A thrombus online detection system and method based on multi-frequency microwave dielectric trajectory and intelligent discrimination

Abstract

The present application relates to a kind of thrombus online detection system and method based on multi-frequency microwave dielectric trajectory and intelligent discrimination, belong to microwave biological detection and blood online monitoring technical field.It includes: extracorporeal circulation pipeline, multi-frequency microwave transmitting and receiving module, multi-frequency microwave detection unit, difference reference unit, intelligent discrimination module.The present application breaks through the limitation of existing detection technology in adaptability, speed, precision and information dimension, provides new technical scheme for constructing high sensitivity, multi-parameter joint, integrated extracorporeal circulation thrombus real-time detection platform.

Description

Technical Field

[0001] This invention relates to an online detection system and method based on multi-frequency microwave dielectric trajectory and intelligent discrimination, belonging to the field of microwave biological detection and online blood monitoring technology. Background Technology

[0002] As extracorporeal circulation support technology, blood management methods, and microwave biodetection technology are all in a phase of rapid development, and the demand for highly sensitive, non-contact online thrombosis monitoring is increasing, traditional detection methods based on pressure, optics, or single impedance have encountered new and prominent bottlenecks in terms of detection speed, information dimension, and system integration capabilities. Specifically, conventional methods such as pressure difference monitoring, visualization observation, optical scattering, and coagulation function testing mostly rely on indirect parameters, offline sampling, or complex external auxiliary conditions. Their detection processes have a certain degree of lag and are sensitive to blood flow velocity, blood color, air bubbles, tubing contamination, and the operating environment. Even when continuous monitoring is achieved in extracorporeal circulation or ECMO tubing, the detection results are still easily affected by individual differences and tubing disturbances.

[0003] Currently, most online thrombosis monitoring methods still rely on pressure changes, optical intensity, or single impedance parameters for analysis. Therefore, their accuracy and stability are limited by variations in these individual parameters. Achieving joint identification and accurate multi-state discrimination of early thrombosis formation, continuous growth, and blockage risk on a single detection platform remains a significant challenge. With the increasing demand for extracorporeal circulation technologies such as ECMO, artificial heart-lung machines, and blood purification, methods using single physical quantities or single detection parameters are no longer sufficient to meet the needs of applications requiring high stability, multi-state identification, and real-time early warning.

[0004] Fortunately, microwave technology has demonstrated unique advantages in recent years in characterizing the dielectric properties of liquids and in label-free detection of biological samples. By measuring parameters such as dielectric constant, loss characteristics, resonant frequency shift, and group delay of blood samples, it can provide a new, multi-dimensional, non-contact detection approach for complex blood systems. Unlike traditional single-parameter detection methods, multi-parameter microwave detection can acquire multi-dimensional information such as resonant frequency shift, amplitude, phase, and group delay without damaging the sealing of the tubing. Therefore, it has a simple structure, fast detection speed, and is also very suitable for clamp-on integration. For example, CN119679366B, an invention entitled "A Tissue Identification Method Based on Multi-Frequency Thermoacoustic Imaging and Joint Feature Extraction," discloses a method of radiating multi-frequency microwaves onto biological tissue, receiving thermoacoustic signals using an ultrasonic transducer, and then using the amplified signals for image reconstruction. The multi-frequency microwaves are pulsed microwaves of various frequencies. During imaging, different frequencies are switched sequentially to excite the tissue under test, resulting in multiple sets of thermoacoustic images under different frequency microwave excitations. These multiple sets of thermoacoustic images are then fused to obtain a multi-frequency thermoacoustic image. By selecting regions of interest from single-frequency images for feature extraction and using a differential algorithm to identify the distribution and content of different tissue components, this multi-frequency microwave thermoacoustic imaging system has been successfully implemented to identify and calculate the content of fat, muscle, and blood vessels. However, it is only suitable for static, one-time imaging. Continuous tracking in dynamic situations is difficult, impacting both measurement accuracy and real-time performance. Therefore, how to introduce a multi-frequency dielectric trajectory recognition mechanism into the same microwave detection platform and achieve collaborative extraction and intelligent analysis of multi-dimensional microwave response features has become an urgent problem to be solved in the fields of microwave biological detection and online thrombosis monitoring.

