Automatic venous and arterial puncture control method, device and system

Abstract

The present invention discloses an automatic venous and arterial puncture control method for achieving high-precision puncture through precise control, comprising: after determining to initiate automatic venous and arterial puncture control, acquiring a vascular ultrasound image; establishing a venous and arterial vessel detection model based on a lightweight deep network and a target detection algorithm for the target vessel ultrasound image data; the vessel detection model is used to detect and locate the vessel; real-time acquisition of the vessel image to be detected and located, and obtaining the vessel position and category based on the trained vessel detection model; confirming the target vessel position based on the vessel positioning and classification, with the target vessel position serving as the puncture target location; planning a puncture path based on the target vessel position, including venous puncture path planning and arterial puncture path planning; and converting the path and control parameters based on the puncture path planning. Also disclosed are a device and system that have low working environment requirements, can adapt to dynamic environments, are easy to promote, reduce puncture difficulty and labor intensity, and are portable.

Description

Technical Field

[0001] The present invention relates to the technical field of medical instruments, venous puncture and arterial puncture control, and in particular to an automatic venous and arterial puncture control method, device and system. Background Art

[0002] For some patients, especially those receiving long-term infusions, it may be difficult to determine the location of the vein during venipuncture, and repeated punctures are often required. A small number of patients have lesions or injuries on the skin surface. Although blood sampling and infusion can be performed, it is very difficult to determine the location of the blood vessels and perform punctures by visual inspection.

[0003] Arteries are located deeper than veins, making it difficult to directly visualize the artery from the superficial layer of the skin. Manual arterial puncture requires palpating the arterial pulse to determine the arterial location and perform the puncture. Arterial puncture is more challenging than venous puncture, and failed punctures may result in subcutaneous hematomas and other problems. Generally, arterial puncture requires more operator experience than venous puncture.

[0004] Automatic vascular puncture equipment based on ultrasonic probe blood vessel acquisition and positioning and the application of mechanical force can significantly reduce the pain of repeated punctures for patients, reduce the risk of puncture failure, and reduce the requirements for operator experience. The venous puncture function can effectively avoid the pain caused by repeated punctures of superficial veins, and can also effectively avoid the damage to blood vessels by anti-tumor drugs and local tissue necrosis caused by drug extravasation. The arterial puncture function can effectively improve the success rate of punctures and reduce the pain and risk of repeated punctures caused by puncture failures. In recent years, robot-assisted puncture of soft tissue punctures has received increasing attention. However, traditional large-scale puncture robots have problems such as high cost, high requirements for the working environment, and limited puncture locations. Robot-assisted puncture is still difficult to achieve widespread application. Summary of the Invention

[0005] In order to solve the problems existing in the prior art, the present invention provides the following technical solutions: an automatic venous and arterial puncture control method, device and system, wherein the automatic venous and arterial puncture control method, and the portable automatic venous and arterial puncture device and system can be used by the operator to perform puncture by hand, and can achieve puncture of different parts of the human body. It has low requirements for the working environment, can adapt to dynamic environments, and is easy to promote and apply; it can achieve high-precision positioning of veins or arteries, and automatically control the puncture device to perform needle insertion, catheterization and needle withdrawal, reducing the difficulty and labor intensity of the operator relying on experience to judge the blood vessel position and perform puncture. The designed puncture device is easy to carry and adapt to dynamic scene applications.

[0006] In one aspect, the present invention provides an automatic venous and arterial puncture control method for achieving high-precision puncture through precise control, comprising:

[0007] S1, after determining to start the automatic vein and artery puncture control, collecting a blood vessel ultrasound image;

[0008] S2, establishing a venous and arterial vessel detection model based on a lightweight deep network and a target detection algorithm for the target vessel ultrasound image data; the vessel detection model is used to detect and locate the vessel;

[0009] S3, collecting images of the blood vessels to be detected and located in real time, and obtaining the blood vessel positions and categories based on the trained blood vessel detection model;

[0010] S4, confirming the target blood vessel position based on the blood vessel positioning and classification, wherein the target blood vessel position is the puncture target position;

[0011] S5, performing puncture path planning based on the target blood vessel position, including: venous puncture path planning and arterial puncture path planning;

[0012] S6: Convert the path and control parameters based on the puncture path planning.

[0013] Preferably, the lightweight deep network includes YOLOv8, NanoDet or NanoDet-Plus.

[0014] Preferably, the S2 includes:

[0015] S21, establishing a basic blood vessel detection model, manually marking blood vessel positions in the collected blood vessel ultrasound image to obtain a manually marked image;

[0016] S22: Performing model training on the basic blood vessel detection model based on the manually labeled image to obtain a trained blood vessel detection model.

[0017] Preferably, the S3 includes:

[0018] S31, using an ultrasound device to capture an image of the blood vessel to be detected and located in real time, inputting the image of the blood vessel to be detected and located into the trained blood vessel detection model to obtain all blood vessels in the image as a first detection result;

[0019] S32, integrating the first detection result with a color ultrasound venous and arterial vessel classification method to obtain high-precision vessel positioning and classification, which is used for subsequent target vessel selection and puncture calculation, including:

[0020] (1) Selecting a detection target area based on the first detection result: selecting a location where a blood vessel may exist in the B-ultrasound image according to the puncture target, and intercepting a portion of the entire image as the detection target area;

[0021] (2) performing blood vessel positioning and rough classification on the first detection result within the detection target area using a deep learning-based blood vessel positioning and rough classification method, including: obtaining a positioning detection result based on the deep learning-based blood vessel positioning method as a circumscribed rectangle or square of the blood vessel; classifying veins and arteries as a relatively rough classification result without precise positioning of the blood vessel edge; and using the classification result as reference information to further perform blood vessel edge fitting within the detection frame;

[0022] (3) performing high-precision classification of veins and arteries based on a color Doppler method or a spectral Doppler method, thereby obtaining high-precision differentiation of veins and arteries corresponding to the coarse classification;

[0023] (4) Venous and arterial classification and vascular edge fitting, including:

[0024] A. If the coarse classification based on veins or arteries is consistent with the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method, the vascular type is determined. If the coarse classification based on veins or arteries is inconsistent with the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method, and the confidence level of the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method is high, the high-precision classification of veins and arteries is directly adopted. If the coarse classification based on veins or arteries is inconsistent with the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method, and the confidence level of the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method is low, further confirmation is prompted.

