A vascular intervention surgery robot control method and system

Abstract

The present invention discloses a vascular interventional surgery robot control method and system, the method includes angiography image processing; including image enhancement, image segmentation and image correction after segmentation, feature extraction of the processed image, combining vascular image features and other patient information to obtain vascular state index values, obtaining guidewire control parameters according to the vascular state index values, and correcting the control parameters, and controlling the vascular interventional surgery robot based on the corrected PID control parameter values. The present invention realizes the adjustment of guidewire control parameters according to different vascular states, so that the guidewire travel control matches the vascular state, and ensures the safety of guidewire entry and travel efficiency.

Description

Technical Field

[0001] The present invention relates to the field of medical devices, and in particular to a vascular intervention surgery robot control method and system. Background Art

[0002] In the field of vascular surgery, vascular interventional surgery is an efficient minimally invasive treatment method. During the operation, the doctor stands next to the patient lying on the operating table to perform minimally invasive surgery, while observing the position of the catheter / guidewire in the body through DSA images. The guidewire and catheter are usually inserted into the femoral artery or radial artery. After arterial puncture, the doctor advances / withdraws and twists the guidewire and catheter into the target artery, and then reaches the lesion along the blood vessel for treatment. However, traditional interventional surgical methods require doctors to wear protective lead clothing for a long time and there is a risk of exposure to radiation, which not only poses a health hazard to doctors, but may also have an adverse effect on the efficiency of the operation. With the development and application of intravascular interventional surgical robots, it can reduce the radiation exposure of doctors while providing higher operating accuracy and stability than traditional methods.

[0003] However, most of the existing intravascular interventional surgical robots are fully controlled by doctors at a remote control console, and the operation of the instrument during the operation is entirely determined by the doctor. This tests the doctor's operational ability and experience. Unskilled doctors often cannot control the operating table well, which introduces more risks to the operation process. In addition, there are huge differences in the health of blood vessels for patients of different ages and causes of disease. The speed of advancing the guidewire and other operations often rely on the doctor's experience and judgment, and cannot be objectively evaluated. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In order to solve the above technical problems, the present invention provides a vascular interventional surgery robot control method and system.

[0006] (II) Technical solution

[0007] In order to solve the above-mentioned technical problems and achieve the purpose of the invention, the present invention is implemented by the following technical solutions:

[0008] A vascular interventional surgery robot control method comprises the following steps:

[0009] S1: Angiography image processing; including image enhancement, image segmentation and post-segmentation image correction steps;

[0010] S2: Extract features from the processed image to obtain the vessel lumen diameter, vessel wall thickness, and whether there is vascular artery stenosis or occlusion;

[0011] S3: combining the vascular image features and other patient information to obtain a vascular status index value, wherein the vascular status index value is divided into k levels according to the health status of the blood vessels;

[0012] S4: obtaining guidewire control parameters according to the blood vessel status index and modifying the control parameters;

[0013] S5: Controlling the vascular interventional surgery robot based on the corrected PID control parameter values.

[0014] Furthermore, the image enhancement in step S1 includes the following steps:

[0015] a. Use mean filtering to estimate the noise of the original image, and the estimated value is ;

[0016] b. Select a window with a length and width of win and divide the image into blocks;

[0017] in,

[0018] Where k is the adjustment coefficient, , ;

[0019] is the standard deviation of the image; are the width and height of the original image respectively;

[0020] c. Filter the sub-blocks that have not overflowed after adding noise;

[0021] d. Calculate the standard deviation of the sub-blocks that have not overflowed, and sort the standard deviations from small to large;

[0022] e. Calculation of statistical parameters ;

[0023] f. Calculate the noise standard deviation estimate: , where MN is the total number of sub-blocks, is the number of overflow sub-blocks;

[0024] g. De-noise the initial image according to the noise standard deviation estimate calculated in step f.

[0025] Furthermore, the post-segmentation image correction in step S1 refers to correcting the hollowed-out area, which specifically includes the following steps:

[0026] a. Copy the segmented image to generate a new image to fill the hollowed-out area;

[0027] b. Create a zero matrix with an aspect ratio n larger than the original input image as a mask based on the image size;

[0028] c. Traverse the pixel values ​​of the image and find the first point with a pixel value of 0 as the starting point for filling. Then use the filling function of OpenCV to fill it. The result is an image filled with 255.

[0029] d. Invert the image to get the inverse image, then perform a bitwise OR operation on the original image and the inverse image to get the image with the holes filled in.

