Eddy current type vein pot blood coagulation early warning method and system

Abstract

The invention discloses a vortex type vein pot blood coagulation early warning method and system, and belongs to the technical field of data acquisition and identification, a macro lens with a near-infrared LED light source and a CMOS camera are adopted for data acquisition and identification, data processing of a continuous dynamic blood flow environment is effectively achieved, and the accuracy of blood coagulation early warning is improved. And the collected data is analyzed and processed to generate an event stream, so that the interference of dynamically flowing blood on thrombus size and position detection can be eliminated, and the thrombus generation process can be tracked in real time, so that medical personnel can find thrombus formation in time, and the thrombus detection efficiency is improved. The device can predict the clogging condition of the vein pot according to the size and position of thrombus and prompt medical staff to replace the vein pot, so that the patrol pressure of the medical staff is reduced, the extra cost caused by frequent replacement of the vein pot is reduced, and meanwhile, the probability that the vein pot is clogged by the thrombus is greatly reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of data acquisition and recognition, and particularly relates to a vortex venous chamber blood coagulation warning method and system. Background Art

[0002] In hemodialysis treatment, the venous chamber, as a key component of extracorporeal circulation, plays an important role in collecting and filtering the blood for reinfusion. However, due to factors such as blood-air contact, hemodynamic changes, and the action of anticoagulant factors, thrombus is extremely likely to form in the venous chamber, which will not only cause the dialysis treatment to be unable to proceed normally, but also the line blockage caused by the thrombus will lead to blood loss, causing great harm to the patient's physiology and psychology. Therefore, it is of great significance to timely detect the coagulation phenomenon occurring in the delivery line and early warn the medical staff to take measures.

[0003] For a long time, the monitoring of venous chamber blood coagulation has relied on medical staff to identify thrombus through naked-eye observation. The naked-eye observation by medical staff is affected by factors such as light and experience, making it difficult to accurately judge the early signs of thrombus formation and prone to missed diagnosis or misdiagnosis. Moreover, thrombus formation is a dynamic process, and it is difficult to achieve 24-hour uninterrupted monitoring by naked-eye observation, unable to timely detect the trend of thrombus formation and delaying the best intervention time. In addition, regularly replacing the venous chamber not only increases the workload of medical staff, but also causes waste of medical resources and increases the economic burden on patients. For this reason, currently, the dialysis machine will use an internal pressure sensor to continuously monitor the air pressure change in the pipeline in real time. When coagulation occurs in the pipeline, it will cause an increase in internal pressure and trigger an alarm. However, the venous chamber needs to be connected to the external atmospheric environment and cannot generate air pressure change characteristics. Therefore, there is an urgent need for a real-time, non-contact thrombus recognition and warning system suitable for vortex venous chambers. Summary of the Invention

[0004] The present invention provides a vortex venous chamber blood coagulation warning method and system to solve the problem that the prior art cannot realize data processing in a continuous dynamic blood flow environment.

[0005] On the one hand, the present invention provides a vortex venous chamber blood coagulation warning system, including: A venous chamber fixing device for being installed on the vortex venous chamber to realize fixation to the dialysis device; A macro lens with a near-infrared LED light source for taking close-up photos and providing a near-infrared LED light source, and the intensity and irradiation angle of the near-infrared LED light source are adjustable; A data line for data transmission; A near-infrared CMOS camera for collecting image information of the vortex venous chamber through the macro lens and transmitting the image information to the data processing module through the data line for vortex venous chamber blood coagulation warning; A bracket for mounting a macro lens and a near-infrared CMOS camera, so that the data acquisition directions of the macro lens and the near-infrared CMOS camera face the eddy current venous ampulla directly, ensuring that the eddy current venous ampulla is within the effective acquisition area; A fixture for mounting the bracket on the venous ampulla fixing device to enable the near-infrared CMOS camera to stably acquire image information.