[0025] B. After the vessel detection results are confirmed and accurate vein and artery classification is obtained, high-precision vessel edge fitting is performed. This includes: within the detected vascular rectangular region, the ultrasound image and the colored image are binarized and fused, followed by Hough transform circle or ellipse detection to obtain a high-precision circle or ellipse fitting result for the vessel edge. The method for binarizing and fusing the ultrasound image and the colored image is as follows: both the ultrasound image and the colored image are converted into consistent binary or grayscale images, with 0 representing the vessel region and 1 representing other regions. The two images are then added together to make the vessel region more distinct.

[0026] Preferably, the prompt further confirms that:

[0027] Perform mutual human-machine pressure-based venous confirmation, including: when the blood vessel type cannot be automatically determined with high precision, prompt for manual pressure confirmation, apply appropriate pressure to the probe to observe the image, and if the blood vessel does not change with the pressure, it is judged to be an artery; if it gradually closes with the increase of pressure, it is judged to be a vein; determine the target blood vessel by manually clicking on the screen, that is, input the system vein and artery information.

[0028] Preferably, the S4 includes:

[0029] S41, select arteriovenous targets according to working status, including:

[0030] (1) If the current working mode is venipuncture, select an available venous vessel with a suitable distance from the skin as the target vessel. If there are multiple available vessels, they constitute a candidate set. The suitable distance from the skin requires the selection of the vessel type according to the puncture task;

[0031] (2) If the current working mode is arterial puncture, the artery closest to the skin is selected as the target vessel. If there are multiple available vessels, they constitute the candidate set;

[0032] S42, optimal target vessel selection based on the location and correlation of the vessels;

[0033] 1. Venous target vessel selection strategy: The thickest vessel is selected for puncture; there are no other vessels between the punctured vessel and the skin; there is no overlapping of vessels; the venous target vessel selection strategy is completed through geometric calculation and comparison;

[0034] l Optimal strategy for selecting target arterial vessels: When the thickness meets the requirements, the vessel should be as close to the skin as possible; there should be no other vessels between the punctured vessel and the skin; and there should be no overlapping of vessels.

[0035] Preferably, the venipuncture path planning includes:

[0036] (1) Connect the center of the target vessel with the center of the needle direction and determine whether there are other veins in the connection path;

[0037] (2) If there are no other veins, the planning is successful and the angle between the needle and the skin and the insertion length can be calculated;

[0038] (3) If there are other veins, determine whether the veins in the path are other veins that meet the puncture requirements, replace the other veins that meet the puncture requirements with the target vessels, and re-plan the puncture path. If the puncture requirements are not met, return to select other veins that meet the puncture requirements as the target vessels; the puncture requirements include: the veins and the thickness of the vessels meet the puncture requirements;

[0039] (4) If all other veins cannot meet the puncture requirements, it indicates that the position is inappropriate and the operator needs to adjust it;

[0040] The arterial puncture path planning is a planning strategy for searching for feasible paths from the vertical direction to both sides, including:

[0041] (1) Connect the center of the target vessel with the center of the needle direction selection to determine whether there are other veins or arteries in the connection path;

[0042] (2) If there are no other veins or arteries, the planning is successful and the angle between the needle and the skin and the insertion length can be calculated;

[0043] (3) If there are other veins or arteries, it will be prompted that puncture cannot be performed safely and the fixed position of the needle needs to be readjusted; if the needle has a lateral motion control mechanism, the needle position is adjusted laterally according to a fixed step length for searching; if the needle does not have a lateral motion control mechanism, it will be prompted that the position is inappropriate and the operator needs to adjust it.

[0044] Preferably, the S6 includes:

[0045] S61, after the puncture path is successfully planned, the control device compares the current angle and needle tip position information of the needle body to be punctured with the planned puncture path to form displacement vector information; the displacement vector information includes axial displacement and angular displacement;

[0046] S62, the control device converts the displacement vector information into a control signal and transmits it to the driving device;

[0047] S63, the driving device receives the displacement vector information, outputs an electrical signal, and controls the motor movement of the automatic venous and arterial puncture device to perform puncture.

[0048] A second aspect of the present invention is to provide an automatic venous and arterial puncture device, which performs automatic venous and arterial puncture based on the control method described in the first aspect; the device comprises: a fixing device, a guide device, and a drive device; the fixing device is slidably connected to the guide device, and the fixing device is fixedly connected to the drive device; wherein:

[0049] The fixing device is used to fix the needle body to be punctured and the hose to be inserted; the hose to be inserted is fixed to the fixing device through the needle body to be punctured;

[0050] The guide device is used to determine the target blood vessel position, that is, the puncture target position;

[0051] The driving device is used to drive the fixing device to move along the direction guided by the guiding device;

[0052] The fixing device includes: a syringe, an inner cylinder, an outer cylinder, a first connecting piece, a fixing plate, a needle clamp and a needle withdrawal button; the hose to be inserted is fixedly arranged on the syringe through the needle body to be punctured; the syringe is arranged in the inner cylinder and fixedly connected to the inner cylinder; the inner cylinder is arranged in the outer cylinder through the first connecting piece, and the outer cylinder can be slidably connected to the inner cylinder; the outer cylinder is fixedly arranged on the fixing plate through the first connecting piece; the needle clamp cooperates with the needle withdrawal button to trigger the needle withdrawal button under the action of external force, so that the needle body to be punctured and the hose to be inserted are separated from the fixing device; after the needle body to be punctured is fixed on the fixing device, the relative spatial position between its initial position and the ultrasonic probe is determined and no longer changes. After the fixing device is calibrated and used for the control algorithm, puncture control based on ultrasonic image positioning can be realized;

