalman filtering algorithm is called to perform denoising and analog-to-digital conversion to generate a real-time nose temperature value, a target reference temperature and an allowable fluctuation range are obtained, a temperature deviation value is calculated and generated with the real-time nose temperature value, if the temperature deviation value is larger than the allowable fluctuation range, a body temperature abnormal state mark is generated, the pixel color of a display interface is adjusted to be a red RGB (red, green and blue) coding value in response to the mark, a PID (proportion integration differentiation) control algorithm is called to perform feedback control on the real-time nose temperature value and the target reference temperature to generate a cooling instruction. According to the invention, the real-time nose temperature is obtained through Kalman filtering denoising, the abnormal visual early warning and PID closed-loop control are realized based on temperature deviation, the noise interference is effectively eliminated, and the monitoring data identification accuracy and the clinical response speed are improved.

Inventors

QU CHUNHONG
HUANG XUHUA
CHEN WEIHONG
Jin Wangyan
GAO FUYANG

Assignees

杭州市萧山区第一人民医院

Dates

Publication Date: 20260508
Application Date: 20260123

Claims (10)

1. The brain intensive care data abnormality identification method is characterized by comprising the following steps of: S1, acquiring an analog voltage signal through a temperature sensor positioned in an intensive brain monitoring environment, and calling a Kalman filtering algorithm to execute denoising processing and analog-to-digital conversion calculation on the analog voltage signal to generate a real-time nose temperature value; s2, acquiring a preset target reference temperature and an allowable fluctuation range, and performing numerical difference calculation on the real-time nose temperature value and the target reference temperature to generate a temperature deviation value; S3, executing the numerical comparison judgment of the temperature deviation numerical value and the allowable fluctuation range, and generating a body temperature abnormal state mark if the temperature deviation numerical value is larger than the allowable fluctuation range; and S4, in response to the body temperature abnormal state mark, adjusting the pixel color attribute of the display interface of the monitoring terminal to be a red RGB (red, green and blue) coding value, calling a PID (proportion integration differentiation) control algorithm to execute feedback control operation on the real-time nose temperature value and the target reference temperature, and generating a cooling equipment control instruction.
2. The method for identifying abnormal brain critical care data according to claim 1, wherein the step of S1 specifically comprises: S11, starting a precise thermistor sensor positioned in the deep part of the nasal cavity of the brain, continuously capturing weak potential change generated along with the intracranial heat dissipation of a human body at a preset kilohertz sampling rate, and performing gain adjustment and band-pass filtering processing on the captured original electric signal by using a preposed low-noise amplifier so as to remove power frequency interference and retain effective frequency band information, so as to generate an analog voltage signal capable of reflecting instantaneous temperature change; S12, constructing a state space model based on a linear random system, acquiring a system state estimated value and a covariance matrix at the last moment, combining the analog voltage signal at the current moment as an observation input, calculating a Kalman gain coefficient by using a recursive least square method, and carrying out iterative correction on the system state, thereby filtering the random interference of environmental electromagnetic noise on the signal and generating purified voltage data subjected to smooth denoising treatment; s13, invoking a high-precision analog-to-digital converter to execute discretization quantization coding operation on the purified voltage data, and according to a voltage-temperature characteristic curve calibrated by a blackbody radiation source in advance, linearly mapping and converting the discretization digital voltage value into a physical quantity which accords with the definition of an international temperature standard ITS-90 to generate a real-time nose temperature value for subsequent medical analysis.
3. The method for identifying abnormal brain critical care data according to claim 1, wherein the step of S2 specifically comprises: S21, accessing an electronic medical record database of a hospital information management system through an encryption communication protocol, retrieving historical physiological parameter records of a patient and current brain protection treatment scheme setting, extracting an ideal temperature baseline required by brain metabolism rate control, and generating a target reference temperature and an allowable fluctuation range according to a safe floating interval determined by a clinical low-temperature treatment protocol; S22, acquiring the real-time nose temperature value generated in the S1 and the target reference temperature generated in the S21, calculating an algebraic difference value between the real-time nose temperature value and the target reference temperature by using a high-precision subtracter, and performing absolute value operation on the algebraic difference value to eliminate the directional influence of temperature drift and generate a temperature deviation value representing the deviation of the current body temperature from an ideal baseline; s23, carrying out integral accumulation calculation based on a time dimension on the temperature deviation value, constructing a trend feature vector capable of reflecting the body temperature deviation duration and accumulated energy effect, and storing the feature vector and the current temperature deviation value in a correlated way to generate a comprehensive deviation data set for assisting in the subsequent abnormal severity judgment.
4. The method for identifying abnormal brain critical care data according to claim 1, wherein the step S3 specifically comprises: S31, acquiring the temperature deviation value and the allowable fluctuation range generated in the step S2, executing instantaneous value comparison operation by using a digital amplitude comparator, judging whether the current deviation amplitude exceeds a preset safety boundary on a physical value, and triggering a primary warning logic if the current deviation amplitude exceeds the preset safety boundary, so as to generate a preliminary out-of-limit trigger signal; S32, responding to the out-of-limit trigger signal, starting a time sliding window checking mechanism, counting the sampling point duty ratio of the temperature deviation value continuously exceeding the allowable fluctuation range in a preset sliding time window, and only when the duty ratio exceeds a preset confidence coefficient threshold value, confirming that the current state is non-sporadic pathological body temperature fluctuation and generating an abnormal confirmation signal; S33, calling an alarm control bit in a system state register based on the abnormality confirmation signal, forcibly switching the logic state from a normal monitoring mode to an emergency intervention mode, recording accurate time stamp and deviation peak value data of abnormality occurrence, and generating a body temperature abnormal state mark capable of triggering subsequent audible and visual alarm and equipment control logic.