[0030] Furthermore, in step S2, feature extraction is performed based on the YOLOv5 target detection algorithm.

[0031] Furthermore, other patient information in step S3 includes age, gender, BMI index, blood pressure, total cholesterol, fasting blood sugar, cholesterol, triglycerides, and thyroid stimulating hormone.

[0032] Furthermore, the step S4 includes:

[0033] S41: prediction of guidewire introduction position, speed, and guide angle;

[0034] Assume that the state quantity and control quantity of the guide wire are expressed as follows:

[0035] in, ,in, is the horizontal coordinate of the guide wire position, is the ordinate of the guidewire position, is the guide wire angle, is the guidewire advancement speed;

[0036] The set value of the output state is obtained through the prediction model, that is, the state quantity and control quantity of the next state are obtained, including the guide wire introduction position, speed, and guide angle;

[0037] The input of the prediction model includes the vascular state index value, the current guidewire state value and the control value. , the output of the prediction model is the state quantity and control quantity of the next state;

[0038] S42: Calculating motion control parameters based on a feedback learning network algorithm;

[0039] S43: Modify the control parameters based on the gravity of the guidewire catheter and the friction coefficient of the blood vessel.

[0040] Furthermore, the step S42 includes:

[0041] a. Initialize the feedback learning network parameters, set the network related parameters and PID parameter initial values, and set the PID parameter range;

[0042] b. Calculate the current reward function value and put the reward function value and current state information into the buffer area;

[0043] The reward function is expressed as follows: , , , Reward for safety, is the damage factor, col is the impact strength value, and its value comes from the force sensor at the front end of the guidewire; Reward for stability, is the adjustment factor, a is the acceleration, is the performance value, which may be the mean value of motion control deviation;

[0044] c. Calculate the objective function value and update the network parameters according to the direction of the optimal objective function to complete the network iterative learning;

[0045] The objective function is as follows:

[0046] in, is the motion error, t is the time, is the time of a sampling cycle;

[0047] d. Output PID parameters to the PID controller, perform a wire movement, calculate the performance value of this movement process, and input it into the network;

[0048] e. Determine whether the performance of this motion process meets the control performance requirements. If so, output the PID control parameters and end the iterative learning. Otherwise, repeat steps b~d.

[0049] Further, the step S43 includes a: obtaining the total mass of the guidewire catheter and calculating the total gravity

[0050] in, is the mass per unit length of the guide wire, is the mass per unit length of the catheter;

[0051] b: Calculate the vascular friction coefficient, specifically, calculate the vascular friction coefficient based on the vascular status index value

[0052] Among them, K is the vascular status index value, r is the adjustment coefficient, which is greater than 1 and is determined according to the severity of vascular arterial stenosis and occlusion;

[0053] c. Calculate the corrected PID parameters

[0054] Said , , is the PID parameter adjustment amount, , , is the PID parameter before correction.

[0055] The present invention also provides a vascular interventional surgery robot control system, which includes:

[0056] Angiography image processing module, which is used for image enhancement, image segmentation and image correction after segmentation;

[0057] Image feature extraction module, which is used to extract features from the processed image based on the YOLOv5 target detection algorithm to obtain the vessel lumen diameter, vessel wall thickness, and whether there is vascular artery stenosis or occlusion;

[0058] A blood vessel status index value calculation module, which is used to obtain a blood vessel status index value according to the blood vessel image characteristics and other patient information, wherein the blood vessel status index value is divided into k levels according to the health status of the blood vessel;

[0059] A guidewire control parameter calculation module is used to obtain guidewire control parameters according to the blood vessel status index, including guidewire introduction position, speed, and guide angle prediction, control the motion parameters based on the feedback learning network algorithm, and modify the control parameters in combination with the guidewire catheter gravity and friction;

[0060] A control module is used to control the vascular interventional surgery robot based on the corrected PID parameter values.

[0061] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which program instructions of a vascular interventional surgery robot control method are stored. The program instructions of the vascular interventional surgery robot control can be executed by one or more processors to implement the steps of the vascular interventional surgery robot control method as described above.

[0062] (III) Beneficial effects

[0063] Compared with the prior art, the present invention has the following beneficial effects:

[0064] (1) The present invention adjusts the guidewire control parameters according to different vascular conditions, so that the guidewire advancement control matches the vascular conditions, thereby ensuring the safety of guidewire entry and advancement efficiency.

