A method for selecting optimal puncture point of blood vessels based on blood vessel size characteristics

Abstract

The present invention discloses a method for selecting an optimal puncture point for a blood vessel based on the characteristics of the blood vessel size, by selecting the central area of ​​an infrared blood vessel image of a vein at an arm, back of hand or wrist as an area of ​​interest for denoising, contrast-limited adaptive histogram equalization algorithm enhancement processing, adaptive threshold binarization, image morphology processing, blood vessel contour extraction, traversal to find the blood vessel contour area with the largest area among all blood vessel contours, distance transformation algorithm to extract the blood vessel centerline and the radius of each position of the blood vessel, selecting the point on the blood vessel centerline where the blood vessel radius is the largest and the relative curvature of the blood vessel centerline is less than a set threshold as the optimal blood vessel puncture point, and being able to obtain the tangent direction of the point on the blood vessel centerline. The present invention realizes the automatic selection of blood vessel puncture points on infrared blood vessel images, has high blood vessel puncture accuracy, and improves the safety and efficiency of blood vessel puncture.

Description

Technical Field

[0001] The present invention relates to the technical field of biometric information perception and recognition, and in particular to a method for selecting an optimal blood vessel puncture point based on blood vessel size characteristics. Background Art

[0002] The manual venipuncture method is limited by the fatigue of medical staff when facing a large number of patients, and its efficiency is low. During the peak period of medical treatment, a large number of patients are likely to wait for medical treatment. In addition, because venipuncture is limited by the work experience of medical staff, it is difficult for medical staff who are new to venipuncture to perform venipuncture. In addition, the patient's condition is unknown. If the protection is not appropriate during the venipuncture, it is easy to cause cross infection, which increases the risk of medical staff's work. In addition, the physical constitution of patients seeking medical treatment is different. Some patients cannot see the clear blood vessel contour position due to vascular sclerosis, thin blood vessels, thick subcutaneous fat, dark skin color, etc., which invisibly increases the difficulty of venipuncture operation for medical staff, and then greatly reduces the success rate of venipuncture, which brings a painful medical experience to patients. Therefore, a new venipuncture point selection algorithm is needed, which is designed according to the principle of experienced medical staff selecting the optimal venipuncture point, with higher selection accuracy and better safety. Summary of the invention

[0003] The present invention provides a method for selecting an optimal vascular puncture point based on vascular size characteristics, which realizes automatic extraction of vascular puncture points on infrared vascular images, has high vascular puncture accuracy, and improves the safety and efficiency of vascular puncture. The method comprises:

[0004] The central area of ​​the infrared blood vessel image is selected as the region of interest for selecting the optimal puncture point of the blood vessel;

[0005] Performing denoising and Gaussian filtering on the image of the region of interest to obtain a denoised image;

[0006] Performing image enhancement processing on the denoised image to obtain an enhanced image;

[0007] Performing image binarization processing on the enhanced image to obtain a binary blood vessel image;

[0008] Performing image morphology processing on the binary blood vessel image to obtain a blood vessel contour image;

[0009] Performing contour detection processing on the blood vessel contour image to obtain a blood vessel maximum contour image;

[0010] The maximum contour image of the blood vessel is processed by a distance transformation algorithm, and the position of each point on the blood vessel centerline and the radius of each blood vessel position are extracted;

[0011] According to the positions of each point on the blood vessel centerline and the radius of each position of the blood vessel, a point on the blood vessel centerline whose curvature meets the threshold requirement and whose blood vessel radius is as large as possible is selected as the optimal blood vessel puncture point.

[0012] Specifically, the image enhancement process is as follows:

[0013] The image enhancement is performed using the contrast-constrained adaptive histogram equalization algorithm.

[0014] Specifically, the image binarization process is as follows:

[0015] Adaptive threshold binarization method is used to perform image binarization.

[0016] Specifically, the image morphological processing is as follows:

[0017] The image is processed by using image morphology opening and closing operations, with opening operation performed first and closing operation second, or closing operation performed first and opening operation second. The effect achieved is to remove background noise in the image and fill the holes in the blood vessel contour area to make the blood vessel contour smooth.