[0053] The guide device comprises three parts: an ultrasonic probe, a needle guide, and an outer cylinder guide, which is used to guide the fixing device and the needle to be punctured to move in a fixed direction based on ultrasound positioning of the blood vessel; the ultrasonic probe is used to collect data information of a preselected target blood vessel through ultrasound and send the data information to the control device; the needle guide is fixedly mounted on the ultrasonic probe, and the needle to be punctured is mounted on the needle guide; the outer cylinder guide is fixedly mounted on the fixing plate, and the outer cylinder is mounted on the outer cylinder guide;

[0054] The ultrasonic probe is used in conjunction with the ultrasonic coupling sleeve, which is installed at the front end of the ultrasonic probe in an installation manner that is removable multiple times and easy to remove;

[0055] The driving device is connected to the control device and is used to drive the fixing device to move along the direction guided by the guide device; the driving device is divided into a needle control part driving device and a tube placement part driving device; the needle control part driving device includes: a needle control motor and a second connecting piece; the needle control motor is fixedly arranged on the fixing plate and is fixedly connected to the inner cylinder through the second connecting piece; the tube placement part driving device includes: a tube placement motor, a third connecting piece, an outer cylinder connecting rod and a motor connecting rod; the tube placement motor is fixedly arranged on the fixing plate through the motor connecting rod, the tube placement motor is fixedly connected to the third connecting piece, and the third connecting piece is connected to the outer cylinder through the outer cylinder connecting rod.

[0056] A third aspect of the present invention provides a venous arterial puncture system, which is completed by a control device and the automatic venous arterial puncture device according to the second aspect, wherein the automatic venous arterial puncture device is connected to the control device, wherein:

[0057] The automatic venous and arterial puncture device collects data information of a preselected target blood vessel through an ultrasonic probe and sends the data information of the preselected target blood vessel to the control device;

[0058] The control device is used to receive data information of the preselected target blood vessel, calculate the position information of the target blood vessel based on the data information, and generate a control signal for controlling the operation of the automatic venous and arterial puncture device based on the position information; the automatic venous and arterial puncture device controls the needle body to be punctured and the soft tube to be inserted to puncture the skin based on the control signal;

[0059] The driving device accurately controls the movement step lengths of the needle control motor and the tube placement motor; the control device converts the blood vessel information collected by the ultrasonic probe into the position information and puncture path of the target blood vessel based on algorithm identification and optimization; the control device compares the current coordinate information of the puncture point and the needle body to be punctured, calculates the displacement vector, and transmits it to the driving device in the form of a control signal; the driving device receives the control signal and moves to the puncture point according to the control signal.

[0060] A fourth aspect of the present invention provides an electronic device, comprising a processor and a memory, wherein the memory stores a plurality of instructions, and the processor is configured to read the instructions and execute the method described in the first aspect.

[0061] A fifth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a plurality of instructions, and the plurality of instructions can be read by a processor to execute the method described in the first aspect.

[0062] The control method, device, and system provided by the present invention have the following beneficial effects:

[0063] Compared with manual puncture or mechanically assisted medical robot-operated puncture in the prior art, the venipuncture device of this patent has the following advantages:

[0064] 1. The overall cost is lower than that of the robotic medical system, it is more practical, all-in-one, portable, and adaptable to dynamic environments.

[0065] 2. The venous puncture device forms a mechanical connection with the ultrasonic probe and cooperates with the control device to realize automatic control of the puncture operation. The operator does not need to participate in the puncture action, which improves efficiency, reduces dependence on the operator's experience, and is not affected by the operator's operating habits.

[0066] 3. The blood vessel positioning algorithm based on ultrasound images is divided into venous and arterial control. The guide device in the venous puncture device uses an ultrasonic probe to locate the veins and arteries. Compared with manual puncture that relies on visual observation and hand touch to determine the position, the blood vessel positioning accuracy is greatly improved, which can ensure the accuracy of puncture.

[0067] 4. The ultrasonic probe is connected to the ultrasonic coupling sleeve. The coupling sleeve is a disposable item and features an easily removable, reusable connection. The coupling sleeve replaces the need for applying conventional ultrasonic coupling agent to the skin, preventing contamination of the puncture site. The coupling sleeve ensures that the ultrasound probe imaging requirements are met while maintaining the sterility of the puncture site.

[0068] 5. The ultrasound coupling sleeve is a disposable, sterile product that can be disinfected and aligned on the outer surface. The skin contact area can still be punctured, making it convenient to use. The ultrasound probe does not come into direct contact with the patient's skin, so it can be kept clean and help extend the service life of the ultrasound probe.

[0069] 6. Error compensation between ultrasound probe viewing angle and needle viewing angle. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 This is a flow chart of the automatic venous and arterial puncture control method described in the present invention.

[0071] Figure 2 This is a schematic diagram of the YOLOv8 network structure described in the present invention.

[0072] Figure 3 Schematic diagram of the NanoDet network structure described in the present invention.

[0073] Figure 4 Schematic diagram of the NanoDet-Plus network structure described in the present invention.

[0074] Figure 5 This is a schematic diagram of the manually labeled veins and arteries in the figure according to the present invention.

[0075] Figure 6 Schematic diagram of the positioning and classification process of veins and arteries according to the present invention.

[0076] Figure 7 Schematic diagram of the classification results of veins and arteries according to the present invention.

[0077] Figure 8 This is a schematic diagram of the venous blood vessel positioning according to the present invention.