5. The method for identifying abnormal brain critical care data according to claim 1, wherein the step S4 specifically comprises: S41, responding to the abnormal body temperature state mark, calling a rendering engine of a graphical user interface of a monitoring terminal, retrieving a UI control handle for displaying the current body temperature value, modifying pixel filling parameters of a control background area into red RGB (red, green and blue) coding values with high warning degree, synchronously triggering a high-frequency flicker effect, and generating a visual warning picture; S42, acquiring the real-time nose temperature value as a process variable, acquiring the target reference temperature as a set point, calculating a control error between the two values, respectively calculating a proportional term, an integral term and a differential term of the error, and synthesizing a control output quantity by using a weighted summation algorithm to realize quick response to temperature change and steady-state precision control and generate a PID operation result; And S43, generating a corresponding pulse width modulation signal according to the PID operation result, transmitting the signal to a power control unit of an externally connected medical cooling blanket or head cooling instrument through a hardware driving interface, and regulating the duty cycle and the circulating liquid flow rate of the refrigeration unit to generate a cooling equipment control instruction for executing physical cooling operation.
6. The brain critical care data abnormality identification method according to claim 2, characterized in that the purge voltage data generation process of S12 specifically includes: Constructing a state transition equation and an observation equation for describing the dynamic characteristics of the temperature sensor, and calculating the Kalman gain Using the analog voltage signal Updating a priori state estimates According to the formula Calculating a posterior state estimated value, wherein the estimated value is the purifying voltage data; Wherein, the Representative of the first The purge voltage data after the time passes the optimal estimation, Representing the first time state prediction based on the previous time The time of day a priori voltage estimate, Representing said kalman gain for balancing the prediction error covariance with the measurement noise covariance, Represents the first The analog voltage signal actually collected by the time sensor, Representing an observation matrix mapping the state space to the observation space.
7. The brain critical care data abnormality identification method according to claim 3, characterized in that the target reference temperature generation process of S21 specifically includes: Acquiring historical body temperature sequence data of a patient in twenty four hours in the past, removing outlier noise points in the sequence by using a quartile range method, calculating an arithmetic average value of the remaining effective data to be used as a basic physiological line, correcting the basic physiological line by combining an induced low temperature coefficient corresponding to a cerebral edema treatment level, generating a dynamic self-adaptive reference value capable of adapting to individual metabolic differences of the patient, and establishing the value as the target reference temperature.
8. The method for identifying brain critical care data abnormalities according to claim 5, characterized in that said PID operation result generation process of S42 specifically comprises: Calculating an error between the real-time nose temperature value at the current moment and the target reference temperature Discretizing the error according to the formula Calculating a control output quantity; Wherein, the Represents the first The PID operation results generated at the moment, Representing the temperature control deviation at the current moment, Representing a scaling factor for responding to the current error magnitude, Representing the integral coefficient used to eliminate steady state errors, Representing the differential coefficient for predicting the error variation trend and suppressing overshoot, Represents the total amount of accumulated error from the control start time to the current time, Representing the rate of change of error over time.
9. The method for identifying abnormalities in brain critical care data according to claim 4, wherein said abnormality confirmation signal generation process of S32 specifically comprises: And immediately starting a sliding time window with the length of thirty seconds after the single out-of-limit is detected, calculating the first derivative of the temperature deviation value corresponding to all sampling points in the window, and if the average change slope of the data in the window shows a forward divergence trend and exceeds a preset deterioration rate threshold, judging that the current condition is acute body temperature out-of-control, and eliminating false fluctuation caused by poor contact of a sensor, thereby generating the abnormal confirmation signal.
10. A brain critical care data anomaly identification system for implementing the brain critical care data anomaly identification method of any one of claims 1 to 9, the system comprising: The signal acquisition and filtering processing module is used for acquiring an analog voltage signal through a temperature sensor positioned in the brain intensive care environment, calling a Kalman filtering algorithm to execute denoising processing and analog-to-digital conversion calculation on the analog voltage signal, and generating a real-time nose temperature value; the temperature deviation analysis and calculation module is used for acquiring a preset target reference temperature and an allowable fluctuation range, performing numerical difference calculation on the real-time nose temperature value and the target reference temperature, and generating a temperature deviation value; the abnormal state judgment and identification module is used for executing the numerical comparison judgment of the temperature deviation numerical value and the allowable fluctuation range, and generating a body temperature abnormal state mark if the temperature deviation numerical value is larger than the allowable fluctuation range; And the feedback control and interaction module is used for responding to the body temperature abnormal state mark, adjusting the pixel color attribute of the display interface of the monitoring terminal to be a red RGB (red, green and blue) coding value, calling a PID (proportion integration differentiation) control algorithm to execute feedback control operation on the real-time nose temperature value and the target reference temperature, and generating a cooling equipment control instruction.