[0005] Therefore, there is an urgent need to propose an online thrombosis detection system and method based on multi-frequency microwave dielectric trajectory and intelligent discrimination to solve the above-mentioned technical problems. Summary of the Invention

[0006] To address the aforementioned problems, an online thrombosis detection system and method based on multi-frequency microwave dielectric trajectory and intelligent discrimination are provided. A brief overview of the invention is given below to provide a basic understanding of certain aspects of the invention. It should be understood that this overview is not an exhaustive summary of the invention. It is not intended to identify key or essential parts of the invention, nor is it intended to limit the scope of the invention.

[0007] The technical solution of the present invention: An online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination includes: The extracorporeal circulation tubing is used to continuously deliver blood, and the blood in the extracorporeal circulation tubing flows sequentially through the multi-frequency microwave detection unit and the multi-frequency microwave detection unit. The multi-frequency microwave transmitter and receiver module is used to transmit and receive microwave signals in multiple frequency bands; The multi-frequency microwave detection unit is connected to the multi-frequency microwave transmitting and receiving module. The multi-frequency microwave detection unit couples microwave signals to the blood in the extracorporeal circulation tubing to collect the thrombosis status signal of the blood. Differential reference cell: provides a benchmark reference. Intelligent discrimination module: Receives the output signal of the differential reference unit for discrimination.

[0008] Preferred configuration: The multi-frequency microwave transmitter and receiver module outputs 1-30GHz multi-frequency microwave excitation signals and receives and collects signals.

[0009] Preferably, a detection unit is formed by a multi-frequency microwave detection unit and a differential reference unit, and one or more detection units are arranged along the extracorporeal circulation pipeline.

[0010] Preferably, the multi-frequency microwave detection unit includes microwave resonators respectively installed on the upper and lower sides of the extracorporeal circulation tubing to apply microwave near-field coupling to the blood.

[0011] Preferred: Thrombus status is normal flow, early coagulation, or mural thrombus / blockage risk.

[0012] Preferably, the intelligent discrimination module receives the output signal of the differential reference unit and performs differential processing, feature extraction, feature fusion, and intelligent discrimination.

[0013] Preferred key feature parameters for feature extraction include: resonant frequency shift, transmission / reflection amplitude, phase delay, group delay, and dielectric spectrum.

[0014] Preferably, it also includes a result display and early warning module, which is connected to the intelligent discrimination module and displays the normal status, early warning, thrombus growth or high risk alarm accordingly.

[0015] A method for online thrombus detection based on multi-frequency microwave dielectric trajectory and intelligent discrimination, employing the aforementioned online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination, includes the following steps: S1. Multiple detection units are sequentially coupled and set on the outside of the pipeline. Blood samples are introduced into the pipeline to allow blood to flow continuously in each detection area and to achieve real-time monitoring. S2. The multi-band microwave excitation signal output by the multi-frequency microwave transmitting and receiving module is sent to the clamp-type multi-frequency microwave detection unit. Based on the microwave near-field coupling effect, the interaction between the microwave electromagnetic field and the flowing blood is established to create an overall microwave response model of the blood, and the response characteristics of the blood are regarded as an equivalent dielectric trajectory mapping operator. S3. Introduce at least two dielectric response recognition mechanisms corresponding to thrombus formation states in different detection areas. The multi-frequency microwave detection unit performs reflection and transmission acquisition, and the differential reference unit performs differential operations on multiple microwave signals through a reference compensation mechanism. S4, the signal processing and intelligent discrimination module extracts features from multiple sets of differential signals, obtains resonant frequency shift, transmission amplitude, reflection amplitude, phase delay, group delay and dielectric trajectory parameters, and obtains an equivalent feature matrix including multi-dimensional detection information, forming multiple controllable dielectric trajectory distribution regions in the feature space; S5. By adjusting the feature parameter weight ratio, compensation coefficient and analysis frequency band, the corresponding thrombosis state recognition mode is activated in different feature dimensions, and the signal fusion relationship of each detection unit is adjusted so that independent analysis or multi-dimensional collaborative discrimination can be achieved between multiple feature spaces. S6. When the detected feature parameters meet the preset blood thrombosis discrimination threshold conditions, multiple stable discrimination channels are formed in the feature space, and the recognition results remain stable even in the presence of blood flow fluctuations or system disturbances, thereby realizing multi-parameter joint online detection and hierarchical discrimination based on multi-frequency microwave dielectric trajectories.