[0065] (2) The present invention uses a noise estimation method to perform image enhancement on vascular images, thereby improving the recognizability of vascular image features and providing favorable assistance for subsequent image recognition.

[0066] (3) The present invention segments the vascular image based on a multi-feature image segmentation algorithm, without knowing the dependency between the features of all channels, thus reducing the computational network complexity. The segmented image is corrected to repair the hollow areas in the blood vessels.

[0067] (4) In the invention, a feedback learning network algorithm is used to control the motion parameters of the guide wire, thereby optimizing the control parameters and ensuring the stability and safety of the controlled motion.

[0068] (5) The present invention combines the gravity and friction of the guidewire catheter to correct the control parameters, thereby achieving adaptive guidewire motion control and avoiding damage to the blood vessels by the guidewire. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0070] FIG1 is a schematic flow chart of a control method for a vascular interventional surgery robot according to an embodiment of the present application;

[0071] FIG. 2 is a schematic diagram of a feedback learning network algorithm flow according to an embodiment of the present application. DETAILED DESCRIPTION

[0072] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0073] The following describes the embodiments of the present disclosure through specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.

[0074] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The drawings only show components related to the present disclosure rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0075] Referring to FIG1 , a vascular interventional surgery robot control method includes the following steps:

[0076] S1: Angiography image processing; including image enhancement, image segmentation and image correction after segmentation. The specific steps are as follows:

[0077] S11: Image Enhancement

[0078] Enhancement of angiographic images is an important part of medical image processing. In the process of medical imaging, due to the existence of complex conditions such as uneven distribution and attenuation of contrast agents and uneven X-ray exposure, the characteristics of vascular images are low contrast and blur, and they are disturbed by complex background structures. At different imaging angles and distances, the vascular structure presents different forms, which makes it difficult to enhance angiographic images. Therefore, it is necessary to enhance the images. Image enhancement is performed according to the following steps:

[0079] a. Use mean filtering to estimate the noise of the original image, and the estimated value is ;

[0080] b. Select a window with a length and width of win and divide the image into blocks;

[0081] in,

[0082] Where k is the adjustment coefficient, , ; is the standard deviation of the image; are the width and height of the original image respectively.

[0083] c. Filter the sub-blocks that have not overflowed after adding noise;

[0084] d. Calculate the standard deviation of the sub-blocks that have not overflowed, and sort the standard deviations from small to large;

[0085] e. Calculation of statistical parameters ;

[0086] f. Calculate the noise standard deviation estimate: , where MN is the total number of sub-blocks, is the number of overflow sub-blocks.

[0087] g. De-noise the initial image according to the noise standard deviation estimate calculated in step f.

[0088] S12: Image Segmentation

[0089] The present invention segments the blood vessel image based on a multi-feature image segmentation algorithm, which specifically includes the following steps:

[0090] a. Perform cross fusion of input features at all levels. The fusion feature map of the i-th layer is shown as follows:

[0091] in, is the feature input of the p-th layer, J represents the convolution operation, and C represents the sampling operation.

[0092] b. Perform a global average pooling operation and compress the feature image after pooling. The compression process is expressed as:

[0093] in, are the fusion feature maps of the i-th layer respectively. The width and height of the pixel; m and n represent the coordinates of the pixel point. Represents the fusion feature map of the i-th layer at the (m,n)-th position;

[0094] c. Use one-dimensional convolution to complete the cross-channel interaction between adjacent feature vectors. The interaction operation is as follows:

[0095] in, represents the activation function, , is the convolution weight and K is the kernel.

[0096] The above method does not need to know the dependencies between the features of all channels, thus reducing the computational network complexity.

[0097] S13: Image correction after segmentation

[0098] Since there may be a hollow area in the middle of the blood vessel after the blood vessel image is segmented, which will affect the subsequent image analysis results, the present invention corrects the hollow area, specifically including the following steps:

[0099] a. Copy the segmented image to generate a new image to fill the hollowed-out area;

[0100] b. Create a zero matrix with an aspect ratio n larger than the original input image as a mask based on the image size;

[0101] c. Traverse the pixel values ​​of the image and find the first point with a pixel value of 0 as the starting point for filling. Then use the filling function of OpenCV to fill it. The result is a 255-filled image.

[0102] d. Invert the image to get the inverse image, then perform a bitwise OR operation on the original image and the inverse image to get the image with the holes filled in.