[0006] Further, the wavelength of the near-infrared LED light source on the macro lens is 850 nm.

[0007] Further, the focal length of the macro lens is 2.7 mm, and the viewing angle is 120 degrees.

[0008] Further, the narrow band of the near-infrared CMOS camera is 850 nm, the sampling speed is 30 frames per second, and the output resolution is 1920×1080.

[0009] Further, the data processing module is set as a Raspberry Pi, and the data processing module is connected to the near-infrared CMOS camera using a USB communication protocol, and the data processing module communicates with the host computer through VNC to utilize the data processing ability of the host computer for coagulation warning of the eddy current venous ampulla.

[0010] On the other hand, the present invention provides a method for coagulation warning of an eddy current venous ampulla, including: Turn on the near-infrared LED light source on the macro lens to irradiate the eddy current venous ampulla, and continuously capture images of the eddy current venous ampulla through the near-infrared CMOS camera to obtain image information; Process the image information of two adjacent frames of the eddy current venous ampulla through an event recognition algorithm to obtain first image event information, and combine multiple pieces of first image event information with timestamps to form a first event stream; Based on the image information, use an image segmentation algorithm to identify and extract the eddy current venous ampulla area, and construct a second event stream corresponding to the eddy current venous ampulla area according to the eddy current venous ampulla area and the first event stream; Classify the second event stream corresponding to the extracted eddy current venous ampulla area through an eddy current venous ampulla thrombus classification model to determine the data classification result, and based on the data classification result and a preset warning condition, perform coagulation warning of the eddy current venous ampulla.

[0011] Further, processing the image information of two adjacent frames of the eddy current venous ampulla through an event recognition algorithm to obtain first image event information includes: The image information of two adjacent frames of the vortex venous chamber is processed using a differential algorithm or an optical flow analysis method to obtain a differential event of the difference between the image information of two adjacent frames of the vortex venous chamber, and the first image event information is obtained.

[0012] Further, based on the image information, an image segmentation algorithm is used to identify and extract the vortex venous chamber region, including: The image information is segmented using an image segmentation algorithm based on a convolutional neural network to obtain the vortex venous chamber region corresponding to the image information.

[0013] Further, according to the vortex venous chamber region and the first event stream, a second event stream corresponding to the vortex venous chamber region is constructed, including: The vortex venous chamber region is multiplied by the first image event information in its corresponding first event stream to obtain the second image event information; The second image event information is used to construct a second event stream corresponding to the vortex venous chamber region.

[0014] Further, the second event stream corresponding to the extracted vortex venous chamber region is classified by a vortex venous chamber thrombus classification model to determine the data classification result, and based on the data classification result and a preset warning condition, a vortex venous chamber coagulation warning is performed, including: Each second image event information in the second event stream corresponding to the extracted vortex venous chamber region is classified using a vortex venous chamber thrombus classification model based on a convolutional neural network to determine the data classification result corresponding to each second image event information; When the number of data classification results that meet the preset warning condition exceeds the recognition warning threshold, a warning signal is output, thereby performing a vortex venous chamber coagulation warning.

[0015] A vortex venous chamber coagulation warning method and system provided by the present invention effectively realizes data processing in a continuous dynamic blood flow environment by using a macro lens with a near-infrared LED light source and a CMOS camera for data collection and recognition, and analyzes and processes the collected data to generate an event stream, which can eliminate the interference of dynamically flowing blood on the detection of thrombus size and position, and real-time track the thrombus formation process, enabling medical staff to timely detect the formation of thrombus. The device can predict the blockage condition of the venous chamber according to the size and position of the thrombus, prompt medical staff to replace the venous chamber, reduce the inspection pressure of medical staff, reduce the additional cost caused by frequent replacement of the venous chamber, and at the same time greatly reduce the occurrence probability of thrombus blocking the venous chamber. Description of the Drawings

[0016] The accompanying drawings here are incorporated into the specification and form a part of this specification, showing embodiments consistent with the present invention, and are used together with the specification to explain the principles of the present invention.

