Visualizing arterial blood sampling and hemostatic compression integrated system

Abstract

The present application relates to blood sampling control technical field, specifically to a kind of visual arterial blood sampling and hemostasis pressing integrated system, system includes blood vessel size extraction module, puncture window limiting module, blood flow rate analysis module, blood sampling state determination module and hemostasis pressing adjustment module.In the present application, the blood vessel projection in the extraction imaging video is converted into real size, combined with the inner diameter size of the guide assembly, the effective puncture window area is limited by the space cross ratio and the center line alignment deviation, the blood flow color time sequence of sampling detection unit is analyzed, the abnormal blood flow area is identified according to the distribution difference of color dynamic change rate, and the blood sampling process stability state is determined according to the spatial distribution proportion of abnormal unit, combined with the skin deformation variable and deformation gradient under the action of multi-stage hemostasis loading force, the target control fine tuning interval is generated, the puncture positioning accuracy, blood sampling continuity and hemostasis pressing reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of blood collection control technology, and in particular to a visual integrated system for arterial blood collection and hemostasis pressure. Background Technology

[0002] The field of blood collection control technology involves technologies related to the regulation and management of blood sample acquisition and hemostasis processes during clinical testing and treatment. Its core aspects include puncture location, vessel identification, blood collection path guidance, blood collection flow control, and pressure hemostasis after blood collection. It typically requires selecting puncture points based on the structural characteristics of human arteries and extracting blood using needles and catheters. After blood collection, pressure is applied to the puncture site manually or using auxiliary devices to control bleeding. This technology spans multiple application scenarios, including emergency monitoring, critical care, and laboratory testing, and involves aspects such as optical observation, mechanical structure design, and standardized operating procedures.

[0003] The traditional integrated system for visual arterial blood collection and hemostasis involves configuring a transparent window or light source in the blood collection device to assist in observing the artery location, and combining it with a fixed structure to guide the blood collection needle. At the same time, a pressure plate or elastic pad is set on the outside of the device. After blood collection is completed, the pressure plate is moved above the puncture point and pressure is applied by manual or mechanical means. The specific methods usually include using the outer shell to fix the angle of the blood collection needle, observing the blood return through the transparent tubing, and after blood collection, the operator pushes the sliding part to drive the pressure block to press down on the skin surface, or using a spring structure to provide continuous pressure, thus completing the continuous operation of blood collection and hemostasis.

[0004] Traditional integrated systems often rely on observation windows and fixed guide structures in conjunction with manual judgment of puncture location and blood return status. During blood collection, it is difficult to simultaneously identify changes in blood vessel diameter and the adaptation relationship of the guide space, which can easily lead to deviations in puncture window selection and difficulty in timely detection of abnormal blood flow fluctuations. As a result, the stability of blood collection is greatly affected by the operator's experience. After blood collection, the pressure applied is mostly adjusted by feel, and there is a lack of continuous judgment between the continuous pressure state and skin deformation, which can easily lead to problems such as insufficient hemostasis or excessive local pressure. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides an integrated system for visual arterial blood sampling and hemostasis compression. The technical solution is as follows:

[0006] On the one hand, a visualized arterial blood sampling and hemostasis compression integrated system is provided, the system comprising: The blood vessel size extraction module acquires the imaging video sequence, extracts the lateral projection pixel values ​​of blood vessels in the imaging video sequence, constructs the projection scaling function and calculates the actual cross-sectional distance scale to generate the true lateral size of blood vessels. The puncture window definition module calculates the spatial cross-sectional ratio between the actual lateral dimension of the blood vessel and the inner diameter of the guide component based on the inner diameter of the guide component, and combines the spatial alignment deviation between the coordinates of the blood vessel centerline and the coordinates of the guide hole center axis to filter and generate an effective puncture window area. The blood flow rate analysis module, based on the effective puncture window area, divides the sampling and detection units, collects the blood flow color time series, calculates the color dynamic change rate and its rate distribution difference with the sorted reference median value, and filters to generate a set of abnormal blood flow regions. The blood collection status determination module traverses the set of abnormal blood flow regions, analyzes the spatial distribution ratio of abnormal units, reads the preset classification hyperplane weight vector and classification bias parameter, performs a product operation on the spatial distribution ratio of abnormal units and the classification hyperplane weight vector, and generates a stable identifier for the blood collection process. The hemostasis pressure adjustment module gradually increases the hemostasis loading force when the blood collection process is stable and meets the preset state parameters, collects the skin pressure deformation, constructs a smooth deformation sequence, and generates a target control fine-tuning range based on the pressure value corresponding to the gradient decrease and turning point of the deformation degree change.

[0007] As a further aspect of the present invention, the blood vessel size extraction module includes: The video pixel extraction submodule acquires the imaging video sequence, extracts the coordinates of continuously distributed blood vessel edge pixels within the imaging video sequence, locates the outer wall boundary points of the blood vessel based on the image grayscale gradient, calculates the difference in the horizontal axis coordinates of the blood vessel edge pixels at both ends, and generates the horizontal projection pixel values ​​of the blood vessel. The scaling function construction submodule collects the probe calibration distance parameters, reads the actual pixel conversion coefficient of the probe in the preset hardware based on the probe calibration distance parameters, performs scalar product calculation on the probe calibration distance parameters and the actual pixel conversion coefficient of the probe, extracts the spatial mapping slope, establishes a linear correspondence equation between pixels and distance scale based on the spatial mapping slope, and generates the projection scaling function. The size conversion calculation submodule calls the lateral projection pixel value of the blood vessel and the projection scaling function, substitutes the lateral projection pixel value of the blood vessel into the projection scaling function to perform constant term multiplication calculation, obtains the actual cross-sectional distance scale, and generates the true lateral size of the blood vessel.

[0008] As a further aspect of the present invention, the process of locating the outer wall boundary points of blood vessels based on image gray-level gradient specifically involves: extracting the image gray-level gradient from the imaging video sequence; calculating the absolute value of the pixel difference of the image gray-level gradient; extracting a preset edge detection operator; performing a two-dimensional convolution operation between the preset edge detection operator and the absolute value of the pixel difference to generate a gray-level jump response sequence; extracting a preset gray-level step threshold, which is set by statistically fitting the historical average gray-level change of the blood vessel boundary region in the sample image library; comparing the gray-level jump response sequence with the gray-level step threshold; filtering out pixels whose gray-level jump response sequence is greater than the gray-level step threshold to generate a candidate set of outer wall boundary points of blood vessels; and extracting the pixel coordinates of the outermost connected region in the candidate set of outer wall boundary points of blood vessels as the outer wall boundary points of blood vessels. The process of establishing a linear correspondence equation between pixels and distance scale based on the spatial mapping slope is as follows: extracting a preset reference pixel intercept constant, which is calibrated and set by the inherent imaging dead zone pixel value of the ultrasound probe in zero-distance contact state; using the spatial mapping slope as the coefficient of the first-order term of the linear correspondence equation; using the reference pixel intercept constant as the constant term of the linear correspondence equation; and constructing a linear correspondence equation with the lateral projection pixel value of the blood vessel as the independent variable and the actual physical distance as the dependent variable.