[0018] Specifically, the contour detection process is as follows:

[0019] The blood vessel contour is extracted from the image, and the blood vessel contour area with the largest blood vessel contour area is found and separated.

[0020] Specifically, the distance transformation algorithm is processed as follows:

[0021] The distance transformation algorithm used can obtain the pixel distance from each point in the vascular area to the image background in the vascular contour image, and select the point with the maximum distance in each row as the point on the vascular centerline. The value obtained by the distance transformation algorithm for each point on the centerline is also the vascular radius value at each vascular position.

[0022] Specifically, the specific process of selecting a point on the blood vessel centerline whose curvature meets the threshold requirement and whose blood vessel radius is as large as possible as the optimal blood vessel puncture point is:

[0023] Selecting the point on the centerline of the blood vessel where the blood vessel radius is the largest as the preselected puncture point;

[0024] Obtaining a blood vessel centerline curve equation according to the positions of each point on the blood vessel centerline, calculating the tangent direction of the preselected puncture point on the blood vessel centerline, and calculating the curvature of the blood vessel centerline curve equation at the preselected puncture point;

[0025] When the curvature at the preselected puncture point is greater than a set threshold, the point is excluded, and the optimal puncture point of the blood vessel is reselected from the remaining points on the center line of the blood vessel according to the principle of maximum blood vessel radius;

[0026] Until a puncture point with a curvature less than a set threshold is obtained, the puncture point with a curvature less than the set threshold is selected as the final optimal puncture point of the blood vessel.

[0027] Specifically, the blood vessel centerline curve equation is obtained by fitting the image coordinates of each point on the blood vessel centerline using the least squares method based on a cubic polynomial, and its expression is as follows:

[0028] f(x)=a0+a1x+a2x 2 +a3x 3 (1)

[0029] Where f(x) is the fitting curve equation of the vascular centerline; x is the image coordinate of a point on the vascular centerline; a0~a3 are polynomial coefficients;

[0030] According to the least squares method, the sum of squares of the residuals between the fitting curve equation and the actual points is required to be minimal, and its expression is as follows:

[0031]

[0032] Where E is the residual sum of squares, which is required to be the minimum; f(x i ) is the coordinate of the point on the vascular centerline fitting curve; x i ,y i is the image coordinates of the point on the centerline of the blood vessel;

[0033] Specifically, the tangent direction of the optimal puncture point of the blood vessel on the center line of the blood vessel is selected according to the puncture direction of the blood collection needle on the optimal puncture point of the blood vessel with a set number of points to calculate the average slope of the tangent, and the calculation expression is as follows:

[0034]

[0035] Where k is the average slope at the optimal puncture point; N is the number of points on the center line of the blood vessel set along the puncture direction at the optimal puncture point of the blood vessel; f′(x i ) is the first-order derivative function of the curve equation obtained by fitting the points on the center line of the blood vessel; x i are the image coordinates of each point on the centerline of the blood vessel along the puncture direction of the optimal puncture point of the blood vessel, where x0 is the image coordinates of the selected optimal puncture point of the blood vessel.

[0036] Specifically, the curvature calculation expression of the selected optimal puncture point on the blood vessel centerline curve equation is as follows:

[0037]

[0038] Where K is the curvature, y′ and y″ are the first-order derivative function and the second-order derivative function of the fitted blood vessel centerline curve equation, respectively.

[0039] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0040] (1) The present invention proposes an optimal venous puncture point selection algorithm based on the venous vessel size characteristics for infrared vascular images. The algorithm imitates the principle of selecting the puncture point for manual venous puncture and extracts the blood vessels with the largest contour area in the infrared vascular image. This algorithm conforms to the principle of selecting superficial subcutaneous blood vessels, the thickest and clearest blood vessels, and the straightest and most convenient blood vessels for puncture by medical staff. It realizes the automatic selection of venous puncture points and improves the accuracy, reliability and safety of venous puncture in selecting the optimal venous puncture point.