[0103] After this processing, the accuracy and stability of the image segmentation results can be effectively improved, the problem of rings in the center line of the blood vessels can be avoided, and a more reliable foundation can be provided for the subsequent construction and analysis of the blood vessel tree.

[0104] S2: Extract features from the processed image to obtain the vessel lumen diameter, vessel wall thickness, and whether there is vascular artery stenosis or occlusion. The feature extraction step is based on the YOLOv5 target detection algorithm. The specific steps are as follows:

[0105] S21. Establishing a sample library, the sample library includes image samples of different vascular lumen diameters, vascular wall thicknesses, the presence of vascular arterial stenosis, occlusion, and the absence of vascular arterial stenosis, occlusion;

[0106] S22. Perform learning and training based on the YOLOv5 detection algorithm, including:

[0107] Transfer learning is performed based on the trained YOLOv5 detection model. The trained YOLOvS detection model feature extraction network parameters are used to initialize the feature extraction network parameters in the hierarchical network. On this basis, the feature extraction network parameters of the hierarchical network are frozen, and the feature fusion and classification head parameters are retrained using back propagation. The parameters of the fine-tuning model are updated as the verification loss value decreases. Finally, the feature extraction network parameters are unfrozen and training continues until the loss value no longer decreases.

[0108] S23. Loss function design

[0109] Based on the multi-objective cross entropy loss function, the loss function is expressed as follows:

[0110] Where n is the number of samples, c is the number of categories, is the weight of category j, is the true label of the i-th sample belonging to category j, The probability of category j predicted by the i-th sample network.

[0111] Furthermore, It can be calculated as follows:

[0112] in, is the number of samples of input category j.

[0113] The feature extraction algorithm achieves high accuracy, fast speed, and lightweight target detection, and can quickly and accurately identify vascular features. The algorithm adopts a lightweight design, achieves high performance and detection accuracy in a small storage space, and solves the problem of inaccurate classification results caused by class imbalance in the hierarchical network, which is particularly important for feature detection.

[0114] S24: Extract features of the vascular image based on the trained model to obtain the vascular lumen diameter, vascular wall thickness, and whether there is vascular artery stenosis or occlusion.

[0115] S3: Combining the vascular image features and other patient information to obtain a vascular status index value, wherein the vascular status index value is divided into k levels according to the health status of the blood vessels.

[0116] The blood vessel image features include the diameter of the blood vessel lumen, the thickness of the blood vessel wall, and whether there is stenosis or occlusion of the blood vessel artery;

[0117] Other patient information included age, gender, BMI index, blood pressure, total cholesterol, fasting blood glucose, cholesterol, triglycerides, and thyroid-stimulating hormone.

[0118] Optionally, the vascular status index value is calculated based on a random forest algorithm.

[0119] S4: Obtain guidewire control parameters based on vascular status indicators

[0120] Since different blood vessel states can withstand different guidewire speeds, for example, the blood vessel state of the elderly is poor, and the blood vessel elasticity and other properties are poor. If the guidewire is introduced too quickly, it is easy to cause blood vessel damage, etc. In addition, for people with cardiovascular diseases, the blood flow velocity in the blood vessels is slow and the blood flow resistance is large, which also has a great impact on the control parameters of the guidewire movement. Therefore, the present invention calculates the guidewire control parameters in combination with the blood vessel state index value, including the following steps:

[0121] S41: Prediction of guidewire introduction position, speed, and guide angle

[0122] Assume that the state quantity and control quantity of the guide wire are expressed as follows:

[0123] in, ,in, is the horizontal coordinate of the guide wire position, is the ordinate of the guidewire position, is the guide wire angle, is the guidewire advancement speed.

[0124] Taylor expand the above equation and omit the higher-order terms:

[0125] Subtract the above two equations to get the linearized kinematic model of the guide wire:

[0126] in, , , , , Discretize the above formula and convert it into:

[0127] in, , , , , T is the sampling time.

[0128] In order to ensure the speed stability, the above formula is transformed to obtain the output state expression to be predicted:

[0129] in, , , , m and n are the dimensions of state quantity and control quantity respectively, is the identity matrix.

[0130] The set value of the output state is obtained through the prediction model , that is, the state quantity and control quantity of the next state are obtained, including the guide wire introduction position, speed, and guide angle.

[0131] The input of the prediction model includes the vascular state index value, the current guidewire state value and the control value. , the output of the prediction model is .