[0016] The present invention has the following beneficial effects: 1. In this invention, because a multi-frequency dielectric trajectory recognition mechanism and a multi-parameter feature fusion method are introduced into the microwave detection device, the designed device can form independent dielectric trajectory distributions under different thrombosis states. Therefore, it greatly improves the discrimination accuracy compared with the single-parameter detection method, and naturally overcomes the various limitations encountered by traditional methods in terms of information dimension and stability. It also has excellent flexibility in configuring the detection strategy.

[0017] 2. In this invention, since the overall device adopts a modular design that integrates the clamp-type sensing structure and the differential reference unit, it can naturally and reasonably utilize the microwave dielectric response to achieve non-contact online detection. Therefore, it does not rely on complex reagents or large equipment, has a compact structure, and is conducive to clinical tubing integration. This also greatly improves its engineering adaptability in extracorporeal circulation, ECMO and blood management application scenarios.

[0018] 3. This invention breaks through the limitations of existing detection technologies in terms of adaptability, speed, accuracy, and information dimensions, and provides a new technical solution for building a highly sensitive, multi-parameter combined, and integrable real-time detection platform for extracorporeal circulation thrombosis. Attached Figure Description

[0019] Figure 1 A schematic diagram of an online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination; Figure 2A diagram of an online thrombus detection method based on multi-frequency microwave dielectric trajectory and intelligent discrimination; Figure 3 Schematic diagram of online monitoring and multi-parameter feature extraction of extracorporeal circulation tubing based on clamp-on microwave resonance detection; Figure 4 Spatial distribution diagram of multi-frequency microwave dielectric trajectory characteristics under different thrombotic states; Figure 5 Comparison of microwave transmission response under single-parameter detection and multi-parameter joint detection conditions; Figure 6 Distribution diagram of confidence characteristics of detection device A and detection device B based on multi-parameter dielectric trajectory; Figure 7 Statistical graph of robustness of online thrombosis detection device under random perturbation conditions. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention is described below with reference to specific embodiments shown in the accompanying drawings. However, it should be understood that these descriptions are merely exemplary and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0021] Specific implementation method one: Combining Figure 1-7 This embodiment describes an online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination, comprising: Extracorporeal circulation tubing: Microwave resonance detection module (clamp-type microwave resonance detection structure) and differential reference unit (differential reference structure) are sequentially arranged on the outside of the extracorporeal circulation blood tubing along the flow direction. The extracorporeal circulation blood tubing serves as a blood flow channel and is used to continuously transport blood. The blood in the extracorporeal circulation tubing flows sequentially through the multi-frequency microwave detection unit and the multi-frequency microwave detection unit. Multi-frequency microwave transmitting and receiving module: transmits and receives microwave signals of multiple frequency bands, providing signals for the multi-frequency microwave detection unit; Multi-frequency microwave detection unit: The multi-frequency microwave detection unit is connected to the multi-frequency microwave transmitting and receiving module. The multi-frequency microwave detection unit couples the microwave signal to the blood in the extracorporeal circulation tubing to achieve microwave field coupling and collects the feedback signal corresponding to the thrombosis status of the blood in the extracorporeal circulation tubing. Differential Reference Unit: Provides a reference item. The differential reference unit is connected to the multi-frequency microwave transmitter and receiver module. The reference reference item signal and the measured signal are differentially transported. The reference item of each detection unit is composed of a differential reference unit or an empty pipe reference area. The reference parameters are set according to the target detection frequency band to compensate for pipeline environment differences for stable detection results. Intelligent discrimination module: The intelligent discrimination module is signal-connected to the differential reference unit and receives the output signal of the differential reference unit for discrimination. In the microwave detection system, the present invention utilizes the resonant sensing structure and extracorporeal circulation pipeline to naturally and rationally construct an online detection platform suitable for flowing blood samples. More importantly, based on the difference in dielectric properties between normal flowing blood, early coagulated blood, and mural thrombi, a dielectric trajectory response with discriminative capability can be formed in the multi-frequency microwave range. This directly and effectively realizes the real-time discrimination of thrombus formation status, solving the problems of limited information dimension of existing single detection parameters, insufficient early discrimination accuracy, and difficulty in achieving non-contact real-time monitoring.

[0022] Specific Implementation Method Two: Combining Figure 1-7 This embodiment describes an online thrombosis detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination. The multi-frequency microwave transmitting and receiving module outputs a 1-30GHz multi-frequency microwave excitation signal and receives the acquisition parameter signal S.