[0041] (2) The algorithm method for selecting the optimal puncture point of a blood vessel based on the size characteristics of a vein is simple to operate and calculate, easy to implement, and highly reliable, and effectively improves the efficiency and accuracy of puncture point extraction. The present invention is applied to the field of intelligent medical devices, such as intelligent intravenous infusion robots and intravenous blood sampling robots, which can free medical staff from complicated intravenous puncture work and high-risk working environments, and is conducive to the intelligent selection of the optimal puncture point by the intelligent intravenous infusion robots and intravenous blood sampling robots, and enables the robots to automatically complete intravenous infusion and blood sampling actions, which is conducive to the promotion and application of some intelligent medical instruments and equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic flow chart of a method for selecting an optimal puncture point for a blood vessel based on blood vessel size characteristics provided by an embodiment of the present invention;

[0043] Figure 2 The original image of the blood vessels of the arm from which blood is drawn (a) and the final selected puncture point effect image (b) provided in an embodiment of the present invention;

[0044] Figure 3 A schematic diagram of a process for selecting an optimal puncture point for a blood vessel based on blood vessel size characteristics provided by an embodiment of the present invention;

[0045] Figure 4 A schematic diagram of a blood vessel contour, a blood vessel centerline, the selection of an optimal puncture point, and a tangent line of the blood vessel centerline at the puncture point provided in an embodiment of the present invention.

[0046] Among them: 1 is the blood vessel contour; 2 is the blood vessel centerline; 3 is the selected optimal puncture point; 4 is the tangent of the selected optimal puncture point on the blood vessel centerline. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and examples. The present invention is further described in detail. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0048] See also Figure 1 , Figure 2 , Figure 3 and Figure 4 ,in Figure 2 and Figure 4 The numbers represent: 1 is the blood vessel contour; 2 is the blood vessel centerline; 3 is the selected optimal puncture point; 4 is the tangent of the selected optimal puncture point on the blood vessel centerline.

[0049] The preferred embodiment of the present invention provides an optimal puncture point selection algorithm method based on the venous blood vessel size characteristics, and the optimal puncture point selection algorithm method mainly includes the following steps:

[0050] Step S110: Select Figure 2 The central area of ​​the infrared blood vessel image shown in (a) is used as the region of interest for selecting the optimal puncture point for blood vessel puncture to eliminate the influence of image background interference.

[0051] Specifically, first, a near-infrared camera is used to photograph the skin of the blood vessel to be punctured, and the blood vessel direction is required to be along the vertical direction of the image, and an infrared blood vessel image is obtained, such as Figure 2 Then, the region of interest (target region) in the central part of the infrared image is extracted to remove the influence of the skin contour on the image.

[0052] Step S120: De-noising and Gaussian filtering are performed on the selected infrared blood vessel region of interest image; in the infrared blood vessel image, there are noise points such as skin pores, hair, spots, etc. near the blood vessel area. In view of this situation, the infrared blood vessel image is de-noised and Gaussian filtered to make the blood vessel contour smooth while denoising and fading the noise points.

[0053] Specifically, Figure 2 The infrared blood vessel image of the region of interest (a) is subjected to denoising and Gaussian filtering normalization processing, so that the infrared blood vessel image of the region of interest is denoised and the noise points are reduced while the blood vessel contour becomes smooth. Figure 3 (a).

[0054] Step S130: Using the Contrast Limited Adaptive Histogram Equalization (CLAHE) algorithm to Figure 3The image enhancement processing is performed on the vascular area of ​​(a); since the image contrast will be reduced after denoising and Gaussian filtering of the infrared vascular image, the contrast-limited adaptive histogram equalization algorithm (CLAHE) is used to Figure 3 (a) The blood vessel area is enhanced, such as Figure 3 as shown in (b).

[0055] Specifically, Figure 3 (a) The image is enhanced by the Contrast Limited Adaptive Histogram Equalization (CLAHE) algorithm, with a convolution kernel size of 8×8 and a contrast threshold of 4.

