CN-121983272-A - Multi-mode monitoring method, system, equipment and storage medium based on hardware cooperation

CN121983272ACN 121983272 ACN121983272 ACN 121983272ACN-121983272-A

Abstract

The invention relates to the field of medical monitoring and discloses a multimode monitoring method, a system, equipment and a storage medium based on hardware cooperation, wherein the method comprises the following steps of quickly starting core monitoring after power-on to support hot plug and plug-and-play access of medical equipment; the method comprises the steps of multi-source access, clock synchronization and resampling to construct a unified time axis, support break-link merging, motion characteristic driving self-adaptive anti-interference, abnormal authenticity identification, low-quality channel intelligent judgment and shielding based on redundant observation and space-time alignment, weighted fusion parameters through consistency analysis, event trigger point capturing, holographic structured event package packaging and bidirectional indexing, full time sequence evolution view construction, correlation attribution and trend early warning execution, and multiple disk feedback optimization algorithm playback. The invention solves the problems of low data quality, unexplained alarm and missing trace under extreme dynamic environment, and meets the requirements of pre-hospital emergency high-quality monitoring and quick decision.

Inventors

LIN PINGPING
XIE XU
Zhan Lijia
CHEN WEIJIA
CAI ZEHANG

Assignees

汕头市超声仪器研究所股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260403

Claims (10)