[0023] Specific implementation method three: Combining Figure 1-7 This embodiment describes an online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination. A multi-frequency microwave detection unit and a differential reference unit form a detection unit, with one or more detection units arranged along the extracorporeal circulation tubing (extracorporeal blood tubing). The online thrombus detection device of this invention mainly consists of a clamp-type microwave resonant structure, extracorporeal blood tubing, and a differential reference unit, integrated on an external detection platform. It does not require damage to the original tubing structure or direct contact with blood, resulting in a compact and versatile structure suitable for extracorporeal circulation tubing and bedside monitoring scenarios. Each detection unit consists of a resonant structure and a corresponding thrombus-sensitive near-field coupling region forming a basic detection unit. The module; since different detection units are interconnected through extracorporeal circulation tubing and multi-channel signal acquisition structure, each unit can independently set the detection frequency band or discrimination parameters, thus naturally and reasonably forming a grouped multi-state detection structure. Each unit forms a detection array according to the design sequence, and different detection areas are used to achieve collaborative acquisition and feature reconstruction of thrombosis information. Since the device forms a stable dielectric trajectory discrimination mode in the multi-dimensional feature space under the detection conditions, the discrimination results have strong tolerance to individual blood differences, environmental disturbances and device parameter deviations. It can still maintain stable recognition performance during multi-state detection, thus significantly reducing the requirements for system consistency and experimental conditions.

[0024] Specific implementation method four: Combination Figure 1-7This embodiment describes an online thrombosis detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination. The multi-frequency microwave detection unit includes microwave resonators respectively arranged on the upper and lower sides of the extracorporeal circulation pipeline to apply microwave near-field coupling to the blood. Since a non-contact multi-frequency microwave detection method is used to achieve multi-parameter coupling analysis, online monitoring can be completed without offline sampling or complex biochemical labeling procedures. Therefore, it is naturally beneficial to shorten the detection time and improve the system response efficiency and integration. The device introduces at least two or more types of thrombus formation-related dielectric response features and adjusts the resonant frequency and near-field coupling parameters of each detection unit to achieve dielectric trajectory distribution in a multi-dimensional feature space. Under specific detection conditions, a robust discrimination channel can be formed in the corresponding feature region. When the detection is closed or the blood returns to the baseline state, the device exhibits no significant abnormal response. Each detection unit's reference term consists of a differential reference unit or an empty tube reference region. The reference parameters are set according to the target detection frequency band. The detection units are connected through extracorporeal blood tubing and a signal acquisition module. Different detection regions are equipped with different microwave resonant structures to detect the dielectric response of specific thrombus formation states.

[0025] Specific Implementation Method Five: Combining Figure 1-7 This embodiment describes an online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination, where the thrombus status is normal flow, early coagulation, or wall-attached thrombus / blockage risk. After multi-parameter feature extraction, the reflected and transmitted signals obtained during microwave detection are subjected to frequency domain and phase analysis. By constructing a multi-frequency microwave dielectric trajectory space, the mapping relationship between microwave response and thrombus formation state is realized, thereby forming a multi-parameter joint thrombus discrimination structure. By changing the weight ratio or compensation parameters between different detection features, a specific discrimination result distribution is formed in the detection device. When the resonant frequency shift feature dominates, the first thrombosis state distribution is output. When the phase or group delay feature dominates, another thrombosis state distribution is output. When multiple features work together, the device presents a multi-parameter joint discrimination state and has robustness.

[0026] Specific Implementation Method Six: Combination Figure 1-7 This embodiment describes an online thrombosis detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination. The intelligent discrimination module receives the output signal from the differential reference unit and sequentially performs differential processing comparison, feature extraction, feature fusion, and intelligent discrimination.

[0027] Specific implementation method seven: Combining Figure 1-7This embodiment describes an online thrombosis detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination. The key feature parameters for feature extraction include: extracted resonant frequency shift, transmission / reflection amplitude, phase delay, group delay, and dielectric spectrum.

[0028] Specific implementation method eight: Combination Figure 1-7 This embodiment describes an online thrombosis detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination. The system utilizes a signal processing and control module to change the detection frequency band, feature extraction method, and compensation parameter allocation, thereby controlling the opening and closing of different thrombosis discrimination channels. The discrimination channels are closed when the detection signal is removed or the parameters are restored to their initial state. When the multi-parameter detection meets the set conditions, a thrombosis discrimination result with stable recognition capability is generated. It also includes a result display and early warning module, which is connected to the output of the signal processing and intelligent discrimination module to display normal status, early warning, thrombus growth or high risk alarm.