[0078] Figure 9 Schematic diagram of the basilic vein of the present invention.

[0079] Figure 10 These are the great saphenous vein map and the small saphenous vein map described in the present invention.

[0080] Figure 11 This is a schematic structural diagram of the ultrasonic coupling sleeve described in the present invention.

[0081] Figure 12 The figure is a schematic structural diagram of an embodiment of the electronic device according to the present invention. DETAILED DESCRIPTION

[0082] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0083] like Figure 1 As shown, the present invention provides an automatic venous and arterial puncture control method for achieving high-precision puncture through precise control, comprising:

[0084] S1, after determining to start the automatic vein and artery puncture control, collecting a blood vessel ultrasound image;

[0085] As a preferred embodiment, S1 includes: acquiring ultrasound images of blood vessels of multiple body parts that may be punctured, such as the back of the hand, forearm, upper arm, etc., and the acquisition is performed based on multiple people and multiple adjustments.

[0086] S2, establishing a venous and arterial vessel detection model based on a lightweight deep network and a target detection algorithm for the target vessel ultrasound image data; the vessel detection model is used to detect and locate the vessel;

[0087] In this embodiment, to enable portable device deployment and ensure puncture efficiency, a lightweight deep network is used to detect and locate blood vessels. For example, YOLOv8, NanoDet, or NanoDet-Plus are used. The lightweight deep network is the basis of the venous and arterial blood vessel detection and positioning algorithm.

[0088] (1) If Figure 2 The figure shows a schematic diagram of the YOLOv8 network structure used in this embodiment.

[0089] YOLOv8 improves and upgrades the previous seven versions of the YOLO detection algorithm. The main improvements include: using the C2f module in the backbone, using the anchor-free + Decoupled-head detection head, using the BCELoss loss function for classification and a combination of CIoU and VFL for regression, changing the box matching strategy from static matching to Task-AlignedAssigner matching, and disabling Mosaic in the last 10 epochs.

[0090] (2) If Figure 3 Shown is a schematic diagram of the NanoDet network structure used in this embodiment.

[0091] NanoDet is a FCOS-style single-stage anchor-free target detection model that uses ATSS for target sampling and Generalized FocalLoss loss function for classification and box regression. The network structure diagram is as follows Figure 3 NanoDet uses the GeneralizedFocalLoss loss function. This function removes the Centerness branch of FCOS, eliminating a large number of convolutions on this branch, thereby reducing the computational overhead of the detection head and making it very suitable for lightweight deployment on mobile devices.

[0092] (3) If Figure 4 Shown is a schematic diagram of the NanoDet-Plus network structure used in this embodiment.

[0093] The NanoDet-Plus detection model has designed a simpler and lighter training auxiliary module AssignGuidanceModule (AGM) and a dynamic soft label assignment strategy DynamicSoftLabelAssigner (DSLA) to solve the optimal label matching problem in the lightweight model. The overall architecture of NanoDet-Plus is as follows Figure 4 As shown in Figure 2, it improves Ghost-PAN in feature fusion, improves the convolution kernel size and downsampling layer in the detection head, and modifies the training optimization strategy.

[0094] As a preferred embodiment, the S2 includes:

[0095] S21, establishing a basic blood vessel detection model, manually marking blood vessel positions in the collected blood vessel ultrasound image to obtain a manually marked image;

[0096] In this embodiment, the veins and arteries in the figure are manually marked as follows. Figure 5 As shown in Figure 2, it is difficult to distinguish between venous vessels and arterial vessels in a single static short-axis B-ultrasound image frame.

[0097] S22: Performing model training on the basic blood vessel detection model based on the manually labeled image to obtain a trained blood vessel detection model.

[0098] S3, collecting images of the blood vessels to be detected and located in real time, and obtaining the blood vessel positions and categories based on the trained blood vessel detection model.

[0099] As a preferred embodiment, the S3 includes:

[0100] S31, using an ultrasound device to capture an image of the blood vessel to be detected and located in real time, inputting the image of the blood vessel to be detected and located into the trained blood vessel detection model to obtain all blood vessels in the image as a first detection result;

[0101] S32, integrating the first detection result with the color ultrasound venous and arterial blood vessel classification method to obtain high-precision blood vessel positioning and classification, which is used for subsequent target vessel selection and puncture calculation.

[0102] In this embodiment, in actual application, the above S32 can also be combined with the operator's pressing of the vein to confirm the blood vessels to further ensure the safety of the operation. The process design of this part is shown in the figure below.

[0103] like Figure 6 As shown, as a preferred embodiment, the S32 includes:

[0104] (1) Selecting the target detection area based on the first detection result: According to the puncture target, select the location where the blood vessels may exist in the B-ultrasound image, and intercept a portion of the entire image as the target detection area. The advantage of this is to reduce false detection and reduce the complexity of post-processing;

[0105] (2) The deep learning-based vascular positioning and coarse classification method performs vascular positioning and coarse classification on the first detection result within the detection target area, including: the positioning detection result obtained by the deep learning-based vascular positioning method is a circumscribed rectangle or square of the blood vessel; the classification of veins and arteries is a relatively coarse classification result without precise positioning of the blood vessel edge. The classification result is used as reference information to further perform blood vessel edge fitting within the detection frame;

[0106] (3) Perform high-precision vein and artery classification based on color Doppler or spectral Doppler methods, thereby obtaining high-precision vein and artery distinction corresponding to the rough classification; in this embodiment, as follows Figure 6 As shown in the figure, different regions generally present irregular edges. In addition to real blood vessels, some non-vascular regions are also detected. Combining the results of deep learning vascular localization, the real veins and arteries can be screened out, such as Figure 7 shown.