[0009] As a further aspect of the present invention, the puncture window limiting module includes: The proportional calculation submodule obtains the inner diameter of the guide component, divides the actual lateral dimension of the blood vessel by the inner diameter of the guide component to generate a spatial cross-sectional ratio, determines whether the spatial cross-sectional ratio is within the preset size matching ratio set, extracts the matching coordinates, and generates a size matching candidate region. The deviation extraction submodule collects the coordinates of the vessel centerline and the coordinates of the guide hole center axis of the size matching candidate region, calculates the horizontal Euclidean distance between the vessel centerline coordinates and the guide hole center axis coordinates, and generates the spatial alignment deviation. The window filtering submodule obtains the preset alignment deviation limit parameter, subtracts the spatial alignment deviation from the preset alignment deviation limit parameter, eliminates invalid puncture areas with a difference greater than zero, extracts the coordinates of areas with a difference not greater than zero, and generates an effective puncture window area.

[0010] As a further aspect of the present invention, the method for obtaining the size matching ratio set is as follows: obtaining the outer diameter of the blood collection needle tube and the elastic safety margin of the arterial wall, calculating the sum of the outer diameter of the blood collection needle tube and twice the elastic safety margin of the arterial wall, dividing the sum of the dimensions by the inner diameter of the guide component to generate the lower limit value of the matching ratio, extracting the maximum boundary width of the guide hole structure and dividing it by the inner diameter of the guide component to generate the upper limit value of the matching ratio, combining the lower limit value of the matching ratio and the upper limit value of the matching ratio to form a closed continuous numerical interval, and generating a preset size matching ratio set; The alignment deviation limit parameter is obtained by extracting the current actual transverse dimension of the blood vessel and the outer diameter of the needle tube of the assembled needle, calculating the cross-sectional difference between the actual transverse dimension of the blood vessel and the outer diameter of the needle tube, dividing the cross-sectional difference by a constant two, performing a 1 / 2 ratio operation to obtain the theoretical maximum offset margin on one side, reading the inherent mechanical vibration tolerance parameter of the puncture drive mechanism, subtracting the inherent mechanical vibration tolerance parameter from the theoretical maximum offset margin on one side, and generating the alignment deviation limit parameter.

[0011] As a further aspect of the present invention, the blood flow rate analysis module includes: The rate extraction submodule divides the effective puncture window area into multiple sampling and detection units, collects the blood flow color time series within the sampling and detection units, subtracts the color component values ​​of adjacent time nodes in the blood flow color time series and divides them by the time interval parameter to generate the color dynamic change rate. The median construction submodule calls the color dynamic change rate, sorts the color dynamic change rates in all sampling detection units in ascending order of numerical value, constructs a rate sequence, extracts the value at the middle position of the rate sequence, and generates a sorting reference median value. The anomaly screening submodule calculates the rate distribution difference by subtracting the sorting reference median value from the color dynamic change rate, compares the rate distribution difference with a preset difference limit parameter, extracts the coordinates of units where the rate distribution difference is greater than the preset difference limit parameter, and generates a set of abnormal blood flow regions.

[0012] As a further aspect of the present invention, the method for obtaining the difference limit parameter is as follows: obtaining a preset healthy blood flow sample dataset, extracting the baseline color change rate of the sample objects in the healthy blood flow sample dataset, calculating the numerical standard deviation of all baseline color change rates, multiplying the numerical standard deviation with a preset discrete distribution weighting coefficient, and generating a preset difference limit parameter.

[0013] As a further aspect of the present invention, the blood collection status determination module includes: The abnormal number statistics submodule traverses each data node within the abnormal blood flow region set, performs cumulative counting on each independent detection grid defined within the abnormal blood flow region set, extracts the cumulative count value, and generates the number of abnormal units. The spatial distribution calculation submodule calls the number of abnormal units to obtain the total number of sampling and detection units, performs a division operation with the number of abnormal units as the numerator and the total number of sampling and detection units as the denominator, extracts the quotient obtained from the division, and generates the spatial distribution ratio of abnormal units. The stability identification submodule reads the preset classification hyperplane weight vector and classification bias parameter according to the spatial distribution ratio of the abnormal units, performs a product operation on the spatial distribution ratio of the abnormal units and the classification hyperplane weight vector, adds the product result to the classification bias parameter, extracts the numerical sign term, and generates a stability identifier for the blood collection process.

[0014] As a further aspect of the present invention, the hemostatic pressure adjustment module includes: The data processing submodule, based on the stability indicator of the blood collection process, increases the hemostatic loading force when the stability indicator of the blood collection process meets the preset state parameters, collects the skin pressure deformation corresponding to the hemostatic loading force, reads the preset state transition matrix, performs product calculation on the skin pressure deformation and the preset state transition matrix, and generates a smooth deformation sequence. The gradient calculation submodule calls the smooth deformation sequence, subtracts the previous level parameter from the subsequent level smooth deformation sequence parameter to extract the deformation difference term, divides the deformation difference term by the pressure increment value of the two adjacent levels, and generates the deformation degree change gradient. The interval filtering submodule compares the values ​​of the deformation degree change gradients in ascending order with the values ​​of consecutive levels of deformation degree change gradients, filters the critical pressure parameters corresponding to the downward turning point of the deformation degree change gradients, extracts the pressure fluctuation limits, and generates the target control fine-tuning interval.

[0015] As a further aspect of the present invention, the process of screening the critical pressure parameter corresponding to the decrease in the deformation degree change gradient is specifically as follows: when the value of the deformation degree change gradient at the current level is less than the value of the deformation degree change gradient at the previous level, the hemostatic loading force corresponding to the current level is extracted as the critical pressure parameter.

[0016] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: This invention converts the lateral projection information of blood vessels in the imaging video to form a correspondence between the actual size of the blood vessels and the guiding space. Combined with the center position deviation, it screens the effective area for puncture, transforming the determination of the puncture entry point from empirical observation to a collaborative constraint of scale matching and spatial alignment. At the same time, it identifies abnormal flow distribution based on the characteristics of blood flow color changes over time, and completes the determination of blood collection stability based on the proportion of abnormal areas. After entering the hemostasis stage, it collects the pressure deformation and extracts the gradient turning position according to the progressive loading process, which can limit the pressure control to a range that better fits the tissue response, taking into account both the continuity of blood collection and the reliability of hemostasis. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of a visual arterial blood collection and hemostasis compression integrated system provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the system framework of the present invention; Figure 3 This is a flowchart of the blood vessel size extraction module in this invention; Figure 4 This is a flowchart of the puncture window limiting module in this invention; Figure 5 This is a flowchart of the blood flow rate analysis module in this invention; Figure 6 This is a flowchart of the blood collection status determination module in this invention; Figure 7 This is a flowchart of the hemostasis pressure adjustment module in this invention. Detailed Implementation