[0029] Specific Implementation Method Nine: Combining Figure 1-7 This embodiment describes an online thrombus detection method based on multi-frequency microwave dielectric trajectory and intelligent discrimination. The method is characterized by employing the aforementioned online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination, comprising the following steps: S1. Multiple detection units are sequentially coupled and set on the outside of the pipeline. Each detection unit corresponds to a detection area. Blood samples are introduced into the pipeline to allow blood to flow continuously in each detection area and to achieve real-time monitoring. S2. The multi-band microwave excitation signal output by the multi-frequency microwave transmitting and receiving module is sent to the clamp-type multi-frequency microwave detection unit. Based on the microwave near-field coupling effect, the interaction between the microwave electromagnetic field and the flowing blood is established to create an overall microwave response model of the blood, and the response characteristics of the blood are regarded as an equivalent dielectric trajectory mapping operator. S3. Introduce at least two dielectric response recognition mechanisms corresponding to thrombus formation states in different detection areas. The multi-frequency microwave detection unit performs reflection and transmission acquisition, and the differential reference unit performs differential operations on multiple microwave signals through a reference compensation mechanism. S4, the signal processing and intelligent discrimination module extracts multi-parameter features such as frequency domain, phase, and group delay from multiple sets of differential signals, and obtains resonant frequency shift, transmission amplitude, reflection amplitude, phase delay, group delay and dielectric trajectory parameters to obtain an equivalent feature matrix including multi-dimensional detection information, thereby forming multiple controllable dielectric trajectory distribution regions in the feature space. S5. By adjusting the feature parameter weight ratio, signal compensation coefficient and microwave analysis frequency band, the corresponding thrombosis status recognition mode can be activated in different feature dimensions, and the signal fusion relationship of each detection unit can be further adjusted to enable independent analysis or multi-dimensional collaborative discrimination between multiple feature spaces. S6. When the detected feature parameters meet the preset blood thrombosis discrimination threshold conditions, multiple stable discrimination channels are formed in the feature space, and the recognition results remain stable even in the presence of blood flow fluctuations or system disturbances, thereby realizing the non-invasive extracorporeal circulation thrombosis multi-parameter joint online detection and grading discrimination function based on multi-frequency microwave dielectric trajectory.

[0030] Specific Implementation Method Ten: Combining Figure 1-7 This embodiment describes an online thrombus detection method based on multi-frequency microwave dielectric trajectory and intelligent discrimination. First, a non-contact thrombus sensing unit is constructed using a clamp-on microwave resonant sensing structure, the blood flow channel within the extracorporeal circulation tubing, and a local thrombus-sensitive detection area. This naturally forms a multi-frequency, multi-parameter coupled microwave dielectric response network. The reflection, transmission, phase, and group delay characteristics of this network under conditions of no thrombus, early thrombus formation, thrombus growth, and thrombus maturation are then systematically analyzed. More importantly, each detection unit in the system consists of a multi-frequency microwave resonant structure, a differential reference structure, and a thrombus-sensitive near-field coupling region. By utilizing different frequency bands, different resonant modes, and different detection positions, at least two or more types of thrombus formation-related dielectric characteristics are introduced, thus enabling a direct and efficient multi-parameter joint detection mechanism for thrombus status. Adjacent detection units can be arranged sequentially along the extracorporeal circulation tubing, allowing blood samples to flow continuously between different detection areas and different frequency band response channels and be monitored in real time. The detection process is reliably provided by a multi-frequency microwave transmitter and receiver module, combined with differential signal processing and intelligent discrimination modules to select different feature combinations and discrimination strategies, thereby constructing a scalable thrombus dielectric trajectory space. By reasonably adjusting the detection frequency band, resonant frequency shift, quality factor, phase delay, group delay, and feature weight parameters, distinguishable and independently analyzable thrombus identification patterns are formed in multiple feature dimensions. Therefore, when the blood in the pipeline is in different states such as normal flow, early coagulation, mural thrombus formation, or increased risk of blockage, the system can naturally realize functions such as online monitoring, early warning, multi-state grading, and risk discrimination. This invention breaks through the fundamental limitations of traditional pressure, optical, ultrasonic, or single impedance detection, which have limited information dimensions, are easily affected by the pipeline environment, and are difficult to make early judgments. Therefore, it provides a new engineering-feasible solution for constructing a highly sensitive, non-contact, multi-parameter combined, scalable, highly integrated, and robust online thrombus detection system, and opens up new technical paths for the application of microwave dielectric detection technology in extracorporeal circulation, ECMO, artificial heart and lungs, and blood management.