[0107] Vein-artery classification and vascular edge fitting, including:

[0108] A. If the coarse classification based on veins or arteries is consistent with the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method, the vascular type is determined. If the coarse classification based on veins or arteries is inconsistent with the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method, and the confidence level of the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method is high, the high-precision classification of veins and arteries is directly adopted. If the coarse classification based on veins or arteries is inconsistent with the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method, and the confidence level of the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method is low, further confirmation is prompted.

[0109] B. After the blood vessel detection results are confirmed and the accurate classification of veins and arteries is obtained, high-precision blood vessel edge fitting is performed. Figure 8 For example, within the detected rectangular region of a blood vessel, the ultrasound image and the colored image are binarized and fused, then Hough transform circle or ellipse detection is performed to obtain a high-precision circle or ellipse fitting result for the vessel edge. Hough transform circle or ellipse detection methods are relatively mature and can be used in common applications. The fitting results are used in subsequent steps.

[0110] The method for binarizing and fusing the ultrasound image and the colored image is as follows: both the ultrasound image and the colored image are converted into a consistent binary or grayscale image, with the blood vessel area represented by 0 and the other areas represented by 1. The two images are added together to make the blood vessel area more prominent.

[0111] The Hough transform is a method used to detect specific shapes in images. It can detect a variety of geometric shapes, including lines, circles, and ellipses. Its principle is based on the mapping relationship between image space and parameter space. For Hough circle or ellipse detection, the principle can be summarized as follows:

[0112] Edge detection: Edge detection is required for the image in order to determine all potential circular or elliptical edge points in the image.

[0113] Parametric mapping: Mapping edge points in image space to parameter space. For a circle, this means mapping each edge point to a point (x, y, r) in three-dimensional space, where (x, y) are the coordinates of the edge point and r is the radius from the center of the image to that point. For an ellipse, the parameter space is more complex because additional parameters need to be considered, such as the lengths of the major and minor axes and the rotation angle of the ellipse.

[0114] Voting mechanism: For each possible center position in parameter space, the gradient vectors from all edge points to that center are calculated. The intersection of these vectors is the potential center position. For ellipses, this process is more complex because more parameters need to be considered to determine the specific position and shape of the ellipse.

[0115] Peak detection: Find the points in parameter space with the most votes, which correspond to the centers of the most likely circular or elliptical shapes in the image.

[0116] Shape Verification: Finally, the detected shapes are verified to conform to the expected circular or elliptical properties by checking for peaks in the parameter space.

[0117] As a preferred embodiment, the prompt further confirms that:

[0118] Perform mutual human-machine pressure-based venous confirmation, including: when the blood vessel type cannot be automatically determined with high precision, prompt for manual pressure confirmation, apply appropriate pressure to the probe to observe the image, and if the blood vessel does not change with the pressure, it is judged to be an artery; if it gradually closes with the increase of pressure, it is judged to be a vein; determine the target blood vessel by manually clicking on the screen, that is, input the system vein and artery information.

[0119] After the above steps are completed, the detailed information of the blood vessels in the ultrasound image at the current position can be obtained, and the process proceeds to step S4.

[0120] S4, confirming the target blood vessel position based on the blood vessel positioning and classification, wherein the target blood vessel position is the puncture target position;

[0121] In this embodiment, multiple different types of blood vessels may be detected in the image. It is necessary to first determine the target blood vessel range and then select the target blood vessel.

[0122] As a preferred embodiment, the S4 includes:

[0123] S41, select arteriovenous target according to working status:

[0124] In this embodiment:

[0125] (1) If the current working mode is venipuncture, select an available vein with a suitable distance from the skin (displayed as a circle with the largest radius in the image) as the target vessel. If there are multiple available vessels, they constitute a candidate set. The suitable distance from the skin here mainly requires the selection of the vessel type according to the puncture task. Currently, venipuncture is mainly performed on the basilic vein, such as Figure 9 Generally, the great saphenous vein or small saphenous vein that is very close to the skin is not punctured. Figure 10 As shown. Figure 10 The left side shows the greater saphenous vein, and the right side shows the lesser saphenous vein.

[0126] Since the distribution positions of the basilic vein and the great and small saphenous veins on B-ultrasound images are quite different, a simple threshold range can be used. For example, if the coordinates on the image are greater than a threshold (this threshold is related to the puncture site, color ultrasound equipment characteristics and parameter adjustment, and can be preset after the equipment and site are determined), the great and small saphenous veins can be excluded.

[0127] (2) If the current working mode is arterial puncture, the artery closest to the skin (the image shows the artery target closest to the top) is selected as the target vessel. If there are multiple available vessels, they constitute a candidate set.

[0128] S42, selecting the optimal target blood vessel based on the position and correlation of the blood vessels in the graph.

[0129] Target venous vessel selection: Generally, the largest vessel is selected for puncture; there should be no other vessels between the punctured vessel and the skin; and there should be no overlapping vessels. These selection strategies can be achieved through geometric calculation and comparison.

[0130] l Target arterial vessel selection: The target artery should be as close to the skin as possible while meeting the required thickness; there should be no other blood vessels between the punctured vessel and the skin; and there should be no overlapping blood vessels.

[0131] S5, performing puncture path planning based on the target blood vessel position, including: venous puncture path planning and arterial puncture path planning.

[0132] In this embodiment, the steel needle is a rigid object and is planned as a line. The objectives of puncture needle path planning include: the skin penetration position, the angle between the steel needle and the skin, and the penetration length (determining the penetration depth).

[0133] As a preferred embodiment, the venipuncture path planning includes:

[0134] (1) Connect the center of the target vessel with the center of the needle direction and determine whether there are other veins in the connection path;

[0135] (2) If there are no other veins, the planning is successful and the angle between the needle and the skin and the insertion length can be calculated;

[0136] (3) If there are other veins, determine whether the veins in the path are other veins that meet the puncture requirements (the two necessary points of the puncture requirements are: the vein and the thickness of the vein meet the puncture requirements). Replace the other veins that meet the puncture requirements with the target vessels and re-plan the puncture path. If the puncture requirements are not met, return to select other veins that meet the puncture requirements as the target vessels.