[0017] Figure 1 It is a schematic structural diagram of a vortex venous pot coagulation warning system provided for an embodiment of the present invention.

[0018] Figure 2 It is a flowchart of a vortex venous pot coagulation warning method provided for an embodiment of the present invention.

[0019] Wherein, 1 - venous pot fixing device, 2 - macro lens, 3 - data line, 4 - near-infrared CMOS camera, 5 - bracket, 6 - fixture, 7 - vortex venous pot.

[0020] Through the above accompanying drawings, specific embodiments of the present invention have been shown, and there will be more detailed descriptions hereinafter. These drawings and written descriptions are not intended to limit the scope of the inventive concept in any way, but to illustrate the concept of the present invention to those skilled in the art by reference to specific embodiments. Detailed Embodiments

[0021] Here, the exemplary embodiments will be described in detail, and the examples are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0022] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] In the existing venous pot thrombosis monitoring technologies, the main means include artificial visual observation method, pressure / flow sensor monitoring, ultrasonic examination for auxiliary judgment, etc. However, these methods have many drawbacks: The artificial visual observation method depends on the experience of medical staff, has strong subjectivity, and cannot achieve continuous monitoring. The professional levels and clinical experiences of different medical staff directly affect the judgment accuracy. For example, newly recruited nurses may have insufficient ability to identify early fibrin deposition in the venous pot, while senior nurses may miss detections due to visual fatigue. In addition, the manual inspection is usually carried out once every one to two hours, and during this period, the patient may be at risk of having the venous pot blocked due to rapid blood coagulation.

[0024] Existing dialysis equipment usually uses a venous pressure sensor for early warning of blood clot blockage in the venous chamber. There is a significant lag in the monitoring of the pressure sensor, and it only triggers a pressure alarm after the blood clot blocks the venous chamber. In addition, the false alarm rate of the pressure sensor is high, and factors such as patient body position changes and blood transfusion tube bending may all cause false alarms, affecting the normal work of medical staff.

[0025] Although ultrasonic examination technology can detect blood flow velocity, it requires external equipment and is costly, making it difficult to integrate into a real-time monitoring system. Currently, there is no application of ultrasonic technology in the field of blood clot monitoring in the venous chamber, and there is still much room for improvement in this technology.

[0026] As Figure 1 shown, an embodiment of the present invention provides an eddy current venous chamber blood clot early warning system, including: A venous chamber fixing device 1, which is used to be installed on the eddy current venous chamber 7 to achieve fixation to the dialysis equipment; A macro lens 2 with a near-infrared LED light source, which is used for close-range shooting and providing a near-infrared LED light source, and the intensity and irradiation angle of the near-infrared LED light source can be adjusted to avoid the appearance of light source reflections on the surface of the venous chamber affecting image acquisition; A data line 3, which is used for data transmission; the data line 3 can be set as a USB data line, which connects the near-infrared CMOS camera 4 to the data processing module, issues an image acquisition instruction to the near-infrared CMOS camera 4, and transmits the obtained data to process the data, obtain event stream information, and then evaluate the blood clot formation situation and send a blood clot early warning through a wireless communication protocol; A near-infrared CMOS camera 4, which is used to collect image information of the eddy current venous chamber 7 through the macro lens 2, and transmits the image information to the data processing module through the data line 3 for eddy current venous chamber blood clot early warning; A bracket 5, which is used to install the macro lens 2 and the near-infrared CMOS camera 4, so that the data collection directions of the macro lens 2 and the near-infrared CMOS camera 4 are directly facing the eddy current venous chamber 7, and there is a guide groove at the camera fixing place, and the height of the camera fixing position can be adjusted to ensure that the eddy current venous chamber 7 is within the effective collection area; A clamp 6, which is used to install the bracket 5 on the venous chamber fixing device 1 to enable the near-infrared CMOS camera 4 to stably collect image information.