[0031] The present invention proposes an online thrombosis detection device and method based on multi-frequency microwave dielectric trajectory and intelligent discrimination mechanism, which is mainly used for extracorporeal circulation blood management, ECMO tubing thrombosis early warning, and microwave intelligent biosensing and real-time discrimination platform.

[0032] Example 1: The device is integrated into a clamp-on microwave resonant sensing and extracorporeal circulation tubing integrated detection platform. The microwave sensing substrate is made of Rogers RO4350B dielectric substrate with a dielectric constant εr≈3.48 and a substrate thickness of approximately 0.508 mm. The microwave resonant structure adopts a combination of complementary open-loop resonator (CSRR) and microstrip coupling line, with a copper layer thickness of approximately 35 μm, a unit length of approximately 18 mm, and a coupling gap of approximately 0.30 mm. The corresponding operating frequency band is 1–30 GHz, with a preferred analysis frequency band of 4–12 GHz. The extracorporeal circulation tubing uses medical-grade silicone tubing with an inner diameter of approximately 6.0 mm and an outer diameter of approximately 8.0 mm. Non-contact coupling between the microwave near field and the blood inside the tubing is achieved through a clamp-on upper and lower probe structure. The detection device also includes a differential reference unit. The reference area uses an empty tubing section of the same material or a stable anticoagulated blood section to subtract the effects of temperature, tubing wall thickness, and system drift. The basic parameters of the detection system are as follows: input microwave power approximately 0 dBm, sweep frequency step approximately 10 MHz, detection distance approximately 0.5 mm, blood flow rate approximately 80 mL / min, and detection temperature controlled at 37.0 ± 0.5 ℃. The blood samples are categorized into three states: normal flow state, early coagulation state, and mural thrombus state, with typical fibrin / aggregate volume fractions of approximately 0%, 5%, and 15%, respectively. Multi-state detection is achieved through different degrees of coagulation and thrombus load.

[0033] Depend on Figure 1 The overall structure of the microwave thrombosis online detection device and the flow of blood samples in the extracorporeal circulation tubing can be clearly and naturally observed. The upper part shows the multi-frequency microwave excitation and signal acquisition path, while the lower part shows the clamp-type resonant detection unit and the thrombus-sensitive near-field coupling region. From this, it can be clearly concluded that different thrombus formation states exhibit different electromagnetic response characteristics within the microwave frequency band, and form distinguishable dielectric trajectory characteristics within the target frequency band. More importantly, when the degree of blood coagulation and the mural thrombus load reach a certain level, the resonant frequency shifts significantly. For example, the resonant frequency under normal flow conditions is approximately 6.42 GHz, while in the early coagulation state and the mural thrombus state, it shifts to approximately 6.31 GHz and 6.08 GHz, respectively. Simultaneously, the corresponding transmission amplitude changes are approximately 0.9 dB and 2.6 dB, thus yielding frequency domain response characteristics with good separation.

[0034] The microwave detection device is rationally mapped onto a multi-parameter feature space, and then the reflection spectrum, transmission spectrum, phase, and group delay response under different thrombotic states are constructed, see [reference needed]. Figure 3 The figure clearly shows significant differences between the different states across multiple parameter dimensions. For example, at 6.20 GHz, the phases are approximately... 38° 56° and The group delays for the 83° values ​​are approximately 0.31 ns, 0.54 ns, and 0.88 ns, respectively. More importantly, the separability between different states can be significantly improved by introducing differential reference and multi-parameter joint analysis methods, thus clearly demonstrating its online detection characteristics based on multi-frequency microwave dielectric trajectories.