Description

Brain intensive care data abnormality identification method and system Technical Field The invention relates to the technical field of medical data processing, in particular to a brain intensive care data abnormality identification method and system. Background The technical field of medical data processing mainly relates to the technical category of digitally acquiring, storing, transmitting and analyzing physiological parameters, image data and medical record texts generated in the clinical medical process by utilizing a computer hardware and software system, wherein the traditional brain intensive care data abnormality identification method is characterized in that a multi-parameter monitor arranged at the bedside of an intensive care unit is connected with an intracranial pressure sensor or an electroencephalogram electrode slice to acquire brain physiological signal values of a patient in real time, medical staff manually inputs upper and lower limit thresholds of safety ranges of various physiological indexes on a control panel of monitoring equipment, and when the real-time values acquired by the sensor exceed a preset threshold interval, the monitor triggers an internal buzzer to send out sound alarm and flash an indicator lamp, or the medical staff regularly patrols and manually refers to vital sign data curves on a paper care record list to compare the fluctuation condition of the values. The traditional brain intensive care data anomaly identification relies on medical staff to manually set upper and lower limit thresholds of a physiological index safety range, the fixed threshold judgment mode is difficult to adapt to dynamic change of a physiological state of a patient, and a sensor collects original signals, often mixes environment noise interference, is extremely easy to cause false alarm or missing report phenomenon by directly comparing the original signals, relies on audible and visual alarm or manual timing to consult a paper record list, has information acquisition hysteresis, cannot realize real-time accurate monitoring and automatic closed loop control of the body temperature state of the patient, leads to low clinical response efficiency, increases the workload of medical staff, and is difficult to meet the severe requirements of brain intensive care on the timeliness and accuracy of data processing. Disclosure of Invention The invention aims to solve the defects in the prior art and provides a brain intensive care data abnormality identification method and system. In order to achieve the above purpose, the invention adopts the following technical scheme that the method for identifying the abnormality of the brain intensive care data comprises the following steps: s1, acquiring an analog voltage signal through a temperature sensor positioned in a brain intensive care environment, and calling a Kalman filtering algorithm to perform denoising processing and analog-to-digital conversion calculation on the analog voltage signal to generate a real-time nose temperature value; S2, acquiring a preset target reference temperature and an allowable fluctuation range, and performing numerical difference calculation on the real-time nose temperature value and the target reference temperature to generate a temperature deviation value; S3, performing numerical comparison judgment of the temperature deviation numerical value and the allowable fluctuation range, and generating a body temperature abnormal state mark if the temperature deviation numerical value is larger than the allowable fluctuation range; And S4, in response to the abnormal body temperature state mark, adjusting the pixel color attribute of the display interface of the monitoring terminal to be a red RGB (red, green and blue) coding value, calling a PID (proportion integration differentiation) control algorithm to execute feedback control operation on the real-time nose temperature value and the target reference temperature, and generating a cooling equipment control instruction. As a further scheme of the invention, the step S1 specifically comprises the following steps: S11, starting a precise thermistor sensor positioned in the deep part of the nasal cavity of the brain, continuously capturing weak potential change generated along with the intracranial heat dissipation of a human body at a preset kilohertz sampling rate, and performing gain adjustment and band-pass filtering processing on the captured original electric signal by using a preposed low-noise amplifier so as to remove power frequency interference and retain effective frequency band information, so as to generate an analog voltage signal capable of reflecting instantaneous temperature change; S12, constructing a state space model based on a linear random system, acquiring a system state estimated value and a covariance matrix at the last moment, combining an analog voltage signal at the current moment as an observation input, calculating a Kalman gain coefficient by u