[0027] Optionally, the clamp 6 has a U-shaped structure, can straddle the fixing rod of the venous chamber fixing device, and can be fixed under the bracket with screws, playing a role in installing the bracket on the venous chamber fixing device.

[0028] Blood is injected from the side of the eddy current venous pot 7, forms an eddy current in the eddy current venous pot 7, and then flows out from the lower outlet of the eddy current venous pot 7.

[0029] By adopting the above technical solution, image data inside the venous pot can be obtained by using the reflection of blood on near-infrared light. Analyzing and processing the collected data to generate an event stream can eliminate the interference of dynamically flowing blood on the detection of thrombus size and location, and can track the thrombus formation process in real time, enabling medical staff to detect the formation of thrombus in a timely manner. This device can predict the blockage condition of the venous pot according to the size and location of the thrombus, and prompt medical staff to replace the venous pot, reducing the inspection pressure on medical staff, reducing the additional costs caused by frequent replacement of the venous pot, and at the same time greatly reducing the occurrence probability of thrombus blocking the venous pot.

[0030] In an embodiment of the present invention, the wavelength of the near-infrared LED light source on the macro lens 2 is 850 nm.

[0031] In an embodiment of the present invention, the focal length of the macro lens 2 is 2.7 mm, and the viewing angle is 120 degrees.

[0032] In an embodiment of the present invention, the narrow band of the near-infrared CMOS camera 4 is 850 nm, the sampling speed is 30 frames per second, and the output resolution is 1920×1080.

[0033] In an embodiment of the present invention, the data processing module is set as a Raspberry Pi, and the data processing module is connected to the near-infrared CMOS camera 4 using a USB communication protocol, and the data processing module communicates with the upper computer through VNC to utilize the data processing ability of the upper computer for early warning of coagulation in the eddy current venous pot.

[0034] The data acquisition module in the present invention utilizes the transmission phenomenon of blood on near-infrared light. The main components of blood are hemoglobin and water, and the absorption of near-infrared light in the 700-900 nm band by hemoglobin and water is the weakest compared to other bands. Therefore, this band is used as the wavelength of the near-infrared light source in the data acquisition module. Further, analyzing the energy-saving lamps used in the ward, it is found that there is no light in the 850 nm band in the energy-saving lamps. Therefore, the near-infrared light source in the data acquisition module is preferably 850 nm, which greatly reduces the interference of the environment on the operation of the device.

[0035] Generally speaking, under different blood states, the absorption of light by the blood in the vortex venous chamber is also different. Therefore, an event stream can be generated based on the near-infrared light image information to analyze the blood state. This solution uses a near-infrared light source for illumination and generates an image by near-infrared light reflection to clearly present the internal state of the vortex venous chamber. The absorption of near-infrared light by blood is very low, while thrombus cannot transmit near-infrared light. Therefore, the internal state of the vortex venous chamber can be obtained more clearly.

[0036] In the present invention, the shape of the fixing device 1 of the vortex venous chamber matches the shape of the vortex venous chamber and can be sleeved outside the vortex venous chamber. The overall structure of the housing does not affect the direct observation of the vortex venous chamber from the outside.

[0037] As Figure 2 shown, the present invention provides a method for warning of blood clotting in a vortex venous chamber, including: S1. Turn on the near-infrared LED light source on the macro lens 2 to irradiate the vortex venous chamber 7, and continuously capture images of the vortex venous chamber 7 through the near-infrared CMOS camera 4 to obtain image information; S2. Process the image information of two adjacent frames of the vortex venous chamber 7 through an event recognition algorithm to obtain first image event information, and combine multiple pieces of first image event information with timestamps to form a first event stream; S3. Based on the image information, use an image segmentation algorithm to identify and extract the vortex venous chamber area, and construct a second event stream corresponding to the vortex venous chamber area according to the vortex venous chamber area and the first event stream; S4. Classify the second event stream corresponding to the extracted vortex venous chamber area through a vortex venous chamber thrombus classification model to determine the data classification result, and perform a warning of blood clotting in the vortex venous chamber based on the data classification result and preset warning conditions.