[0035] from Figure 4 The distribution and classification results of various thrombus states in the feature space can be observed very naturally and clearly, with the horizontal axis representing the resonant frequency shift feature and the vertical axis representing the group delay or phase feature. This makes it easy to identify the cluster distribution of different states in the feature space. Normal blood samples cluster in the low-frequency shift, low-group-delay region, while mural thrombus samples are distributed in the high-frequency shift, high-group-delay region. Early agglutination samples provide a good contrast, with their distribution falling between the two. More importantly, by reasonably adjusting the detection frequency band and feature weights, the classification boundary can be further optimized and the discrimination accuracy improved. Therefore, a multi-state online thrombus detection method based on multi-frequency microwave dielectric trajectories can be naturally constructed. Thus, the conclusion of this embodiment is very clear: the designed microwave online thrombus detection device achieves stable extraction of multi-parameter dielectric trajectory features and multi-state discrimination in a non-contact clamping platform, and exhibits good robustness to blood flow fluctuations. Example 2: In the online thrombosis detection device and method based on a multi-frequency microwave dielectric trajectory detection architecture, two detection schemes, namely detection device A and detection device B, are set up, and their microwave response characteristics and thrombosis discrimination capabilities are compared and analyzed. The device is integrated into a clamp-type microwave resonant sensor and extracorporeal circulation tubing integrated detection platform. The dielectric constant of the microwave sensor substrate is εr≈3.48, the substrate thickness is approximately 0.508 mm, the length of the microwave resonant unit is approximately 18 mm, the coupling gap is approximately 0.30 mm, the operating frequency band is 1–30 GHz, the preferred analysis frequency band is 4–12 GHz, the inner diameter of the extracorporeal circulation tubing is approximately 6.0 mm, the outer diameter is approximately 8.0 mm, and the clamping positioning error is controlled within ±5%. Figure 5 As shown, under the same microwave detection structure and blood sample conditions, detection device A and detection device B can reconstruct the dielectric trajectory distribution simply by changing the thrombus identification strategy and feature extraction method. Figure 5The solid line represents the transmission response distribution under single-parameter detection conditions, while the dashed line represents the response result after multi-parameter joint detection. It can be observed that the original feature distribution range is significantly expanded under multi-parameter detection, local response differences are enhanced, and new effective separation regions are formed in the feature space, demonstrating the programmable feature control capability of the device under different detection configurations.

[0036] This invention rationally maps microwave detection results to a finite sample feature space model. First, it extracts resonant frequency shift, phase, and group delay features under multi-state blood sample conditions. Then, it uses a classification confidence index to naturally and clearly identify the distribution of each sample. See details... Figure 6 Specifically, Figure 6 The left side shows the feature distribution results under the conditions of detection device A, and the right side shows the results of detection device B. The intensity of the color reflects the discrimination confidence of the sample in the corresponding category. It is clearly visible from the figure that high-density clustering exists near certain feature regions, namely the studied normal flow state, early coagulation state, and mural thrombus state. More importantly, compared with detection device A, the multi-parameter detection device B can form clearer classification boundaries in the feature space. Device A has an overall recognition accuracy of approximately 76.0% and an average classification confidence of approximately 0.79 for the three types of samples, while device B improves the overall recognition accuracy to 93.0% and the average classification confidence to 0.91. Therefore, reconstructible discriminative control of thrombus state can be achieved by adjusting the detection parameters and feature combinations.

[0037] This invention naturally and reasonably introduces random disturbances (range ±10%) of blood flow velocity fluctuations and detection noise into existing systems to simulate actual blood variations and system instability factors. A rigorous statistical analysis of the detection stability is then performed. (See [link to relevant documentation]). Figure 7 The horizontal axis represents the overall perturbation ratio, and the vertical axis represents the average classification confidence and overall recognition accuracy. The analysis results are very clear: under a certain range of perturbation conditions, the recognition performance of the multi-parameter detection device remains basically stable. When the perturbation ratio increases from 0% to 10%, the average classification confidence of detection device A decreases from 0.79 to 0.62, and the overall recognition accuracy decreases from 76.0% to 65.0%; while the average classification confidence of detection device B only decreases from 0.91 to 0.84, and the overall recognition accuracy decreases from 93.0% to 87.0%, indicating that the multi-parameter joint detection structure has a stronger tolerance to external perturbations. Therefore, it is quite natural to conclude that the proposed microwave thrombosis online detection device can reliably maintain dielectric trajectory discrimination characteristics even under the presence of blood flow fluctuations and noise perturbations, meaning that its structure has excellent robustness in practical extracorporeal circulation thrombosis monitoring applications.