[0019] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0020] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0021] like Figure 1-2 As shown, this embodiment of the invention provides a visual integrated system for arterial blood collection and hemostasis, which includes a blood vessel size extraction module, a puncture window limitation module, a blood flow rate analysis module, a blood collection status determination module, and a hemostasis adjustment module. The blood vessel size extraction module acquires the imaging video sequence, extracts the horizontal projection pixel values ​​of blood vessels from the imaging video sequence, constructs a projection scaling function, inputs the horizontal projection pixel values ​​of blood vessels into the projection scaling function for calculation, and generates the true horizontal size of blood vessels. The puncture window limitation module obtains the inner diameter of the guide component, calculates the spatial cross-sectional ratio between the actual transverse dimension of the blood vessel and the inner diameter of the guide component, compares the spatial cross-sectional ratio with the preset size matching ratio set, calculates the spatial alignment deviation between the coordinates of the blood vessel centerline and the coordinates of the guide hole center axis, filters the area where the spatial alignment deviation is lower than the preset deviation limit and the spatial cross-sectional ratio is within the preset size matching ratio set, and generates an effective puncture window area. The blood flow rate analysis module divides the effective puncture window area into multiple sampling and detection units, collects the blood flow color time series within the sampling and detection units, calculates the color dynamic change rate of the blood flow color time series, constructs the sorting reference median value of the color dynamic change rate, calculates the rate distribution difference between the color dynamic change rate and the sorting reference median value, filters units whose rate distribution difference exceeds the preset difference limit parameter, and generates a set of abnormal blood flow regions. The blood collection status determination module counts the number of contained units in the abnormal blood flow region set, calculates the spatial distribution ratio of the number of contained units to the total number of sampling and detection units, and determines and generates a stable blood collection process identifier. The hemostasis pressure adjustment module determines that when the blood collection process stability indicator meets the preset stability conditions, it gradually increases the hemostasis loading force, collects the skin pressure deformation corresponding to the multi-level hemostasis loading force, calls the Kalman filter algorithm to calculate the skin pressure deformation, constructs a smooth deformation sequence, calculates the deformation degree change gradient of the smooth deformation sequence under adjacent pressure states, filters the pressure values ​​corresponding to the decrease and inflection of the deformation degree change gradient, and generates the target control fine-tuning range.

[0022] The true transverse dimensions of the blood vessel include the absolute lumen span, the wall expansion margin, and the effective cross-sectional width for blood flow. The effective puncture window area includes the safe needle insertion boundary, the target point projection coordinates, and the allowable offset tolerance. The abnormal blood flow region set includes eddy current accumulation patterns, flow velocity tomographic grids, and chromatic stagnation micro-regions. The stability indicators of the blood collection process include steady-state confidence probability, risk warning codes, and stage judgment scores. The target control fine-tuning range includes the lower limit of elastic compression, the upper limit of plastic deformation, and the pressure optimization step distance.

[0023] Specifically, such as Figure 2 , 3 As shown, the blood vessel size extraction module includes: The video pixel extraction submodule acquires the imaging video sequence, extracts the coordinates of continuously distributed blood vessel edge pixels within the imaging video sequence, locates the outer wall boundary points of the blood vessel based on the image grayscale gradient, calculates the difference in the horizontal axis coordinates of the blood vessel edge pixels at both ends, and generates the horizontal projection pixel values ​​of the blood vessel. The process of locating the outer wall boundary points of blood vessels based on image gray-level gradients is as follows: Extracting the image gray-level gradient from the imaging video sequence; calculating the absolute value of the pixel difference of the image gray-level gradient; extracting a preset edge detection operator; performing a two-dimensional convolution operation between the preset edge detection operator and the absolute value of the pixel difference to generate a gray-level jump response sequence; extracting a preset gray-level step threshold, which is set by statistically fitting the historical average gray-level changes of the blood vessel boundary region in the sample image library; comparing the gray-level jump response sequence with the gray-level step threshold; filtering pixels whose gray-level jump response sequence is greater than the gray-level step threshold to generate a candidate set of outer wall boundary points of blood vessels; and extracting the pixel coordinates of the outermost connected region in the candidate set of outer wall boundary points of blood vessels as the outer wall boundary points of blood vessels. The image acquisition card connected to the medical ultrasound probe reads a 1920x1080 resolution video sequence. An internal frame buffer is used to break down the video sequence frame by frame at a fixed frequency of 30 frames per second, converting the pixel values ​​of each frame into single-channel grayscale data ranging from 0 to 255. An internal filter performs mean filtering on the extracted frame images. The mean filter convolution kernel size is set to 3x3, traversing all pixels within the frame image, replacing the center pixel with the average grayscale value of its eight neighboring pixels, thus removing high-frequency random noise from the image acquisition process. The grayscale gradient of the filtered image is calculated; the absolute value of the pixel difference equals the current pixel's grayscale value minus the absolute value of the grayscale values ​​of its horizontally adjacent pixels. The advantage of this operation is that by directly calculating the absolute value of the grayscale difference between adjacent pixels instead of complex surface fitting calculations, the extraction delay of edge features is significantly reduced. A preset edge detection operator is retrieved from memory. This preset edge detection operator is a horizontal gradient operator, also 3x3 in size, containing matrix parameters with a center weight of 2 and edge weights of 1 and -1. A 2D convolution operation is performed between the preset edge detection operator and the absolute value of pixel differences. The gray-level jump response sequence is equal to the sum of the products of the corresponding elements of the preset edge detection operator and the absolute value of pixel differences. For example, if the absolute value of pixel differences in the current pixel's neighborhood has all elements of 120, substituting them into the horizontal gradient operator for convolution results in a gray-level jump response sequence of 0. For the gray-level jump response sequence, a preset gray-level step threshold is read. This threshold is obtained by retrieving 5000 sample images from a hospital ultrasound image database containing labeled vascular boundary regions, extracting the historical average gray-level changes for these regions, summing these 5000 historical average gray-level changes, and dividing by 5000. The gray-level step threshold is then set to 145. The gray-level jump response sequence is compared with the gray-level step threshold. When the gray-level jump response sequence equals 160 and the gray-level step threshold equals 145, the condition is met, and the pixel coordinates are added to the candidate set of vessel outer wall boundary points. Connectivity analysis is performed on the candidate set of vessel outer wall boundary points to calculate the centroid x-axis coordinates of each connected region. The pixel coordinates of the two outermost connected regions with the largest and smallest x-axis coordinate values ​​are extracted as vessel outer wall boundary points. The lateral projection pixel value of the vessel is equal to the x-axis coordinate value of the right vessel outer wall boundary point minus the x-axis coordinate value of the left vessel outer wall boundary point. When the left coordinate is 200 and the right coordinate is 350, the lateral projection pixel value of the vessel is 150.