[0038] In the embodiment of the present invention, processing the image information of two adjacent frames of the vortex venous chamber 7 through an event recognition algorithm to obtain first image event information includes: Using a difference algorithm or an optical flow analysis method to process the image information of two adjacent frames of the vortex venous chamber 7, obtaining a differential event between the image information of two adjacent frames of the vortex venous chamber 7, and obtaining first image event information.

[0039] The event information △A1 is obtained by processing two frames of images through a differential algorithm, and multiple image event information is combined with timestamps to form an event stream. First, the adjacent two frames of images are grayscaled. The pixel matrices of the adjacent two frames of images after grayscaling are A1 and A2 respectively. The frame difference method is used to perform differential on the corresponding pixels at the same positions of A1 and A2 and take the absolute value to obtain the differential event ΔA1, and the timestamp is attached to the event in combination with the current time t1 to obtain the event (ΔA1, t1). Multiple events are encoded to form the event stream M(ΔA, t).

[0040] Optionally, the optical flow analysis method is used to analyze the pixel matrices A1 and A2 of the adjacent two frames of images to obtain the differential event ΔA1.

[0041] Optionally, the time-frequency method is used to analyze the pixel matrices A1 and A2 of the adjacent two frames of images to obtain the differential event ΔA1.

[0042] Among them, the differential event ΔA1 is a 1980×1080 event matrix, and the elements in the matrix represent the absolute value of the brightness difference between the two frames of grayscale images. By setting the brightness change threshold to identify the changed part of the swirling venous lacuna image, the following formula is used for processing:

[0043] In the embodiments of the present invention, based on the image information, the swirling venous lacuna region is identified and extracted by using an image segmentation algorithm, including: The image information is segmented by an image segmentation algorithm based on a convolutional neural network to obtain the swirling venous lacuna region corresponding to the image information.

[0044] Optionally, the specific principle of segmenting the image by using an image segmentation algorithm based on a convolutional neural network is as follows: First, the swirling venous lacuna region in the dataset is labeled by the labelme software. The labeled data is trained through a neural network, and the image data obtained by the near-infrared camera is input into the trained model to identify the swirling venous lacuna region and output the segmented image result.

[0045] Optionally, the real-time instance segmentation model YOLACT is used for image segmentation, and the principle is as follows: The vortex venous sinus region in the dataset is labeled using the labelme software. The labeled data is trained through a neural network to obtain a trained model. The YOLACT model learns multiple prototype masks (P) for each image. After inputting the image matrix A1 collected by a CMOS near-infrared camera, the mask confidence is predicted through the mask coefficient prediction branch, and the overall predicted output set A (class, x, y, w, h, C) is obtained, where class is the class label, including the vortex venous sinus region and other regions, x, y, w, h are the calibration box regression values, where x and y are the endpoint coordinates of the box, w and h are the width and height of the box, and C is a 1× k matrix. The image segmentation result of the vortex venous sinus is calculated through the following formula:

[0046] where the formula for the sigmoid activation function is:

[0047] The binary cross-entropy loss function is adopted as the loss function of the mask :

[0048] where is the binary label matrix (pixels in the venous sinus region are 1, and pixels in non-venous sinus regions are 0), is the predicted value, that is, the predicted result output by the model.

[0049] Subsequently, the event matrix ΔA1 of the differential of the vortex venous sinus image is processed to obtain the event E1 of the venous sinus region:

[0050] In the embodiment of the present invention, according to the vortex venous sinus region and the first event stream, a second event stream corresponding to the vortex venous sinus region is constructed, including: Multiplying the vortex venous sinus region by the first image event information in its corresponding first event stream to obtain second image event information; Using the second image event information to construct a second event stream corresponding to the vortex venous sinus region.