[0056] Step S140: Adopting an adaptive threshold binarization method to Figure 3 (b) is used for image binarization. Since the quality of infrared vascular images of different patients is different, the skin color factor directly affects the contour range of the vascular area in the image relative to the image skin background. To address this problem, the adaptive threshold binarization method can effectively improve the accuracy and versatility of extracting vascular contours. After adaptive threshold binarization, Figure 3 (c).

[0057] Step S150: Use image morphology opening and closing operations to Figure 3 (c) is subjected to morphological processing, and the infrared vascular image is binarized, as shown in Figure 3 As shown in (c), some skin pores, hair, and spots will form small spots distributed around the blood vessel contour. In view of this situation, image morphology opening and closing operations are used to remove the burr noise points in the image background and fill the holes in the blood vessel contour area to make the blood vessel contour smooth. Figure 3 (d).

[0058] Specifically, the Figure 3 (c) is first closed and then opened to obtain Figure 3 (d), the convolution kernel size of the morphological opening and closing operations is 24×24.

[0059] Step S160: Use contour detection to Figure 3 (d) performs vascular contour extraction, traverses to find the vascular contour area with the largest vascular contour area and separates it to obtain a new vascular contour image, such as Figure 3 As shown in (e); after the vascular contour is extracted from the infrared vascular image, multiple vascular contours of different sizes will be obtained. By traversing the vascular contour area with the largest vascular contour area, the vascular contour with the largest area is obtained to meet the principle of manually selecting the thickest and most convenient blood vessel for needle insertion in the venous puncture point. Then the vascular contour with the largest area is separated to obtain a new vascular image.

[0060] Specifically, the traversal to find the blood vessel contour only traverses the outer contour of the blood vessel, and after finding the blood vessel contour with the largest area, an image containing only the largest blood vessel contour is drawn, and then the image is reversed to black and white to obtain Figure 3 (e).

[0061] Step S170: Use distance transformation algorithm to Figure 3 (e) is processed to obtain the distance from each point in the maximum contour area of ​​the blood vessel to the image background, and then the position of each point on the blood vessel centerline and the radius of each blood vessel position are obtained.

[0062] Specifically, since the principle of the distance transformation algorithm is to calculate the shortest distance from each point in the white foreground area of ​​the binary image to each point in the black background area of ​​the image, it is necessary to invert the black and white of the blood vessel contour to white and the background to black, such as Figure 3 As shown in (e). The distance transformation algorithm used is the Euclidean distance calculation. Figure 3 The distance from each point in the white blood vessel area to the black background in (e).

[0063] Specifically, after the distance transformation algorithm obtains the pixel distance from each point in the blood vessel area to the image background in the blood vessel contour image, the image coordinates of each point on the blood vessel centerline and the radius of each blood vessel position are obtained by scanning the maximum value of the distance from each row of pixels in the blood vessel contour area to the background and the position of the point.

[0064] Step S180: According to the positions of each point on the blood vessel centerline and the radius of each blood vessel position in step S170, the point with the maximum blood vessel radius on the blood vessel centerline is selected as the pre-selected puncture point, such as Figure 4 As shown in the reference numeral 3; the curve equation of the vascular centerline is obtained by fitting the image coordinates of each point on the vascular centerline according to the least squares method based on a cubic polynomial, as shown in Figure 4 As shown in the reference numeral 2, the tangent direction of the preselected puncture point on the center line of the blood vessel is obtained by solving, such as Figure 4 As shown in the label 4; at the same time, the curvature of the pre-selected puncture point on the curve equation of the blood vessel centerline is solved to determine whether the curvature of the pre-selected puncture point meets the threshold requirement; when the curvature of the pre-selected puncture point is greater than the set threshold, the point position is excluded, and the optimal puncture point of the blood vessel is re-selected from the remaining points on the blood vessel centerline according to the principle of maximum blood vessel radius; until a puncture point with a curvature less than the set threshold is obtained, the puncture point with a curvature less than the set threshold is selected as the optimal puncture point of the blood vessel. Specifically, in the step S180, the point with the largest value of the distance transformation result in the point concentration on the blood vessel centerline is selected as the undetermined optimal puncture point, which is the point with the largest blood vessel radius.