1. The multimode monitoring method based on hardware cooperation is characterized by comprising the following steps of: after the system is powered on, self-checking initialization is completed, core physical sign acquisition is started preferentially, a basic monitoring view is built, hot plug and plug-and-play access of the image diagnosis equipment is supported, and seamless fusion of data is realized; The multi-interface sink node is connected to the heterogeneous medical equipment in parallel, analyzes the standardized data frames for unified management, utilizes the hardware clock to execute time base giving and transmission delay compensation, establishes a unified time axis, accurately maps the multi-source data to the same continuous time axis through clock deviation correction and resampling technology, and automatically executes seamless merging of the data after the equipment is disconnected and recovered; The generated motion labels are shared to a signal processing module in real time, a self-adaptive anti-interference algorithm is driven to dynamically switch a filtering strategy, and the motion characteristics are utilized to identify the authenticity of physiological abnormality; Carrying out real-time comparison and confidence judgment on multi-source data by utilizing a dynamic consistency analysis strategy to obtain a comparison result and a confidence judgment result, and carrying out weighted fusion on the multi-physical quantity observation values according to the comparison result and the confidence judgment result to obtain physiological parameters; when capturing an event trigger point, locking full waveforms, images and metadata in a dynamic time window, packaging the full waveforms, images and metadata into a non-tamperable multidimensional holographic structured event package, and establishing bidirectional index association; The method comprises the steps of constructing a full-time-sequence clinical evolution view based on holographic data comprising the multi-dimensional holographic structured event package, executing multi-dimensional data time-sequence association attribution, accurately discriminating pathological changes and environmental interference, implementing macroscopic trend early warning and stage evaluation, realizing upgrading from instantaneous warning to continuous perception, enabling full-process quality control and multi-disc through a visual playback tool, and feeding back an analysis result to a front-end algorithm to optimize an anti-interference and judging strategy.
2. The method for multi-mode monitoring based on hardware collaboration according to claim 1, wherein the system completes self-checking initialization after power-on, preferentially starts core physical sign acquisition and builds a basic monitoring view, and simultaneously supports hot plug and plug-and-play access of an image diagnosis device to realize seamless data fusion, comprising: automatically completing self-checking of an edge terminal and initializing a core module after the system is powered on; The method comprises the steps of preferentially driving a quick response sensor, and immediately starting continuous data acquisition and real-time calculation of key parameters at the moment of signal access, wherein the quick response sensor comprises an electrocardio sensor, an oximetry sensor and/or a respiration sensor; based on the ready core vital sign data, rendering a dynamic monitoring curve in real time and starting a basic data recording service, and quickly constructing a visualized basic monitoring view; In basic monitoring operation, the on-demand dynamic access of the image diagnosis equipment is supported, and equipment identification and time sequence association are automatically completed.
3. The method of claim 1, wherein the multi-mode monitoring method based on hardware collaboration is characterized in that heterogeneous medical equipment is accessed in parallel through a multi-interface sink node, standardized data frames are analyzed to be managed in a unified mode, a unified time axis is established by utilizing a hardware clock to execute time base giving and transmission delay compensation, multi-source data are precisely mapped to the same continuous time axis through clock deviation correction and resampling technology, and seamless merging of the data is automatically executed after equipment disconnection and connection are recovered, and the method comprises the following steps: Configuring an edge terminal as a multi-interface sink node, accessing a plurality of medical sensors and devices in parallel through a wired link and/or a wireless link, and establishing a multi-channel data input environment; Defining and analyzing a predefined data frame containing equipment identification, data type, acquisition sequence number and original load, and realizing unified identification and standardized management of heterogeneous data of different sources; based on a built-in hardware clock, performing receiving stamping at the moment of data physical arrival, and performing reverse compensation calculation by combining a preset interface transmission delay model to generate an accurate acquisition time stamp; correcting deviation among devices through periodic clock synchronization handshake, strictly sequencing data by using a global continuous time axis queue, and solving the problem of time alignment of multi-frequency data by adopting interpolation or resampling technology; detecting continuity of the acquisition sequence number in real time, automatically marking packet loss or abnormal event, and starting a local caching mechanism when equipment communication is interrupted, so as to ensure that data in the disconnection period is not lost; After the equipment is reconnected, the buffer data is seamlessly integrated to the global time axis according to the time stamp sequence, so that the continuity and time sequence consistency of the multi-mode monitoring data in the whole life cycle are ensured.
4. The method for multi-modal monitoring based on hardware cooperation according to claim 1, wherein the steps of extracting multi-dimensional motion characteristics based on the unified time axis and identifying the state of a transportation environment, constructing a physiological motion coupling data system, sharing the generated motion labels to a signal processing module in real time, driving a self-adaptive anti-interference algorithm to dynamically switch a filtering strategy, and identifying the authenticity of physiological anomalies by utilizing the motion characteristics comprise: the method comprises the steps of collecting original data by using an inertial measurement unit, extracting multidimensional features comprising time domain, frequency domain and gesture by using a sliding window algorithm, and judging gesture changes and motion modes in real time by combining a mode identification technology; Converting the identified motion state into independent modal data, synchronously fusing the motion state data and vital sign data based on a uniform time stamp, generating a physiological motion coupling data stream with a motion state label and a quality score, and extracting motion context information from the physiological motion coupling data stream; automatically starting adaptive filtering and baseline drift suppression algorithm under motion interference, or reducing weight of a specific sensor and delaying intermittent measurement to remove motion artifact and prevent false alarm; A bidirectional association mechanism of a motion mode and clinical events is established, physiological abnormality causes are identified by utilizing motion characteristics in an auxiliary mode, and accordingly alarm suppression, degradation or special critical alarm triggering are executed.