[0024] The scaling function construction submodule collects the probe calibration distance parameters, reads the actual pixel conversion coefficient of the probe in the preset hardware based on the probe calibration distance parameters, performs scalar product calculation on the probe calibration distance parameters and the actual pixel conversion coefficient of the probe, extracts the spatial mapping slope, establishes a linear correspondence equation between pixels and distance scale based on the spatial mapping slope, and generates the projection scaling function. The process of establishing a linear correspondence equation between pixels and distance scale based on the spatial mapping slope is as follows: extract a preset reference pixel intercept constant. The reference pixel intercept constant is calibrated and set by the pixel value of the inherent imaging dead zone of the ultrasound probe in the zero-distance contact state. The spatial mapping slope is used as the coefficient of the first-order term of the linear correspondence equation. The reference pixel intercept constant is used as the constant term of the linear correspondence equation to construct a linear correspondence equation with the lateral projection pixel value of the blood vessel as the independent variable and the actual physical distance as the dependent variable. The probe calibration distance parameter, set to 40 mm, is read from the probe's factory calibration file via serial communication. Simultaneously, the actual pixel conversion coefficient is read from the preset internal read-only memory, set to 25 pixels per millimeter. The spatial mapping slope is calculated by multiplying the calibration distance parameter and the actual pixel conversion coefficient by a scalar. Substituting these values, the spatial mapping slope equals 1000. A preset reference pixel intercept constant is read. This constant is calibrated by reading the inherent imaging dead zone pixel values ​​of the ultrasound probe in zero-distance contact mode. Specifically, the probe is placed against a standard sound-absorbing rubber block, and the average pixel value of the black dead zone height in the first 50 frames is read. The reference pixel intercept constant is set to 15. The spatial mapping slope is used as the coefficient of the first-order term in the linear equation, and the reference pixel intercept constant is used as the constant term. A linear equation is constructed with the lateral projection pixel values ​​of the blood vessel as the independent variable and the actual physical distance as the dependent variable. The advantage of this operational logic is that by introducing dead zone pixels in the 0-distance state as intercept compensation, it eliminates the systematic deviation in size measurement caused by near-field blind zones.

[0025] Table 1 Probe Calibration and Mapping Parameters Probe calibration distance parameters actual pixel conversion coefficient of the probe Spatial mapping slope Reference pixel intercept constant 40 25 1000 15 As shown in Table 1, the mapping relationship is constructed based on the parameters in the table, and the calculated projection scaling function is directly used for subsequent scale conversion.

[0026] The size conversion calculation submodule calls the horizontal projection pixel value of the blood vessel and the projection scaling function. It substitutes the horizontal projection pixel value of the blood vessel into the projection scaling function to perform constant term multiplication calculation, obtains the actual cross-sectional distance scale, and generates the true horizontal size of the blood vessel.

[0027] The system receives the lateral projection pixel values ​​of the blood vessel via the bus interface and simultaneously calls the projection scaling function stored in memory. Substituting these lateral projection pixel values ​​into the projection scaling function, a constant term multiplication is performed. The actual cross-sectional distance scale is equal to the lateral projection pixel value of the blood vessel minus the reference pixel intercept constant, divided by the spatial mapping slope. For example, when the lateral projection pixel value of the blood vessel is 150, the reference pixel intercept constant is 15, and the spatial mapping slope is 1000, the actual cross-sectional distance scale is equal to 150 minus 15 divided by 1000, resulting in an actual cross-sectional distance scale of 0.135 mm. Multiplying the actual cross-sectional distance scale by the probe's field of view magnification (set to 10), the actual lateral dimension of the blood vessel is calculated as 0.135 mm multiplied by 10, resulting in an actual lateral dimension of 1.35 mm.

[0028] Specifically, such as Figure 2 , 4 As shown, the puncture window limiting module includes: The proportion calculation submodule obtains the inner diameter of the guide component, divides the actual lateral dimension of the blood vessel by the inner diameter of the guide component to generate the spatial cross-sectional proportion, determines whether the spatial cross-sectional proportion is within the preset size matching proportion set, extracts the matching coordinates, and generates a size matching candidate region. The method for obtaining the size matching ratio set is as follows: obtain the outer diameter of the blood collection needle tube and the elastic safety margin of the arterial wall, calculate the sum of the outer diameter of the blood collection needle tube and twice the elastic safety margin of the arterial wall, divide the sum of the dimensions by the inner diameter of the guide component to generate the lower limit value of the matching ratio, extract the maximum boundary width of the guide hole structure and divide it by the inner diameter of the guide component to generate the upper limit value of the matching ratio, combine the lower limit value of the matching ratio and the upper limit value of the matching ratio to form a closed continuous numerical interval, and generate the preset size matching ratio set. Access the medical device database to retrieve the specifications of the currently assembled guide assembly, obtaining its inner diameter, which is set to 2.0 mm. Divide the actual transverse dimension of the blood vessel (1.35 mm) by the guide assembly's inner diameter (2.0 mm), resulting in a spatial cross-sectional ratio of 0.675 (1.35 divided by 2.0). Read the size matching ratio set, obtained by accessing the blood collection needle configuration table, to retrieve the blood collection needle's outer diameter and the arterial wall's elastic safety margin. The outer diameter is set to 0.8 mm, and the arterial wall's elastic safety margin is set to 0.15 mm. Calculate the sum of the blood collection needle's outer diameter and twice the arterial wall's elastic safety margin. This sum equals 0.8 plus 2 multiplied by 0.15, resulting in a total size of 1.1 mm. Divide this sum by the guide assembly's inner diameter, with a matching ratio lower limit of 0.55 (1.1 divided by 2.0). The maximum boundary width of the guide hole structure is extracted and set to 1.8 mm. Dividing this maximum boundary width by the inner diameter of the guide component, the upper limit of the matching ratio is calculated as 1.8 divided by 2.0, resulting in an upper limit value of 0.9. Combining the lower limit value of the matching ratio (0.55) with the upper limit value of 0.9 forms a closed continuous numerical range, generating a preset size matching ratio set of 0.55 to 0.9. Comparing the spatial cross-sectional ratio (0.675) with the preset size matching ratio set, a match is determined if 0.675 is greater than 0.55 and less than 0.9. The matching coordinates in the current video frame are extracted as 300 on the horizontal axis and 400 on the vertical axis, generating a size matching candidate region. The advantage of this calculation logic is that by comprehensively considering the needle outer diameter and elastic margin to define the ratio range, unsuitable puncture areas with excessively small or large diameters are eliminated, improving the reliability of visual detection and matching in complex tissue environments.