[0137] (4) If all other veins cannot meet the puncture requirements, it indicates that the position is inappropriate and the operator needs to adjust it.

[0138] As a preferred embodiment, the arterial puncture path planning: Since arteries are located deeper, there are often veins between them and the skin. The puncture needle must avoid these veins during the puncture process. Generally, the puncture process is to minimize the length of the needle in the body. Therefore, this embodiment designs a planning strategy that searches for feasible paths from the vertical direction to both sides:

[0139] (1) Connect the center of the target blood vessel with the center of the needle direction selection to determine whether there are other veins or arteries in the connection path.

[0140] (2) If there are no other veins or arteries, the planning is successful and the angle between the needle and the skin and the insertion length can be calculated;

[0141] (3) If there are other veins or arteries, it will be prompted that puncture cannot be performed safely and the fixed position of the needle needs to be readjusted; if the needle has a lateral motion control mechanism, the needle position is adjusted laterally according to a fixed step length for searching; if the needle does not have a lateral motion control mechanism, it will be prompted that the position is inappropriate and the operator needs to adjust it.

[0142] S6: Convert the path and control parameters based on the puncture path planning.

[0143] As a preferred embodiment, the S6 includes:

[0144] S61, after the puncture path is successfully planned, the control device compares the current angle and needle tip position information of the needle body to be punctured with the planned puncture path to form displacement vector information; the displacement vector information includes axial displacement and angular displacement;

[0145] S62, the control device converts the displacement vector information into a control signal and transmits it to the driving device;

[0146] S63, the driving device receives the displacement vector information, outputs an electrical signal, and controls the motor movement of the automatic venous and arterial puncture device to perform puncture.

[0147] A second aspect of the present invention is to provide an automatic venous and arterial puncture device, which performs automatic venous and arterial puncture based on the control method described in the first aspect; the device comprises: a fixing device, a guide device, and a drive device; the fixing device is slidably connected to the guide device, and the fixing device is fixedly connected to the drive device; wherein:

[0148] The fixing device is used to fix the needle body to be punctured and the hose to be inserted; the hose to be inserted is fixed to the fixing device through the needle body to be punctured;

[0149] The guide device is used to determine the target blood vessel position, that is, the puncture target position;

[0150] The driving device is used to drive the fixing device to move along the direction guided by the guiding device.

[0151] (1) Fixing device

[0152] The fixing device includes: a needle cylinder, an inner cylinder, an outer cylinder, a first connecting piece, a fixing plate, a needle clamp and a needle withdrawal button.

[0153] Structural description: The soft tube to be inserted is fixedly set on the syringe through the needle body to be punctured; the syringe is set in the inner cylinder and fixedly connected to the inner cylinder; the inner cylinder is set in the outer cylinder through the first connecting piece, and the outer cylinder can be slidably connected to the inner cylinder; the outer cylinder is fixedly set on the fixed plate through the first connecting piece; the needle card cooperates with the needle withdrawal button to trigger the needle withdrawal button under the action of external force, so that the needle body to be punctured and the soft tube to be inserted are separated from the fixing device; after the needle body to be punctured is fixed on the fixing device, the relative spatial position between its initial position and the ultrasonic probe is determined and no longer changes. After the fixing device is calibrated and used for the control algorithm, puncture control based on ultrasonic image positioning can be realized.

[0154] (2) Guide device

[0155] The guiding device comprises three parts: an ultrasonic probe, a needle guide and an outer cylinder guide, and is used to guide the fixing device and the needle to be punctured to move in a fixed direction based on ultrasound positioning of the blood vessel.

[0156] 1. Ultrasonic probe, used to collect data information of pre-selected target blood vessels through ultrasound and send the data information to the control device.

[0157] 2. A needle guide is fixedly mounted on the ultrasonic probe, and the needle to be punctured is mounted on the needle guide.

[0158] 3. An outer cylinder guide, fixedly arranged on the fixing plate, and the outer cylinder is mounted on the outer cylinder guide.

[0159] As a preferred embodiment, the ultrasonic probe is used in conjunction with the ultrasonic coupling sleeve, which is installed at the front end of the ultrasonic probe in a manner that allows for multiple and easy removal. Figure 11 The figure shows a schematic diagram of the structure of the ultrasonic coupling sleeve, which is mainly used to ensure that the artery is sterile.

[0160] (3) Drive device

[0161] The driving device is connected to the control device and is used to drive the fixing device to move along the direction guided by the guide device; the driving device is divided into a needle control part driving device and a tube placement part driving device.

[0162] 1. The needle control part driving device includes: a needle control motor and a second connecting member; the needle control motor is fixedly arranged on the fixing plate and fixedly connected to the inner cylinder through the second connecting member.

[0163] 2. The driving device of the tube placement part includes: a tube placement motor, a third connecting piece, an outer cylinder connecting rod and a motor connecting rod; the tube placement motor is fixedly arranged on the fixed plate through the motor connecting rod, the tube placement motor is fixedly connected to the third connecting piece, and the third connecting piece is connected to the outer cylinder through the outer cylinder connecting rod.

[0164] A third aspect of the present invention provides a venous arterial puncture system, which is completed by a control device and the automatic venous arterial puncture device according to the second aspect, wherein the automatic venous arterial puncture device is connected to the control device, wherein:

[0165] The automatic venous and arterial puncture device collects data information of a preselected target blood vessel through an ultrasonic probe and sends the data information of the preselected target blood vessel to the control device;

[0166] The control device is used to receive data information of the preselected target blood vessel, calculate the position information of the target blood vessel based on the data information, and generate a control signal for controlling the operation of the automatic venous and arterial puncture device based on the position information; the automatic venous and arterial puncture device controls its own fixed needle body to be punctured and the hose to be inserted to puncture the skin according to the control signal.