5. The multi-mode monitoring method based on hardware collaboration according to claim 1, wherein the multi-physical quantity observation values of the same physiological index are obtained in parallel through a heterogeneous redundant observation framework, the multi-physical quantity observation values are aligned in time and space in combination with the motion labels, the multi-source data are compared in real time and judged by confidence coefficient by utilizing a dynamic consistency analysis strategy to obtain a comparison result and a confidence coefficient judgment result, the multi-physical quantity observation values are subjected to weighted fusion according to the comparison result and the confidence coefficient judgment result to obtain physiological parameters, when an abnormality occurs in the comparison result, an abnormal intelligent judgment mechanism is started, a source of interference is traced, and a low-quality channel is dynamically shielded, and the method comprises the steps of: configuring multi-source sensing channels based on different physical sensing principles, and monitoring the same physiological index in parallel; operating a multi-mode signal processing algorithm in parallel, extracting characteristic parameters from different sensing channels respectively, and uniformly mapping all calculation results to a global continuous time axis to form a multi-dimensional parameter vector for comparison; Calculating the deviation of the multisource observation values in real time and judging consistency by applying a self-adaptive dynamic threshold value; for consistent data, improving the quality level of the consistent data, and carrying out fusion output by adopting a weighted average algorithm based on a signal to noise ratio so as to reduce random errors and improve measurement accuracy; When the difference of the multi-source data exceeds the limit, an abnormal decision mechanism is triggered, and the low-reliability channel is automatically marked, the alarm weight is dynamically adjusted or false alarm is restrained by combining the motion context tracing signal quality.
6. The method of claim 1, wherein the deploying the multi-dimensional event monitoring engine receives the physiological parameter anomaly signal and the confidence level determination result in real time, locks full waveforms, images and metadata in a dynamic time window when an event trigger point is captured, encapsulates the full waveforms, images and metadata into a non-tamperable multi-dimensional holographic structured event package, and establishes a bi-directional index association, and comprises: Based on a unified hardware clock reference, assigning millisecond-level time stamps to all acquired multi-mode data, and additionally writing the multi-mode data into a storage medium according to a strict time sequence to form a continuous data stream retaining a complete time sequence topological relation; Scanning key physiological parameters and system states in real time, and accurately positioning and marking event trigger points according to multiple trigger conditions, wherein the multiple trigger conditions comprise at least one of threshold crossing, nonlinear trend deterioration, multi-mode consistency abnormality and specific clinical operation instructions; At the moment of event triggering, according to the event type, the dynamic time window data of the front and rear time periods are adaptively intercepted, and are copied from a circulating buffer zone to a special read-only event protection zone; Integrating full waveform original data, high frequency parameter trend, image key frame/file, environment motion label and operation metadata, encapsulating the data into an atomized structured event package containing unique ID and index pointer according to a predefined data structure, and safely storing the event package; A bidirectional indexing mechanism of an event package and an original continuous time sequence data stream is constructed in a database, so that the original data position is supported to be rapidly positioned through an event ID, or an event mark point is highlighted when the data stream is played back, and rapid retrieval under mass data is realized; and synchronously reproducing vital sign waveforms, video images, motion gestures and/or operation logs based on the unified time axis.
7. The hardware collaboration-based multi-modal monitoring method according to claim 1, wherein the constructing a full-time clinical evolution view based on holographic data comprising the multi-dimensional holographic structured event package, performing multi-dimensional data time-series association attribution, accurately discriminating pathological changes and environmental disturbances, implementing macroscopic trend early warning and stage assessment, enabling upgrading from instantaneous warning to continuous perception, enabling full-flow quality control and duplication by a visual playback tool, and feeding back analysis results to a front-end algorithm to optimize anti-interference and arbitration strategies, comprises: discrete physiological parameters, imaging evidence and medical operation logs on a continuous time axis are integrated, single-point numerical limitation is broken, and coherent events of a patient in the whole transportation process are reconstructed; automatically analyzing the time delay and correlation between a particular medical procedure and a subsequent physiological response to assess treatment effectiveness and couple the environmental disturbance event with vital sign fluctuations; And (3) identifying gradual change trend which is not out of range but continuously worsened based on long-term data fitting and early warning, dividing the whole flow into different treatment stages, and quantitatively counting parameter stability and event frequency of each stage to form full-line data.
8. A multi-modality monitoring system based on hardware collaboration, comprising: the rapid starting monitoring module is used for completing self-checking initialization after the system is powered on, preferentially starting core physical sign acquisition and constructing a basic monitoring view, and simultaneously supporting hot plug and plug-and-play access of the image diagnosis equipment to realize seamless data fusion; The system comprises a hardware coordination and time alignment module, a time base giving and transmission delay compensation module, a data seamless merging module, a data transmission module and a data transmission module, wherein the hardware coordination and time alignment module is used for accessing heterogeneous medical equipment through a multi-interface sink node in parallel, analyzing a standardized data frame to be managed in a unified way, utilizing a hardware clock to execute time base giving and transmission delay compensation to establish a unified time axis, further precisely mapping multi-source data to the same continuous time axis through a clock deviation correction and resampling technology, and automatically executing data seamless merging after equipment disconnection and recovery; The motion interference identification module is used for extracting multidimensional motion characteristics based on the unified time axis, identifying the state of a transportation environment and constructing a physiological motion coupling data system, sharing the generated motion labels to the signal processing module in real time, driving the self-adaptive anti-interference algorithm to dynamically switch the filtering strategy, and identifying the authenticity of physiological abnormality by utilizing the motion characteristics; The multi-mode cross validation module is used for parallelly acquiring multi-physical quantity observation values of the same physiological index through a heterogeneous redundant observation framework, carrying out space-time alignment by combining the motion labels, carrying out real-time comparison and confidence judgment on multi-source data by utilizing a dynamic consistency analysis strategy to obtain comparison results and confidence judgment results, and carrying out weighted fusion on the multi-physical quantity observation values according to the comparison results and the confidence judgment results to obtain physiological parameters; The traceable event recording module is used for receiving the physiological parameter abnormal signal and the confidence coefficient judgment result in real time by deploying a multidimensional event monitoring engine, locking full waveforms, images and metadata in a dynamic time window when an event trigger point is captured, packaging the full waveforms, images and metadata into a non-tamperable multidimensional holographic structured event package, and establishing bidirectional index association; The clinical analysis and auxiliary decision module is used for constructing a full-time sequence clinical evolution view based on holographic data comprising the multi-dimensional holographic structured event package, executing multi-dimensional data time sequence association attribution, accurately discriminating pathological changes and environmental interference, implementing macroscopic trend early warning and stage evaluation, realizing the upgrade from instantaneous warning to continuous perception, enabling full-flow quality control and multi-disc through a visual playback tool, and feeding back an analysis result to a front-end algorithm to optimize an anti-interference and decision strategy.
9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the hardware collaboration-based multimodal monitoring method according to any of claims 1 to 7 when executing the computer program.
10. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the hardware collaboration-based multimodal monitoring method of any of claims 1-7.