[0051] Optionally, a convolutional neural network-based vortex venous pot thrombus classification model is used to classify the input event stream. First, a data set is collected. The continuous images obtained by the near-infrared CMOS camera are labeled by professional medical staff through labelimg, and are divided into four categories: no thrombus, thrombus but no risk of blockage for the time being, thrombus with risk of blockage, and blockage has occurred. The labeled tags are mapped to the events to obtain a data set composed of event streams, and a classification model is obtained through training by a convolutional neural network. The event stream E is input into the classification model to output the predicted classification, and the classification of the event stream E is output and the multi-class cross-entropy loss is used as the loss function :

[0052] where is the true binary label of the event stream, is the predicted value output by the model.

[0053] In the embodiment of the present invention, the second event stream corresponding to the extracted vortex venous pot region is classified by the vortex venous pot thrombus classification model to determine the data classification result, and based on the data classification result and the preset warning conditions, vortex venous pot coagulation warning is performed, including: Using a convolutional neural network-based vortex venous pot thrombus classification model to classify each second image event information in the second event stream corresponding to the extracted vortex venous pot region, and determining the data classification result corresponding to each second image event information; When the number of data classification results that meet the preset warning conditions exceeds the recognition warning threshold, a warning signal is output, thereby performing vortex venous pot coagulation warning.

[0054] When the classification output by the vortex venous pot thrombus classification model is at risk of blockage and already blocked, it is determined that an alarm signal needs to be output, and a continuous trigger mechanism is used to reduce false alarms. When n consecutive events in the event stream are input into the classification model and the output is at risk of blockage or already blocked, the data processing module outputs an alarm signal.

[0055] Optionally, the event stream of the vortex venous pot is judged. If it is a thrombus classification, the contour extraction algorithm is used to detect the thrombus region contour. Let all pixel sets in the contour be Then the bounding box B is obtained:

[0056] where are respectively the minimum and maximum values of x and y in the pixel set of the set.

[0057] The position of the circumscribed frame of the thrombus is calibrated by the bounding box B, and the size of the thrombus is output according to the total number of thrombus pixels.

[0058] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc. that contain computer-usable program code.

[0059] The present application is described with reference to the flowcharts and / or block diagrams of methods, device systems, and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate for implementing in the process Figure 1 one process or multiple processes and / or blocks Figure 1 a device for the functions specified in one block or multiple blocks.

[0060] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements in the process Figure 1 one process or multiple processes and / or blocks Figure 1 a device for the functions specified in one block or multiple blocks.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Therefore, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 a device for the functions specified in one block or multiple blocks.

[0062] Those of ordinary skill in the art can understand that all or part of the steps in implementing the above facts and methods can be completed by instructing relevant hardware through a program. The program involved or the said program can be stored in a computer-readable storage medium. When the program is executed, it includes the following steps: At this time, the corresponding method steps are introduced. The said storage medium can be ROM / RAM, magnetic disk, optical disc, etc.

[0063] The specific embodiments described above have further elaborated on the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A vortex venous pot coagulation warning system, characterized in that Comprising: A venous chamber fixing device (1) for being mounted on a vortex venous chamber (7) to achieve fixation to a dialysis device; A macro lens (2) with a near-infrared LED light source for taking close-up shots and providing a near-infrared LED light source, and the intensity and irradiation angle of the near-infrared LED light source are adjustable; A data line (3) for data transmission; A near-infrared CMOS camera (4) for collecting image information of the vortex venous chamber (7) through the macro lens (2), and transmitting the image information to a data processing module through the data line (3) for vortex venous chamber coagulation warning; A bracket (5) for mounting the macro lens (2) and the near-infrared CMOS camera (4) so that the data collection directions of the macro lens (2) and the near-infrared CMOS camera (4) are directly facing the vortex venous chamber (7), ensuring that the vortex venous chamber (7) is within the effective collection area; A clamp (6) for mounting the bracket (5) on the venous chamber fixing device (1) to enable the near-infrared CMOS camera (4) to stably collect image information.