[0029] The deviation extraction submodule collects the coordinates of the vessel centerline and the guide hole center axis of the size matching candidate region, calculates the horizontal Euclidean distance between the vessel centerline coordinates and the guide hole center axis coordinates, and generates the spatial alignment deviation. To achieve precise visual inspection and spatial positioning before puncture and assembly, the centroid of the grayscale distribution within the generated size matching candidate region is acquired and defined as the vessel centerline coordinates, with the horizontal axis at 300 and the vertical axis at 400. Simultaneously, by reading the position feedback from the motor encoder of the guide mechanism, the coordinates of the guide hole's central axis are calculated, with the horizontal axis at 305 and the vertical axis at 400. The horizontal Euclidean distance between the vessel centerline coordinates and the guide hole's central axis coordinates is calculated; the spatial alignment deviation is equal to the absolute value of the difference between the horizontal axis of the vessel centerline coordinates and the horizontal axis of the guide hole's central axis coordinates. Substituting the aforementioned horizontal axes 300 and 305 into the calculation, the spatial alignment deviation is calculated to be 5 pixel units. Multiplying each pixel unit by the reciprocal of the probe's actual pixel conversion coefficient (each pixel corresponds to 0.04 mm), the final spatial alignment deviation is 0.2 mm. This quantitative result based on visual inspection technology effectively reflects the relative offset between the physical drive mechanism and the target anatomical structure.

[0030] The window filtering submodule obtains the preset alignment deviation limit parameter, subtracts the spatial alignment deviation from the preset alignment deviation limit parameter, removes invalid puncture areas with a difference greater than zero, extracts the coordinates of areas with a difference not greater than zero, and generates an effective puncture window area. The alignment deviation limit parameter is obtained by extracting the current actual transverse dimension of the blood vessel and the outer diameter of the needle tube of the assembled needle, calculating the cross-sectional difference between the actual transverse dimension of the blood vessel and the outer diameter of the needle tube, dividing the cross-sectional difference by a constant two, performing a 1 / 2 ratio operation to obtain the theoretical maximum offset margin on one side, reading the inherent mechanical vibration tolerance parameter of the puncture drive mechanism, subtracting the inherent mechanical vibration tolerance parameter from the theoretical maximum offset margin on one side to perform numerical deduction, and generating the alignment deviation limit parameter. The preset alignment deviation limit parameter is obtained by extracting the current actual lateral dimension of the blood vessel (1.35 mm) and the outer diameter of the needle tube (set to 0.8 mm). The cross-sectional difference between the actual lateral dimension of the blood vessel and the outer diameter of the needle tube is calculated. This difference equals 1.35 minus 0.8, resulting in a cross-sectional difference of 0.55 mm. Dividing this difference by a constant 2, the theoretical maximum offset margin on one side is calculated as 0.55 divided by 2, resulting in a theoretical maximum offset margin on one side of 0.275 mm. The inherent mechanical vibration tolerance parameter in the puncture drive mechanism control board is read. This parameter is set to 0.05 mm. Subtracting this parameter from the theoretical maximum offset margin on one side, the alignment deviation limit parameter equals 0.275 minus 0.05, resulting in a calculation of 0.225 mm. The advantage of this calculation logic is that by subtracting the mechanical vibration tolerance, unstable puncture points in a critical state are filtered out. Subtracting the aforementioned spatial alignment deviation of 0.2 mm from the preset alignment deviation limit parameter of 0.225 mm, the difference is equal to 0.2 minus 0.225, resulting in a difference of -0.025 mm. Since the condition of a difference less than 0 and not greater than 0 is met, the center coordinates and boundary range of this area are retained, marked, and an effective puncture window area is generated.

[0031] Specifically, such as Figure 2 , 5 As shown, the blood flow rate analysis module includes: The rate extraction submodule divides the effective puncture window area into multiple sampling and detection units, collects the blood flow color time series within the sampling and detection units, subtracts the color component values ​​of adjacent time nodes in the blood flow color time series and divides them by the time interval parameter to generate the color dynamic change rate. Based on the generated effective puncture window area, a dynamic visual detection mechanism for Doppler blood flow imaging is introduced. This area is divided into 25 sampling detection units, each 20 pixels by 20 pixels. Color Doppler blood flow time series within each sampling detection unit are acquired, and the red channel component values ​​of blood flow color at adjacent time points (frame 1 and frame 2) are extracted. The time interval parameter is set to 0.033 seconds according to the frame rate. The color component value of frame 2 is subtracted from the color component value of frame 1 and then divided by the time interval parameter. For example, if the color component value of frame 1 is 150 and the color component value of frame 2 is 183, the color dynamic change rate is equal to 183 minus 150 divided by 0.033, resulting in a color dynamic change rate of 1000 color values ​​per second.

[0032] The median construction submodule calls the color dynamic change rate, sorts the color dynamic change rates in all sampling detection units in ascending order of numerical value, constructs a rate sequence, extracts the value at the middle position of the rate sequence, and generates the sorting reference median value. The system retrieves the color dynamic change rates from the 25 generated sampling detection units and stores these 25 values ​​in a floating-point array. A quicksort algorithm is then run, selecting the first element of the array as the pivot value. Elements smaller than the pivot are placed in the left interval, and elements larger than the pivot are placed in the right interval. This process is recursively repeated to sort the color dynamic change rates in all sampling detection units in ascending order, constructing a rate sequence of length 25. The value at the middle position of the rate sequence, i.e., the value with index 13, is extracted and set to 950 color units per second. This value is then used as the median value for sorting. The advantage of this operational logic is that by extracting the median value instead of the arithmetic mean, it avoids interference from a few isolated local turbulence anomalies in the overall blood flow change assessment.

[0033] The anomaly screening submodule calculates the rate distribution difference by subtracting the sorting reference median value from the rate of color dynamic change, compares the rate distribution difference with a preset difference limit parameter, extracts the coordinates of cells whose rate distribution difference is greater than the preset difference limit parameter, and generates a set of abnormal blood flow regions. The specific method for obtaining the difference limit parameter is as follows: obtain a preset healthy blood flow sample dataset, extract the baseline color change rate of the sample objects in the healthy blood flow sample dataset, calculate the numerical standard deviation of all baseline color change rates, multiply the numerical standard deviation with the preset discrete distribution weighting coefficient, and generate the preset difference limit parameter.

[0034] Based on the color dynamic change rate and the median value of the sorting reference, the rate distribution difference is calculated. The rate distribution difference is equal to the absolute value of the color dynamic change rate minus the absolute value of the median value. Substituting the aforementioned color dynamic change rate of 1000 and the median value of 950, the rate distribution difference is calculated to be 50, which is equal to the absolute value of 1000 minus 950. The rate distribution difference is compared with a preset difference limit parameter. The difference limit parameter is obtained by retrieving an externally mounted healthy blood flow sample dataset, which comes from continuous Doppler acquisition records of 10,000 normal healthy individuals. 10,000 baseline color change rates of the sample objects in the dataset are extracted, and the arithmetic mean of these baseline color change rates is calculated to be 900 color value units per second. Then, the sum of squares of the differences between all baseline color change rates and the mean is calculated. The sum of squares is divided by 10,000 and then the square root is taken, resulting in a standard deviation of 25. The numerical standard deviation is multiplied by the preset discrete distribution weighting coefficient, which is set to 1.5. The preset difference limit parameter is equal to 25 multiplied by 1.5, and the preset difference limit parameter is calculated to be 37.5.