[0167] The driving device accurately controls the movement step lengths of the needle control motor and the tube motor; the control device converts the blood vessel information collected by the ultrasonic probe into the position information and puncture path of the target blood vessel based on algorithm identification and optimization; the control device compares the current coordinate information of the puncture point and the needle body to be punctured, calculates the displacement vector, and transmits it to the driving device in the form of a control signal; the driving device receives the control signal and moves to the puncture point according to the control signal.

[0168] As a preferred embodiment, the venous arterial puncture system also includes a portable industrial computer or an all-in-one computer, and the portable industrial computer or the all-in-one computer and the B-ultrasound machine are placed together to achieve portability; the AI in the portable industrial computer is used to calculate the computing resources of the board / pad.

[0169] The present invention also provides a memory storing a plurality of instructions, wherein the instructions are used to implement the method as in the first embodiment.

[0170] like Figure 12 As shown, the present invention also provides an electronic device, including a processor 301 and a memory 302 connected to the processor 301, the memory 302 stores multiple instructions, and the instructions can be loaded and executed by the processor to enable the processor to execute the method as in embodiment 1.

[0171] Although preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they are aware of the basic inventive concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the invention. Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the invention. Thus, the present invention is intended to include such changes and modifications as fall within the scope of the claims and their equivalents.

Claims

1. An automatic vein and artery puncture control method, characterized in that: For high-precision puncture with precise control, including: S1, after determining to start the automatic vein and artery puncture control, collecting a blood vessel ultrasound image; S2, for the target blood vessel ultrasound image data, establishing a venous and arterial blood vessel detection model based on a lightweight deep network and a target detection algorithm; the blood vessel detection model is used to detect and locate the blood vessel; S3, real-time acquisition of the image of the blood vessels to be detected and located, and obtaining the blood vessel location and category based on the trained blood vessel detection model; S4, confirming the target blood vessel position based on blood vessel positioning and classification, wherein the target blood vessel position is the puncture target position; S5, performing puncture path planning based on the target blood vessel position, including: venous puncture path planning and arterial puncture path planning; S6, converting the path and control parameters based on the puncture path planning; The S2 includes: S21, establishing a basic blood vessel detection model, manually marking blood vessel positions in the collected blood vessel ultrasound image to obtain a manually marked image; S22, performing model training on the basic blood vessel detection model based on the manually annotated image to obtain a trained blood vessel detection model; The S3 includes: S31, using an ultrasound device to capture an image of the blood vessel to be detected and located in real time, inputting the image of the blood vessel to be detected and located into the trained blood vessel detection model to obtain all blood vessels in the image as a first detection result; S32, integrating the first detection result with a color ultrasound venous and arterial vessel classification method to obtain high-precision vessel positioning and classification, which is used for subsequent target vessel selection and puncture calculation, including: (1) Selecting a detection target area based on the first detection result: selecting a location where a blood vessel may exist in the B-ultrasound image according to the puncture target, and intercepting a portion of the entire image as the detection target area; (2) performing blood vessel positioning and rough classification on the first detection result within the detection target area using a deep learning-based blood vessel positioning and rough classification method, including: obtaining a positioning detection result based on the deep learning-based blood vessel positioning method as a circumscribed rectangle or square of the blood vessel; classifying veins and arteries as a relatively rough classification result without precise positioning of the blood vessel edge; and using the classification result as reference information to further perform blood vessel edge fitting within the detection frame; (3) performing high-precision vein and artery classification based on a color Doppler method or a spectral Doppler method, thereby obtaining high-precision vein and artery differentiation corresponding to the coarse classification; (4) Venous and arterial classification and vascular edge fitting, including: A. If the coarse classification based on veins or arteries is consistent with the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method, the vascular type is determined. If the coarse classification based on veins or arteries is inconsistent with the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method, and the confidence level of the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method is high, the high-precision classification of veins and arteries is directly adopted. If the coarse classification based on veins or arteries is inconsistent with the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method, and the confidence level of the high-precision classification of veins and arteries based on the color Doppler effect or spectral Doppler method is low, further confirmation is prompted. B. After the vessel detection results are confirmed and accurate vein and artery classification is obtained, high-precision vessel edge fitting is performed. This includes: within the detected vessel rectangular region, the ultrasound image and the colored image are binarized and fused, followed by Hough transform circle or ellipse detection to obtain a high-precision circle or ellipse fitting result for the vessel edge. The method for binarizing and fusing the ultrasound image and the colored image is as follows: both the ultrasound image and the colored image are converted to consistent binary or grayscale images, with 0 representing the vessel region and 1 representing other regions. The two images are then added together to make the vessel region more distinct. The prompt further confirms that: Performing human-machine mutual pressure-based venous confirmation, including: When the vessel type cannot be automatically determined with high accuracy, prompting manual pressure confirmation, applying appropriate pressure to the probe and observing the image, if the vessel does not change with the pressure, it is judged to be an artery; if it gradually closes with increasing pressure, it is judged to be a vein; manually clicking on the screen to determine the target vessel, that is, inputting system venous and arterial information; The S4 includes: S41, select arteriovenous targets according to working status, including: (1) If the current working mode is venipuncture, select an available venous vessel with a suitable distance from the skin as the target vessel. If there are multiple available vessels, they constitute a candidate set. The suitable distance from the skin requires the selection of the vessel type according to the puncture task; (2) If the current working mode is arterial puncture, the artery closest to the skin is selected as the target vessel. If there are multiple available vessels, they constitute the candidate set; S42, optimal target vessel selection based on vascular location and correlation; including: Venous target vessel selection strategy: The thickest vessel is selected for puncture; there are no other vessels between the punctured vessel and the skin; there is no overlapping of vessels; the venous target vessel selection strategy is completed through geometric calculation and comparison; The optimal strategy for selecting the target artery vessel is as close to the skin as possible while the thickness meets the requirements; there are no other blood vessels between the punctured vessel and the skin; and there is no overlapping of blood vessels.