2. The eddy current venous chamber blood coagulation warning system according to claim 1, characterized in that, The wavelength of the near-infrared LED light source on the macro lens (2) is 850 nm.

3. The eddy current venous pot blood coagulation warning system according to claim 1, wherein The focal length of the macro lens (2) is 2.7 mm, and the viewing angle is 120 degrees.

4. The eddy current venous chamber blood coagulation warning system according to claim 1, wherein, The narrow band of the near-infrared CMOS camera (4) is 850 nm, the sampling speed is 30 frames per second, and the output resolution is 1920×1080.

5. The eddy current venous chamber blood coagulation warning system according to claim 1, characterized in that The data processing module is set as a Raspberry Pi, and the data processing module is connected to the near-infrared CMOS camera (4) using a USB communication protocol, and the data processing module communicates with a host computer through VNC to perform vortex venous chamber coagulation warning using the data processing ability of the host computer.

6. A method for early warning of vortex venous chamber blood coagulation, which is applied to the vortex venous chamber blood coagulation early warning system according to any one of claims 1 to 5, and is characterized in that, Comprising: Turn on the near-infrared LED light source on the macro lens (2) to irradiate the vortex venous chamber (7), and continuously take images of the vortex venous chamber (7) through the near-infrared CMOS camera (4) to obtain image information; Process the image information of two adjacent frames of the vortex venous chamber (7) through an event recognition algorithm to obtain first image event information, and combine multiple first image event information with timestamps to form a first event stream; Based on the image information, use an image segmentation algorithm to identify and extract the vortex venous chamber area, and construct a second event stream corresponding to the vortex venous chamber area according to the vortex venous chamber area and the first event stream; Classify the second event stream corresponding to the extracted vortex venous chamber area through a vortex venous chamber thrombus classification model to determine a data classification result, and based on the data classification result and preset warning conditions, perform vortex venous chamber coagulation warning.

7. The vortex venous pot coagulation warning method according to claim 6, characterized in that, Process the image information of two adjacent frames of the vortex venous chamber (7) through an event recognition algorithm to obtain first image event information, including: Use a differential algorithm or an optical flow analysis method to process the image information of two adjacent frames of the vortex venous chamber (7), obtain a differential event between the image information of two adjacent frames of the vortex venous chamber (7), and obtain first image event information.

8. The eddy current venous pot blood coagulation warning method according to claim 6, wherein Based on the image information, an image segmentation algorithm is used to identify and extract the area of the vortex vein ampulla, including: Segment the image information by an image segmentation algorithm based on a convolutional neural network to obtain the area of the vortex vein ampulla corresponding to the image information.

9. The eddy current venous chamber blood coagulation warning method according to claim 8, wherein According to the area of the vortex vein ampulla and the first event stream, construct a second event stream corresponding to the area of the vortex vein ampulla, including: Multiply the area of the vortex vein ampulla by the first image event information in its corresponding first event stream to obtain second image event information; Use the second image event information to construct a second event stream corresponding to the area of the vortex vein ampulla.

10. The eddy current venous pot blood coagulation warning method according to claim 9, wherein Classify the second event stream corresponding to the extracted area of the vortex vein ampulla through a vortex vein ampulla thrombus classification model to determine the data classification result, and based on the data classification result and a preset warning condition, perform a coagulation warning for the vortex vein ampulla, including: Classify each second image event information in the second event stream corresponding to the extracted area of the vortex vein ampulla by using a vortex vein ampulla thrombus classification model based on a convolutional neural network to determine the data classification result corresponding to each second image event information; When the number of data classification results that meet the preset warning condition exceeds the recognition warning threshold, an alarm signal is output, thereby performing a coagulation warning for the vortex vein ampulla.

Images