[0035] Table 2. Characteristics and Difference Limits of Healthy Blood Flow Average rate of change of baseline color Numerical standard deviation Discrete distribution weighting coefficients Preset difference limit parameters 900 25 1.5 37.5 As shown in Table 2, the judgment criteria are derived according to the parameters in this table. The velocity distribution difference 50 is compared with the preset difference limit parameter 37.5. If the condition 50 is greater than 37.5, the coordinate information of the sampling detection unit is extracted, with the horizontal axis at 300 and the vertical axis at 400, and it is merged into the abnormal blood flow region set.

[0036] Specifically, such as Figure 2 , 6 As shown, the blood collection status determination module includes: The abnormal number statistics submodule traverses each data node within the abnormal blood flow region set, performs cumulative counting on each of the independent detection grids delineated within the abnormal blood flow region set, extracts the cumulative count value, and generates the number of abnormal units. Access the set of abnormal blood flow regions stored in memory, extract the length identifier of the set, initialize an integer counter parameter to 0, traverse each data node within the abnormal blood flow region set, and perform a cumulative count operation on each of the independently delineated detection grids within the abnormal blood flow region set. When the current node coordinates are valid and not repeatedly marked, the counter parameter is equal to the original counter parameter plus 1. Assuming there are a total of 5 nodes satisfying the condition in the abnormal blood flow region set, perform the addition operation 5 times, extract the final accumulated count value, and generate an abnormal cell containing 5 nodes.

[0037] The spatial distribution calculation submodule calls the number of abnormal units, obtains the total number of sampled and detected units, performs a division operation with the number of abnormal units as the numerator and the total number of sampled and detected units as the denominator, extracts the quotient obtained from the division, and generates the spatial distribution ratio of abnormal units. The number of abnormal units is 5. The total number of sampling and detection units is 25. A division operation is performed with the number of abnormal units as the numerator and the total number of sampling and detection units as the denominator. The spatial distribution ratio of abnormal units equals the number of abnormal units divided by the total number of sampling and detection units. Substituting 5 and 25 into the calculation, the spatial distribution ratio of abnormal units equals 5 divided by 25, resulting in a spatial distribution ratio of 0.2, meaning that one-fifth of the area exhibits abnormal blood flow velocity.

[0038] The stability identification submodule reads the preset classification hyperplane weight vector and classification bias parameter based on the spatial distribution ratio of abnormal units, performs a product operation on the spatial distribution ratio of abnormal units and the classification hyperplane weight vector, adds the product result to the classification bias parameter, extracts the numerical sign term, and generates a stability identifier for the blood collection process. The process of generating a stable blood collection identifier involves the following steps: Iteratively solving a pre-set sample dataset to extract orthogonal normal vectors of the data boundaries, configuring these vectors as classification hyperplane weight vectors; calculating the average intercept of edge feature points to the data boundaries, setting this average intercept as a classification bias parameter; performing a dot product operation between the spatial distribution ratio of abnormal units and the classification hyperplane weight vectors to obtain a scalar product; adding the scalar product to the classification bias parameter to generate a linear discriminant value; extracting a positive sign and assigning a numerical sign to a term with a linear discriminant value greater than zero to generate a stable blood collection identifier; and extracting a negative sign and assigning a numerical sign to a term with a linear discriminant value not greater than zero to generate a fluctuating blood collection identifier. The pre-set classification hyperplane weight vector and classification bias parameters are read. Obtaining these parameters relies on iterative solving of the support vector machine model. A sample dataset containing 500 historical blood sampling fluctuation features and corresponding labels is constructed. The Lagrange multiplier method is used to iteratively solve the sample dataset. In each iteration, the distance from each sample point to the segmentation plane is calculated, and the weights are continuously updated to maximize the positive and negative sample classification margin until the loss function converges. The orthogonal normal vectors of the data boundaries are extracted and configured as follows: The classification hyperplane weight vector is set to -5.0. The average intercept of edge feature points to the data boundary is calculated and set as the classification bias parameter, with a value of 1.5. The outlier spatial distribution ratio of 0.2 is multiplied by the classification hyperplane weight vector of -5.0, resulting in a scalar product of 0.2 and -5.0, which is -1.0. This scalar product is then added to the classification bias parameter, and the linear discriminant value is equal to -1.0 plus 1.5, resulting in a linear discriminant value of 0.5. Since 0.5 is greater than 0, the condition is met, and a positive sign is extracted and assigned to the numerical sign term to generate a stable blood sampling identifier. If the linear discriminant value is not greater than 0, a fluctuating blood sampling identifier is generated. The advantage of this operation logic is that by introducing the boundary intercept bias, the classification sensitivity is balanced under different blood vessel thickness conditions.

[0039] Specifically, such as Figure 2 , 7 As shown, the hemostasis pressure adjustment module includes: The data processing submodule, based on the stability indicator of the blood collection process, adds a hemostatic loading force when the stability indicator of the blood collection process meets the preset state parameters, collects the skin pressure deformation corresponding to the hemostatic loading force, reads the preset state transition matrix, performs product calculation on the skin pressure deformation and the preset state transition matrix, and generates a smooth deformation sequence. The process of multiplying the skin pressure deformation with a preset state transition matrix is ​​as follows: First, obtain the preset state prediction covariance parameters and noise reference sequence; second, construct an initialization transition operator for the discrete state space dimension based on the state prediction covariance parameters and noise reference sequence, and set the initialization transition operator to the preset state transition matrix; third, arrange the skin pressure deformation to form a one-dimensional observation column vector; fourth, perform a dot product operation between the one-dimensional observation column vector and the preset state transition matrix, extract the main diagonal elements from the dot product result to form the smoothing filter value for the current level; fifth, concatenate the smoothing filter values ​​from multiple levels according to the acquisition order to generate a smooth deformation sequence. Upon receiving a stable blood collection indicator, and determining that the indicator meets preset state parameters (i.e., the system is in a safe operating state), a control pulse is sent to the force-controlled motor to increase the hemostatic loading force. The hemostatic loading force is set to 2.0 Newtons. A thin-film pressure sensor array connected to the skin contact surface collects the skin pressure deformation corresponding to the hemostatic loading force; the currently collected skin pressure deformation is set to 0.5 mm. A preset state transition matrix is ​​read. The matrix is ​​generated by obtaining a preset state prediction covariance parameter of 0.1 and a noise reference sequence of 0.05. Based on the state prediction covariance parameter and the noise reference sequence, an initialization transition operator for the discrete state space dimension is constructed. The main diagonal element of the transition operator is equal to 1 minus the ratio of the state prediction covariance parameter to the noise reference sequence, i.e., 1 minus 0.1 divided by 0.05. This process... The logic is not a simple addition or subtraction but involves a simplified one-dimensional state transition setting of Kalman filtering. Since the calculated negative value of the main diagonal does not conform to physical reality, the state prediction covariance parameter is adjusted to 0.01. The main diagonal element is equal to 1 minus 0.01 divided by 0.05, and the calculated main diagonal element is 0.8, with the remaining elements being 0. This two-dimensional diagonal matrix is ​​set as the preset state transition matrix. The skin compression deformation of 0.5 is arranged to form a one-dimensional observation column vector. The one-dimensional observation column vector is multiplied by the preset state transition matrix. The result of the dot product is equal to 0.5 multiplied by 0.8, which is 0.4. The result of the dot product is extracted to form the smoothing filter value of the current level, which is 0.4 mm. The smoothing filter values ​​of multiple levels are concatenated and stored in contiguous memory addresses in chronological order to generate a smooth deformation sequence of length 50.