Description

Multi-mode monitoring method, system, equipment and storage medium based on hardware cooperation Technical Field The invention relates to the field of medical monitoring, in particular to a multi-mode monitoring method, a system, equipment and a storage medium based on hardware cooperation. Background Pre-hospital emergency is a key link of medical rescue, emphasizing "platinum ten minutes" and "gold one hour". The existing portable equipment can collect electrocardio, blood oxygen and the like, but lacks a hardware-level motion artifact resisting mechanism under extreme dynamic environments such as open-air no-network, stretcher lifting, vehicle jolting and the like, so that original data is easy to be polluted by noise, high false alarm rate is caused by unstable signals, and clinical judgment is seriously interfered. Existing intelligent diagnostic techniques such as multi-modality AI analysis (CN 120913802 a), multi-dimensional health status assessment (CN 120998475A) and blockchain forensics (CN 121034585A) perform well in static or networked environments, but are difficult to adapt to pre-hospital scenarios. Meanwhile, the block chain consensus delay causes decision delay, the centralized storage has a falsification risk, the data in the pre-hospital is cracked, and the illness state evolution track is lost in the transfer process. In summary, the prior art has significant limitations in terms of data quality under dynamic interference, real-time decision response, and full-flow security traceability. The lack of a contextual evidence chain for alarm results in unexplained, delayed AI-assisted diagnosis response, and inability to meet the urgent need for high quality monitoring and fast trusted decisions in emergency scenarios. Disclosure of Invention The embodiment of the invention provides a multi-mode monitoring method, a system, computer equipment and a storage medium based on hardware cooperation, which are used for solving the problems that the prior art has low data quality, unexplained alarm, AI decision delay and full-flow safety traceability due to lack of an anti-interference mechanism and a high-efficiency cooperation framework in an extreme dynamic environment, and is difficult to meet urgent requirements of pre-hospital first aid on high-quality monitoring and quick trusted decision. A multimode monitoring method based on hardware cooperation comprises the following steps: after the system is powered on, self-checking initialization is completed, core physical sign acquisition is started preferentially, a basic monitoring view is built, hot plug and plug-and-play access of the image diagnosis equipment is supported, and seamless fusion of data is realized; The multi-interface sink node is connected to the heterogeneous medical equipment in parallel, analyzes the standardized data frames for unified management, utilizes the hardware clock to execute time base giving and transmission delay compensation, establishes a unified time axis, accurately maps the multi-source data to the same continuous time axis through clock deviation correction and resampling technology, and automatically executes seamless merging of the data after the equipment is disconnected and recovered; The generated motion labels are shared to a signal processing module in real time, a self-adaptive anti-interference algorithm is driven to dynamically switch a filtering strategy, and the motion characteristics are utilized to identify the authenticity of physiological abnormality; Carrying out real-time comparison and confidence judgment on multi-source data by utilizing a dynamic consistency analysis strategy to obtain a comparison result and a confidence judgment result, and carrying out weighted fusion on the multi-physical quantity observation values according to the comparison result and the confidence judgment result to obtain physiological parameters; when capturing an event trigger point, locking full waveforms, images and metadata in a dynamic time window, packaging the full waveforms, images and metadata into a non-tamperable multidimensional holographic structured event package, and establishing bidirectional index association; The method comprises the steps of constructing a full-time-sequence clinical evolution view based on holographic data comprising the multi-dimensional holographic structured event package, executing multi-dimensional data time-sequence association attribution, accurately discriminating pathological changes and environmental interference, implementing macroscopic trend early warning and stage evaluation, realizing upgrading from instantaneous warning to continuous perception, enabling full-process quality control and multi-disc through a visual playback tool, and feeding back an analysis result to a front-end algorithm to optimize an anti-interference and judging strategy. Optionally, the system completes self-checking initialization after power-on, preferentially starts core physical si