[0132] S42: Calculating motion control parameters based on the feedback learning network algorithm, as shown in FIG2, includes the following steps:

[0133] a. Initialize the feedback learning network parameters, set the network related parameters and PID parameter initial values, and set the PID parameter range;

[0134] b. Calculate the current reward function value and put the reward function value and current state information into the buffer area;

[0135] The reward function is expressed as follows: , , , Reward for safety, is the damage factor, col is the impact strength value, and its value comes from the force sensor at the front end of the guidewire; Reward for stability, is the adjustment factor, a is the acceleration, is a performance value, which may be a mean value of motion control deviation.

[0136] c. Calculate the objective function value and update the network parameters in the direction of the optimal objective function to complete the network iterative learning.

[0137] The objective function is as follows:

[0138] in, is the motion error, t is the time, is the time of one sampling cycle.

[0139] d. Output PID parameters to the PID controller, perform a wire movement, calculate the performance value of this movement process, and input it into the network;

[0140] e. Determine whether the performance of this motion process meets the control performance requirements. If so, output the PID control parameters and end the iterative learning. Otherwise, repeat steps b~d.

[0141] S43: Correction of control parameters

[0142] Since the overshoot and other conditions generated during the control process are related to the blood vessel state, and since the guidewire catheter is affected by its own gravity and the friction with the blood in the blood vessel and the blood vessel wall, the control parameters are affected to a certain extent. Therefore, the present invention corrects the control parameters based on the blood vessel state index value and the guidewire gravity, and the correction method is as follows:

[0143] a: Obtain the total mass of the guidewire catheter and calculate the total gravity ,in, is the mass per unit length of the guide wire, is the mass per unit length of the catheter.

[0144] b: Calculate the vascular friction coefficient, specifically, calculate the vascular friction coefficient based on the vascular status index value

[0145] Among them, K is the vascular status index value, r is the adjustment coefficient, whose value is greater than 1 and is determined according to the severity of vascular artery stenosis and occlusion.

[0146] c. Calculate the corrected PID parameters

[0147] Said , , is the PID parameter adjustment amount, , , is the PID parameter before correction.

[0148] Optionally, the above adjustment amount is obtained through a neural network, and the neural network input includes total gravity and vascular friction coefficient.

[0149] S5: Controlling the vascular interventional surgery robot based on the compensated PID parameter values.

[0150] In this embodiment, the guidewire control parameters are adjusted according to different blood vessel conditions so that the guidewire movement control matches the blood vessel condition, ensuring the safety of guidewire entry and the travel efficiency. The control parameters are corrected in combination with the gravity and friction of the guidewire catheter to achieve adaptive guidewire movement control and avoid damage to the blood vessel caused by the guidewire.

[0151] The embodiment of the present invention further provides a vascular intervention surgery robot control system, which includes:

[0152] Angiography image processing module, which is used for image enhancement, image segmentation and image correction after segmentation;

[0153] Image feature extraction module, which is used to extract features from the processed image based on the YOLOv5 target detection algorithm to obtain the vessel lumen diameter, vessel wall thickness, and whether there is vascular artery stenosis or occlusion;

[0154] The blood vessel status index value calculation module is used to obtain the blood vessel status index value according to the blood vessel image characteristics and other patient information. The blood vessel status index value is divided into k levels according to the health status of the blood vessel.

[0155] A guidewire control parameter calculation module is used to obtain guidewire control parameters according to the blood vessel status index, including guidewire introduction position, speed, and guide angle prediction, control the motion parameters based on the feedback learning network algorithm, and modify the control parameters in combination with the guidewire catheter gravity and friction;

[0156] A control module is used to control the vascular interventional surgery robot based on the corrected PID parameter values.

[0157] In addition, an embodiment of the present invention also proposes a computer-readable storage medium, on which program instructions of a vascular interventional surgery robot control method are stored. The vascular interventional surgery robot control program instructions can be executed by one or more processors to implement the steps of the vascular interventional surgery robot control method as described above.