[0040] The gradient calculation submodule calls the smooth deformation sequence, subtracts the previous level parameter from the subsequent level smooth deformation sequence parameter to extract the deformation difference term, divides the deformation difference term by the pressure increment value of the two adjacent levels, and generates the gradient of deformation degree change. The system retrieves the smooth deformation sequence and corresponding force value record array from memory. The parameter of the next smooth deformation sequence is set to 0.45 mm, and the parameter of the previous sequence is set to 0.4 mm. The deformation difference term equals the parameter of the next smooth deformation sequence minus the parameter of the previous sequence. Substituting 0.45 and 0.4 into the calculation, the deformation difference term equals 0.45 minus 0.4, resulting in a deformation difference term of 0.05 mm. The corresponding loading force of the next stage (2.2 N) and the corresponding loading force of the previous stage (2.0 N) are extracted. The pressure increment between adjacent stages equals 2.2 minus 2.0, resulting in a pressure increment of 0.2 N. Dividing the deformation difference term by the pressure increment between adjacent stages, the deformation gradient equals 0.05 divided by 0.2, resulting in a deformation gradient of 0.25 mm per N.

[0041] Table 3. Record of Skin Pressure and Smooth Deformation Hemostatic loading force Skin deformation under pressure Smoothing filter values Deformation gradient 2.0 0.5 0.4 0.25 2.2 0.55 0.45 0.25 As shown in Table 3, by sequentially calling the data at each level in the table and continuously performing differential operations, a continuous gradient sequence of deformation degree changes is generated.

[0042] The interval filtering submodule compares the values ​​of the deformation degree change gradients in ascending order based on the deformation degree change gradient, filters the critical pressure parameters corresponding to the turning point of the deformation degree change gradient, extracts the pressure fluctuation limit, and generates the target control fine-tuning interval. The process of selecting the critical pressure parameter corresponding to the decrease in the deformation degree gradient is as follows: when the value of the deformation degree gradient at the current level is less than the value of the deformation degree gradient at the previous level, the hemostatic loading force corresponding to the current level is extracted as the critical pressure parameter.

[0043] The algorithm extracts a sequence of deformation gradient arrays, sets a loop variable, and compares the values ​​of deformation gradients at consecutive levels in ascending order. The current level's deformation gradient value is set to 0.15 mm / Newton, and the previous level's value is set to 0.25 mm / Newton. A judgment is made: if the current level's deformation gradient value is less than the previous level's, the corresponding hemostatic loading force of 2.4 Newtons is extracted as the critical pressure parameter. The advantage of this logic is that by monitoring the gradient descent inflection point, the physical inflection point of the transition from elastic deformation to plastic compression in the skin and subcutaneous tissue is located. The critical pressure parameter of 2.4 Newtons is added to a pressure fluctuation limit set to 0.2 Newtons, resulting in an upper limit of 2.6 Newtons. Subtracting the pressure fluctuation limit from the critical pressure parameter yields a lower limit of 2.2 Newtons. By integrating the upper and lower limits, a target control fine-tuning range of 2.2 Newtons to 2.6 Newtons is generated. This target control fine-tuning range is then sent to the underlying motor driver to limit the range of contact force fluctuations during the subsequent puncture and needle insertion stages. This ensures smooth puncture under the dual safety closed-loop protection of visual detection guidance and underlying force feedback.

[0044] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of protection of the technical solution.

Claims

1. A visual integrated system for arterial blood sampling and hemostasis compression, characterized in that, The system includes: The blood vessel size extraction module acquires the imaging video sequence, extracts the lateral projection pixel values ​​of blood vessels in the imaging video sequence, constructs the projection scaling function and calculates the actual cross-sectional distance scale to generate the true lateral size of blood vessels. The puncture window definition module calculates the spatial cross-sectional ratio between the actual lateral dimension of the blood vessel and the inner diameter of the guide component based on the inner diameter of the guide component, and combines the spatial alignment deviation between the coordinates of the blood vessel centerline and the coordinates of the guide hole center axis to filter and generate an effective puncture window area. The blood flow rate analysis module, based on the effective puncture window area, divides the sampling and detection units, collects the blood flow color time series, calculates the color dynamic change rate and its rate distribution difference with the sorted reference median value, and filters to generate a set of abnormal blood flow regions. The blood collection status determination module traverses the set of abnormal blood flow regions, analyzes the spatial distribution ratio of abnormal units, reads the preset classification hyperplane weight vector and classification bias parameter, performs a product operation on the spatial distribution ratio of abnormal units and the classification hyperplane weight vector, and generates a stable identifier for the blood collection process. The hemostasis pressure adjustment module gradually increases the hemostasis loading force when the blood collection process is stable and meets the preset state parameters, collects the skin pressure deformation, constructs a smooth deformation sequence, and generates a target control fine-tuning range based on the pressure value corresponding to the gradient decrease and turning point of the deformation degree change.

2. The integrated system for visual arterial blood collection and hemostasis compression according to claim 1, characterized in that, The blood vessel size extraction module includes: The video pixel extraction submodule acquires the imaging video sequence, extracts the coordinates of continuously distributed blood vessel edge pixels within the imaging video sequence, locates the outer wall boundary points of the blood vessel based on the image grayscale gradient, calculates the difference in the horizontal axis coordinates of the blood vessel edge pixels at both ends, and generates the horizontal projection pixel values ​​of the blood vessel. The scaling function construction submodule collects the probe calibration distance parameters, reads the actual pixel conversion coefficient of the probe in the preset hardware based on the probe calibration distance parameters, performs scalar product calculation on the probe calibration distance parameters and the actual pixel conversion coefficient of the probe, extracts the spatial mapping slope, establishes a linear correspondence equation between pixels and distance scale based on the spatial mapping slope, and generates the projection scaling function. The size conversion calculation submodule calls the lateral projection pixel value of the blood vessel and the projection scaling function, substitutes the lateral projection pixel value of the blood vessel into the projection scaling function to perform constant term multiplication calculation, obtains the actual cross-sectional distance scale, and generates the true lateral size of the blood vessel.