[0065] Preferably, the selected optimal puncture point position and the contour and centerline information of the blood vessel can be superimposed on the original image of the blood vessel ( Figure 2(a)) allows medical staff to more intuitively observe the actual effect of the selected optimal puncture point on the entire vascular original image, such as Figure 2 as shown in (b).

[0066] Specifically, in step S180, the blood vessel centerline curve equation is obtained by fitting the least square method based on a cubic polynomial according to the positions of each point on the blood vessel centerline, and its expression is as follows:

[0067] f(x)=a0+a1x+a2x 2 +a3x 3 (1)

[0068] Where f(x) is the fitting curve equation of the vascular centerline; x is the image coordinate of a point on the vascular centerline; a0~a3 are polynomial coefficients;

[0069] According to the least squares method, the sum of squares of the residuals between the fitting curve equation and the actual points is required to be minimal, and its expression is as follows:

[0070]

[0071] Where E is the residual sum of squares, which is required to be the minimum; f(x i ) is the coordinate of the point on the vascular centerline fitting curve; x i ,y i is the image coordinate of the point on the center line of the blood vessel

[0072] Specifically, in step S180, the tangent direction of the optimal puncture point of the blood vessel on the center line of the blood vessel is selected according to the puncture direction of the blood sampling or infusion needle on the optimal puncture point of the blood vessel. The slope average value is calculated by selecting the tangent slopes of each point of the blood vessel center line curve equation according to the set number of points. The calculation expression is as follows:

[0073]

[0074] Where k is the average slope at the optimal puncture point; N is the number of points on the center line of the blood vessel set along the puncture direction at the optimal puncture point of the blood vessel; f′(x i ) is the first-order derivative function of the curve equation obtained by fitting the points on the center line of the blood vessel; x i are the image coordinates of each point on the centerline of the blood vessel along the puncture direction of the optimal puncture point of the blood vessel, where x0 is the image coordinates of the selected optimal puncture point of the blood vessel.

[0075] Specifically, in step S180, the curvature calculation expression of the selected optimal puncture point on the blood vessel centerline curve equation is as follows:

[0076]

[0077] Where K is the curvature, y′ and y″ are the first-order derivative function and the second-order derivative function of the fitted blood vessel centerline curve equation, respectively.

[0078] The present invention provides an optimal blood vessel puncture point selection algorithm based on blood vessel size characteristics. By processing infrared blood vessel images of veins at the arm, back of hand, elbow and other parts, the blood vessel contour position, blood vessel centerline position, tangent slope and curvature, and blood vessel size characteristic information of the radius of each part of the blood vessel are obtained to select the best needle point. Automatic selection of blood vessel puncture points is achieved, the accuracy and reliability of optimal vein puncture point selection and the safety of vein puncture are improved, and the problem that intelligent infusion or blood sampling robots are difficult to select accurate and reliable vein puncture points is solved.

[0079] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for selecting the optimal puncture point of a blood vessel based on blood vessel size characteristics, characterized in that: The central area of ​​the infrared blood vessel image is selected as the region of interest for selecting the optimal puncture point of the blood vessel; Performing denoising and Gaussian filtering on the image of the region of interest to obtain a denoised image; Performing image enhancement processing on the denoised image to obtain an enhanced image; Performing image binarization processing on the enhanced image to obtain a binary blood vessel image; Performing image morphology processing on the binary blood vessel image to obtain a blood vessel contour image; Performing contour detection processing on the blood vessel contour image to obtain a blood vessel maximum contour image; The maximum contour image of the blood vessel is processed by a distance transformation algorithm, and the position of each point on the blood vessel centerline and the radius of each blood vessel position are extracted; According to the positions of each point on the blood vessel centerline and the radius of each position of the blood vessel, a point on the blood vessel centerline whose curvature meets the threshold requirement and whose blood vessel radius is as large as possible is selected as the optimal blood vessel puncture point; The distance transformation algorithm is specifically processed as follows: The distance transformation algorithm used can obtain the pixel distance from each point in the blood vessel area to the image background in the blood vessel contour image, and select the maximum distance point in each row as the point on the blood vessel centerline. The value of the point on the blood vessel centerline obtained by the distance transformation algorithm is also the blood vessel radius value at each position of the blood vessel. The specific process of selecting the point on the blood vessel centerline whose curvature meets the threshold requirement and whose blood vessel radius is as large as possible as the optimal blood vessel puncture point is: Selecting the point on the centerline of the blood vessel where the blood vessel radius is the largest as the preselected puncture point; Obtaining a blood vessel centerline curve equation according to the positions of each point on the blood vessel centerline, calculating the tangent direction of the preselected puncture point on the blood vessel centerline, and calculating the curvature of the blood vessel centerline curve equation at the preselected puncture point; When the curvature at the preselected puncture point is greater than a set threshold, the point is excluded, and the optimal puncture point of the blood vessel is reselected from the remaining points on the center line of the blood vessel according to the principle of maximum blood vessel radius; Until a puncture point with a curvature less than a set threshold is obtained, the puncture point with a curvature less than the set threshold is selected as the final optimal puncture point of the blood vessel.