[0158] The embodiments described above are only descriptions of the preferred implementation modes of the present invention, and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A vascular interventional surgery robot control system, characterized in that: include: Angiography image processing module, which is used for image enhancement, image segmentation and image correction after segmentation; The image feature extraction module is used to extract features from the processed image based on the YOLOv5 target detection algorithm to obtain the diameter of the blood vessel lumen, the thickness of the blood vessel wall, and whether there is stenosis or occlusion of the blood vessel artery; A vascular status index value calculation module, which is used to obtain a vascular status index value according to vascular image features and other patient information, wherein the vascular status index value is divided into k levels according to the health status of the blood vessels; A guidewire control parameter calculation module is used to obtain guidewire control parameters according to the vascular status index, including guidewire introduction position, speed, and guide angle prediction, control the control parameters based on the feedback learning network algorithm, and correct the control parameters in combination with the gravity of the guidewire catheter and the vascular friction coefficient; The guidewire introduction position, speed, and guide angle prediction specifically include: Assume that the state quantity and control quantity of the guide wire are expressed as follows: ,in, ,in, is the horizontal coordinate of the guide wire position, is the ordinate of the guidewire position, is the guide wire angle, is the guidewire advancement speed; The set value of the output state is obtained through the prediction model, that is, the state quantity and control quantity of the next state are obtained, including the guide wire introduction position, speed, and guide angle; The input of the prediction model includes the vascular state index value, the current guidewire state value and the control value. , the output of the prediction model is the state quantity and control quantity of the next state; A control module is used to control the vascular interventional surgery robot based on the corrected PID parameters.

2. The vascular interventional surgery robot control system according to claim 1, characterized in that: Image enhancement includes the following steps: a. Use mean filtering to estimate the noise of the original image, and the estimated value is ; b. Select a window with a length and width of win and divide the image into blocks; in, , where k is the adjustment coefficient, , ; is the standard deviation of the image; are the width and height of the original image respectively; c. Filter the sub-blocks that have not overflowed after adding noise; d. Calculate the standard deviation of the sub-blocks that have not overflowed, and sort the standard deviations from small to large; e. Calculation of statistical parameters ; f. Calculate the noise standard deviation estimate: , where MN is the total number of sub-blocks, is the number of overflow sub-blocks; g. De-noise the initial image according to the noise standard deviation estimate calculated in step f.

3. The vascular interventional surgery robot control system according to claim 1, characterized in that: Image correction after segmentation refers to correcting the hollowed-out area, which specifically includes the following steps: a. Copy the segmented image to generate a new image to fill the hollowed-out area; b. Create a zero matrix with an aspect ratio n greater than the original input image as a mask based on the image size; c. Traverse the pixel values ​​of the image and find the first point with a pixel value of 0 as the starting point for filling. Then use the filling function of OpenCV to fill it. The result is a 255-filled image. d. Invert the image to get the inverse image, then perform a bitwise OR operation on the original image and the inverse image to get the image with the holes filled in.

4. The vascular interventional surgery robot control system according to claim 1, characterized in that: Other patient information included age, gender, BMI index, blood pressure, total cholesterol, fasting blood glucose, cholesterol, triglycerides, and thyroid-stimulating hormone.

5. The vascular interventional surgery robot control system according to claim 4, characterized in that: The control of the control parameters based on the feedback learning network algorithm includes: a. Initialize the feedback learning network parameters, set the network related parameters and PID parameter initial values, and set the PID parameter range; b. Calculate the current reward function value and put the reward function value and current state information into the buffer area; The reward function is expressed as follows: , , , Reward for safety, is the damage factor, col is the impact strength value, and its value comes from the force sensor at the front end of the guide wire; Reward for stability, is the adjustment factor, a is the acceleration, is the performance value, which is the mean value of motion control deviation; c. Calculate the objective function value and update the network parameters according to the direction of the optimal objective function to complete the network iterative learning; The objective function is as follows: ,in, is the motion error, t is the time, is the time of a sampling cycle; d. Output PID parameters to the PID controller, perform a wire movement, calculate the performance value of this movement process, and input it into the network; e. Determine whether the performance of this motion process meets the control performance requirements. If so, output the PID parameters and end the iterative learning. Otherwise, repeat steps b~d.

6. The vascular intervention surgery robot control system according to claim 5, characterized in that: The correction of the control parameters based on the gravity of the guidewire catheter and the friction coefficient of the blood vessel includes: obtaining the total mass of the guidewire catheter, calculating the total gravity ,in, is the mass per unit length of the guide wire, is the mass per unit length of the catheter; b: Calculate the vascular friction coefficient, specifically, calculate the vascular friction coefficient based on the vascular status index value , where K is the vascular status index value, r is the adjustment coefficient, which is greater than 1 and is determined according to the severity of vascular arterial stenosis and occlusion; c. Calculate the corrected PID parameters , , , is the PID parameter adjustment amount, , , is the PID parameter before correction.

Images