3. The integrated system for visual arterial blood collection and hemostasis compression according to claim 2, characterized in that, The process of locating the outer wall boundary points of blood vessels based on image gray-level gradients specifically involves: extracting the image gray-level gradient from the imaging video sequence; calculating the absolute value of the pixel difference of the image gray-level gradient; extracting a preset edge detection operator; performing a two-dimensional convolution operation between the preset edge detection operator and the absolute value of the pixel difference to generate a gray-level jump response sequence; extracting a preset gray-level step threshold, which is set by statistically fitting the historical average gray-level change of the blood vessel boundary region in the sample image library; comparing the gray-level jump response sequence with the gray-level step threshold; filtering out pixels whose gray-level jump response sequence is greater than the gray-level step threshold to generate a candidate set of outer wall boundary points of blood vessels; and extracting the pixel coordinates of the outermost connected region in the candidate set of outer wall boundary points of blood vessels as the outer wall boundary points of blood vessels. The process of establishing a linear correspondence equation between pixels and distance scale based on the spatial mapping slope is as follows: extracting a preset reference pixel intercept constant, which is calibrated and set by the inherent imaging dead zone pixel value of the ultrasound probe in zero-distance contact state; using the spatial mapping slope as the coefficient of the first-order term of the linear correspondence equation; using the reference pixel intercept constant as the constant term of the linear correspondence equation; and constructing a linear correspondence equation with the lateral projection pixel value of the blood vessel as the independent variable and the actual physical distance as the dependent variable.

4. The integrated system for visual arterial blood collection and hemostasis compression according to claim 1, characterized in that, The puncture window limiting module includes: The proportional calculation submodule obtains the inner diameter of the guide component, divides the actual lateral dimension of the blood vessel by the inner diameter of the guide component to generate a spatial cross-sectional ratio, determines whether the spatial cross-sectional ratio is within the preset size matching ratio set, extracts the matching coordinates, and generates a size matching candidate region. The deviation extraction submodule collects the coordinates of the vessel centerline and the coordinates of the guide hole center axis of the size matching candidate region, calculates the horizontal Euclidean distance between the vessel centerline coordinates and the guide hole center axis coordinates, and generates the spatial alignment deviation. The window filtering submodule obtains the preset alignment deviation limit parameter, subtracts the spatial alignment deviation from the preset alignment deviation limit parameter, eliminates invalid puncture areas with a difference greater than zero, extracts the coordinates of areas with a difference not greater than zero, and generates an effective puncture window area.

5. The integrated system for visual arterial blood collection and hemostasis compression according to claim 4, characterized in that, The method for obtaining the size matching ratio set is as follows: obtain the outer diameter of the blood collection needle tube and the elastic safety margin of the arterial wall, calculate the sum of the outer diameter of the blood collection needle tube and twice the elastic safety margin of the arterial wall, divide the sum of the dimensions by the inner diameter of the guide component to generate the lower limit value of the matching ratio, extract the maximum boundary width of the guide hole structure and divide it by the inner diameter of the guide component to generate the upper limit value of the matching ratio, combine the lower limit value of the matching ratio and the upper limit value of the matching ratio to form a closed continuous numerical interval, and generate the preset size matching ratio set. The alignment deviation limit parameter is obtained by extracting the current actual transverse dimension of the blood vessel and the outer diameter of the needle tube of the assembled needle, calculating the cross-sectional difference between the actual transverse dimension of the blood vessel and the outer diameter of the needle tube, dividing the cross-sectional difference by a constant two, performing a 1 / 2 ratio operation to obtain the theoretical maximum offset margin on one side, reading the inherent mechanical vibration tolerance parameter of the puncture drive mechanism, subtracting the inherent mechanical vibration tolerance parameter from the theoretical maximum offset margin on one side, and generating the alignment deviation limit parameter.

6. The integrated system for visual arterial blood collection and hemostasis compression according to claim 1, characterized in that, The blood flow rate analysis module includes: The rate extraction submodule divides the effective puncture window area into multiple sampling and detection units, collects the blood flow color time series within the sampling and detection units, subtracts the color component values ​​of adjacent time nodes in the blood flow color time series and divides them by the time interval parameter to generate the color dynamic change rate. The median construction submodule calls the color dynamic change rate, sorts the color dynamic change rates in all sampling detection units in ascending order of numerical value, constructs a rate sequence, extracts the value at the middle position of the rate sequence, and generates a sorting reference median value. The anomaly screening submodule calculates the rate distribution difference by subtracting the sorting reference median value from the color dynamic change rate, compares the rate distribution difference with a preset difference limit parameter, extracts the coordinates of units where the rate distribution difference is greater than the preset difference limit parameter, and generates a set of abnormal blood flow regions.

7. The integrated system for visual arterial blood collection and hemostasis compression according to claim 6, characterized in that, The method for obtaining the difference limit parameter is as follows: obtain a preset healthy blood flow sample dataset, extract the baseline color change rate of the sample objects in the healthy blood flow sample dataset, calculate the numerical standard deviation of all baseline color change rates, multiply the numerical standard deviation with the preset discrete distribution weighting coefficient, and generate the preset difference limit parameter.

8. The integrated system for visual arterial blood collection and hemostasis compression according to claim 1, characterized in that, The blood collection status determination module includes: The abnormal number statistics submodule traverses each data node within the abnormal blood flow region set, performs cumulative counting on each independent detection grid defined within the abnormal blood flow region set, extracts the cumulative count value, and generates the number of abnormal units. The spatial distribution calculation submodule calls the number of abnormal units to obtain the total number of sampling and detection units, performs a division operation with the number of abnormal units as the numerator and the total number of sampling and detection units as the denominator, extracts the quotient obtained from the division, and generates the spatial distribution ratio of abnormal units. The stability identification submodule reads the preset classification hyperplane weight vector and classification bias parameter according to the spatial distribution ratio of the abnormal units, performs a product operation on the spatial distribution ratio of the abnormal units and the classification hyperplane weight vector, adds the product result to the classification bias parameter, extracts the numerical sign term, and generates a stability identifier for the blood collection process.

9. The integrated system for visual arterial blood collection and hemostasis compression according to claim 1, characterized in that, The hemostatic pressure adjustment module includes: The data processing submodule, based on the stability indicator of the blood collection process, increases the hemostatic loading force when the stability indicator of the blood collection process meets the preset state parameters, collects the skin pressure deformation corresponding to the hemostatic loading force, reads the preset state transition matrix, performs product calculation on the skin pressure deformation and the preset state transition matrix, and generates a smooth deformation sequence. The gradient calculation submodule calls the smooth deformation sequence, subtracts the previous level parameter from the subsequent level smooth deformation sequence parameter to extract the deformation difference term, divides the deformation difference term by the pressure increment value of the two adjacent levels, and generates the deformation degree change gradient. The interval filtering submodule compares the values ​​of the deformation degree change gradients in ascending order with the values ​​of consecutive levels of deformation degree change gradients, filters the critical pressure parameters corresponding to the downward turning point of the deformation degree change gradients, extracts the pressure fluctuation limits, and generates the target control fine-tuning interval.

10. The integrated system for visual arterial blood collection and hemostasis compression according to claim 9, characterized in that, The process of selecting the critical pressure parameter corresponding to the decrease in the deformation degree gradient is as follows: when the value of the deformation degree gradient at the current level is less than the value of the deformation degree gradient at the previous level, the hemostatic loading force corresponding to the current level is extracted as the critical pressure parameter.

Images