2. The automatic vein and artery puncture control method according to claim 1, characterized in that: The lightweight deep network includes YOLOv8, NanoDet or NanoDet-Plus.

3. The automatic vein and artery puncture control method according to claim 2, characterized in that: The venipuncture path planning includes: (1) Connect the center of the target vessel with the center of the needle direction and determine whether there are other veins in the connection path; (2) If there are no other veins, the planning is successful and the angle between the needle and the skin and the insertion length can be calculated; (3) If there are other veins, determine whether the veins in the path are other veins that meet the puncture requirements, replace the other veins that meet the puncture requirements with the target vessels, and re-plan the puncture path. If the puncture requirements are not met, return to select other veins that meet the puncture requirements as the target vessels; the puncture requirements include: the veins and the thickness of the vessels meet the puncture requirements; (4) If all other veins cannot meet the puncture requirements, it indicates that the position is inappropriate and the operator needs to adjust it; The arterial puncture path planning is a planning strategy for searching for feasible paths from the vertical direction to both sides, including: (1) Connect the center of the target vessel with the center of the needle direction selection to determine whether there are other veins or arteries in the connection path; (2) If there are no other veins or arteries, the planning is successful and the angle between the needle and the skin and the insertion length can be calculated; (3) If there are other veins or arteries, it will be prompted that puncture cannot be performed safely and the fixed position of the needle needs to be readjusted; if the needle has a lateral motion control mechanism, the needle position is adjusted laterally according to a fixed step length for searching; if the needle does not have a lateral motion control mechanism, it will be prompted that the position is inappropriate and the operator needs to adjust it.

4. An automatic venous arterial puncture device, which performs automatic venous arterial puncture based on the control method according to any one of claims 1 to 3, characterized in that: Including: fixing device, guiding device and driving device; The fixing device is slidably connected to the guide device, and the fixing device is fixedly connected to the driving device; wherein: The fixing device is used to fix the needle body to be punctured and the hose to be inserted; the hose to be inserted is fixed to the fixing device through the needle body to be punctured; The guide device is used to determine the target blood vessel position, that is, the puncture target position; The driving device is used to drive the fixing device to move along the direction guided by the guiding device; The fixing device includes: a syringe, an inner cylinder, an outer cylinder, a first connecting piece, a fixing plate, a needle clamp and a needle withdrawal button; the hose to be inserted is fixedly arranged on the syringe through the needle body to be punctured; the syringe is arranged in the inner cylinder and fixedly connected to the inner cylinder; the inner cylinder is arranged in the outer cylinder through the first connecting piece, and the outer cylinder can be slidably connected to the inner cylinder; the outer cylinder is fixedly arranged on the fixing plate through the first connecting piece; the needle clamp cooperates with the needle withdrawal button to trigger the needle withdrawal button under the action of external force, so that the needle body to be punctured and the hose to be inserted are separated from the fixing device; after the needle body to be punctured is fixed on the fixing device, the relative spatial position between its initial position and the ultrasonic probe is determined and no longer changes. After the fixing device is calibrated and used for the control algorithm, puncture control based on ultrasonic image positioning can be realized; The guide device comprises three parts: an ultrasonic probe, a needle guide, and an outer cylinder guide, which is used to guide the fixing device and the needle to be punctured to move in a fixed direction based on ultrasound positioning of the blood vessel; the ultrasonic probe is used to collect data information of a preselected target blood vessel through ultrasound and send the data information to the control device; the needle guide is fixedly mounted on the ultrasonic probe, and the needle to be punctured is mounted on the needle guide; the outer cylinder guide is fixedly mounted on the fixing plate, and the outer cylinder is mounted on the outer cylinder guide; The ultrasonic probe is used in conjunction with an ultrasonic coupling sleeve, which is installed at the front end of the ultrasonic probe and is easily disassembled multiple times. The driving device is connected to the control device and is used to drive the fixing device to move along the direction guided by the guide device; the driving device is divided into a needle control part driving device and a tube placement part driving device; the needle control part driving device includes: a needle control motor and a second connecting piece; the needle control motor is fixedly arranged on the fixing plate and is fixedly connected to the inner cylinder through the second connecting piece; the tube placement part driving device includes: a tube placement motor, a third connecting piece, an outer cylinder connecting rod and a motor connecting rod; the tube placement motor is fixedly arranged on the fixing plate through the motor connecting rod, the tube placement motor is fixedly connected to the third connecting piece, and the third connecting piece is connected to the outer cylinder through the outer cylinder connecting rod.

5. A venous arterial puncture system, characterized in that: The automatic venous arterial puncture device according to claim 4 is used in conjunction with the control device, wherein: The automatic venous and arterial puncture device collects data information of a preselected target blood vessel through an ultrasonic probe and sends the data information of the preselected target blood vessel to the control device; The control device is used to receive data information of the preselected target blood vessel, calculate the position information of the target blood vessel based on the data information, and generate a control signal for controlling the operation of the automatic venous and arterial puncture device based on the position information; the automatic venous and arterial puncture device controls the needle body to be punctured and the soft tube to be inserted to puncture the skin based on the control signal; The driving device accurately controls the movement step lengths of the needle control motor and the tube placement motor; the control device converts the blood vessel information collected by the ultrasonic probe into the position information and puncture path of the target blood vessel based on algorithm identification and optimization; the control device compares the current coordinate information of the puncture point and the needle body to be punctured, calculates the displacement vector, and transmits it to the driving device in the form of a control signal; the driving device receives the control signal and moves to the puncture point according to the control signal.

Images