2. The method for selecting the optimal puncture point of a blood vessel based on blood vessel size characteristics according to claim 1, characterized in that: The image enhancement process is specifically as follows: The image enhancement is performed using contrast-constrained adaptive histogram equalization algorithm.

3. The method for selecting the optimal puncture point of a blood vessel based on blood vessel size characteristics according to claim 1, characterized in that: The image binarization process is specifically as follows: Adaptive threshold binarization method is used to perform image binarization.

4. The method for selecting the optimal puncture point of a blood vessel based on blood vessel size characteristics according to claim 1, characterized in that: The image morphological processing is specifically as follows: The image is processed by using image morphology opening and closing operations, with opening operation performed first and then closing operation, or closing operation performed first and then opening operation. The effect achieved is to remove background noise in the image and fill the holes in the blood vessel contour area to make the blood vessel contour smooth.

5. The method for selecting the optimal puncture point of a blood vessel based on blood vessel size characteristics according to claim 1, characterized in that: The contour detection process is specifically as follows: The blood vessel contour is extracted from the image, and the blood vessel contour area with the largest blood vessel contour area is found and separated.

6. The method for selecting the optimal puncture point of a blood vessel based on blood vessel size characteristics according to claim 1, characterized in that: The blood vessel centerline curve equation is obtained by fitting the image coordinates of each point on the blood vessel centerline using the least squares method based on a cubic polynomial, and its expression is as follows: （1） in Fitting curve equation for the blood vessel centerline; x is the image coordinates of the point on the centerline of the blood vessel; ~ are the polynomial coefficients; According to the least squares method, the sum of squares of the residuals between the fitting curve equation and the actual points is required to be minimal, and its expression is as follows: （2） Where E is the residual sum of squares, and E is required to be the minimum; is the coordinate of the point on the blood vessel centerline fitting curve; x i , y i are the image coordinates of a point on the centerline of the blood vessel.

7. The method for selecting the optimal puncture point of a blood vessel based on blood vessel size characteristics according to claim 1, characterized in that: The direction of the tangent of the optimal puncture point of the blood vessel on the center line of the blood vessel is calculated by selecting points of the curve equation of the blood vessel center line with a set number of points according to the puncture direction of the blood collection needle at the optimal puncture point of the blood vessel, and the average slope of the tangent is calculated as follows: （3） Where k is the average slope at the optimal puncture point; N is the number of points on the centerline of the blood vessel set along the puncture direction at the optimal puncture point of the blood vessel; It is the first-order derivative function of the curve equation obtained by fitting the points on the center line of the blood vessel; x i is the image coordinates of each point on the center line of the blood vessel along the puncture direction at the optimal puncture point of the blood vessel, where is the image coordinate of the optimal puncture point of the selected blood vessel.

8. The method for selecting the optimal puncture point of a blood vessel based on blood vessel size characteristics according to claim 1, characterized in that: The curvature calculation expression of the selected optimal puncture point on the blood vessel centerline curve equation is as follows: （4） in K is the curvature, and are the first-order derivative and second-order derivative functions of the fitted vascular centerline curve equation.

Images