[0038] It should be noted that in the above embodiments, as long as the technical solutions are not contradictory, they can be permuted and combined. Those skilled in the art can exhaust all possibilities based on the mathematical knowledge of permutation and combination. Therefore, the present invention will not describe the technical solutions after permutation and combination one by one, but it should be understood that the technical solutions after permutation and combination have been disclosed by the present invention.

[0039] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A thrombus online detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination, characterized in that: include: Extracorporeal circulation tubing: used for continuous blood delivery, with the blood in the extracorporeal circulation tubing flowing sequentially through the multi-frequency microwave detection unit; Multi-frequency microwave transmitter and receiver module: transmits and receives microwave signals; Multi-frequency microwave detection unit: The multi-frequency microwave detection unit is connected to the multi-frequency microwave transmitting and receiving module. The multi-frequency microwave detection unit couples microwave signals to the blood in the extracorporeal circulation tubing to collect the thrombosis status signal of the blood. Differential reference cell: provides a benchmark reference. Intelligent discrimination module: Receives the output signal of the differential reference unit for discrimination.

2. The online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination according to claim 1, characterized in that: The multi-frequency microwave transmitter and receiver module outputs 1-30GHz multi-frequency microwave excitation signals and receives and collects signals.

3. The online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination according to claim 2, characterized in that: A multi-frequency microwave detection unit and a differential reference unit form a detection unit, and one or more detection units are arranged along the extracorporeal circulation pipeline.

4. The online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination according to claim 3, characterized in that: The multi-frequency microwave detection unit includes microwave resonators installed on the upper and lower sides of the extracorporeal circulation tubing to apply microwave near-field coupling to the blood.

5. The online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination according to claim 4, characterized in that: Thrombosis status includes normal flow, early coagulation, or mural thrombus / blockage risk.

6. The online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination according to claim 5, characterized in that: The intelligent discrimination module receives the output signal from the differential reference unit and performs differential processing, feature extraction, feature fusion, and intelligent discrimination.

7. The online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination according to claim 6, characterized in that: Key feature parameters for feature extraction include: resonant frequency shift, transmission / reflection amplitude, phase delay, group delay, and dielectric spectrum.

8. The online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination according to claim 7, characterized in that: It also includes a result display and early warning module, which is connected to the intelligent discrimination module and displays the normal status, early warning, thrombus growth or high risk alarm.

9. A method for online thrombus detection based on multi-frequency microwave dielectric trajectory and intelligent discrimination, characterized in that: The online thrombus detection system based on multi-frequency microwave dielectric trajectory and intelligent discrimination as described in any one of claims 1-8 includes the following steps: S1. Multiple detection units are sequentially coupled and set on the outside of the pipeline. Blood samples are introduced into the pipeline to allow blood to flow continuously in each detection area and to achieve real-time monitoring. S2. The multi-band microwave excitation signal output by the multi-frequency microwave transmitting and receiving module is sent to the clamp-type multi-frequency microwave detection unit. Based on the microwave near-field coupling effect, the interaction between the microwave electromagnetic field and the flowing blood is established to create an overall microwave response model of the blood, and the response characteristics of the blood are regarded as an equivalent dielectric trajectory mapping operator. S3. Introduce at least two dielectric response recognition mechanisms corresponding to thrombus formation states in different detection areas. The multi-frequency microwave detection unit performs reflection and transmission acquisition, and the differential reference unit performs differential operations on multiple microwave signals through a reference compensation mechanism. S4, the signal processing and intelligent discrimination module extracts features from multiple sets of differential signals, obtains resonant frequency shift, transmission amplitude, reflection amplitude, phase delay, group delay and dielectric trajectory parameters, and obtains an equivalent feature matrix including multi-dimensional detection information, forming multiple controllable dielectric trajectory distribution regions in the feature space; S5. By adjusting the feature parameter weight ratio, compensation coefficient and analysis frequency band, the corresponding thrombosis state recognition mode is activated in different feature dimensions, and the signal fusion relationship of each detection unit is adjusted so that independent analysis or multi-dimensional collaborative discrimination can be achieved between multiple feature spaces. S6. When the detected feature parameters meet the preset blood thrombosis discrimination threshold conditions, multiple stable discrimination channels are formed in the feature space, and the recognition results remain stable even in the presence of blood flow fluctuations or system disturbances, thereby realizing multi-parameter joint online detection and hierarchical discrimination based on multi-frequency microwave dielectric trajectories.

Images