Control method of intelligent surgical robot

Abstract

The invention relates to the field of robot control, in particular to a control method of an intelligent surgical robot. The method comprises the following steps: acquiring preoperative multi-modal medical image data of a patient, and preprocessing to obtain preprocessed medical image data; based on the preprocessed medical image data, generating a three-dimensional anatomical model through a multi-modal medical image registration and segmentation algorithm; based on the three-dimensional anatomical model, a real-time optimized operation path is generated through a multi-scale dynamic disturbance optimization reinforcement learning algorithm; and based on the operation path optimized in real time, the robot is controlled. The technical problems that in control of the intelligent surgical robot, a traditional robot control method is inaccurate in medical image data processing and analysis and poor in surgical path planning flexibility are solved.

Description

Technical Field

[0001] The present invention relates to the field of robot control, and in particular to a control method for an intelligent surgical robot. Background Art

[0002] With the continuous advancement of medical technology, especially the rapid development of robot-assisted surgery, surgical robots are playing an increasingly important role in various types of complex surgeries. Surgical robots can reduce human errors and surgical trauma through highly precise operations, and improve the success rate of surgery and the speed of patient recovery. Especially in some delicate and high-precision surgeries, such as neurosurgery, ophthalmology, spinal surgery, etc., surgical robots have shown irreplaceable advantages. However, during the implementation process, surgical robots face huge challenges in how to accurately navigate and optimize paths in real time.

[0003] At present, surgical robot systems mainly rely on preoperative imaging data to generate a three-dimensional model, and perform path planning and surgical navigation based on the model. Traditional navigation methods mostly rely on a single image data source, such as CT or MRI. However, due to the limitations of different imaging technologies, the anatomical information provided by a single image source is often difficult to fully cover the surgical area, resulting in insufficient accuracy in path planning. In addition, dynamic changes during surgery and individual differences among patients also greatly affect the accuracy of robot navigation.

[0004] The above-mentioned technology has technical problems in the control of intelligent surgical robots, such as inaccurate processing and analysis of medical image data and poor flexibility in surgical path planning. Summary of the invention

[0005] The present invention provides a control method for an intelligent surgical robot to solve the technical problems of inaccurate processing and analysis of medical image data and poor flexibility in surgical path planning in the control of the intelligent surgical robot by traditional robot control methods.

[0006] A control method of an intelligent surgical robot of the present invention specifically includes the following technical solutions: A control method for an intelligent surgical robot comprises the following steps: S1. Obtain the patient's preoperative multimodal medical imaging data and perform preprocessing to obtain preprocessed medical imaging data; based on the preprocessed medical imaging data, introduce an adaptive transformation model based on morphological gradient consistency through a multimodal medical imaging registration and segmentation algorithm, optimize the registration accuracy of multimodal medical imaging, establish a morphologically consistent mapping relationship, and generate a three-dimensional anatomical model; S2. Based on the three-dimensional anatomical model, a real-time optimized surgical path is generated through multi-scale dynamic perturbation optimization reinforcement learning algorithm; based on the real-time optimized surgical path, the robot is controlled.

[0007] Preferably, the S1 specifically includes: In the process of implementing the multimodal medical image registration and segmentation algorithm, the preprocessed medical image data is subjected to preliminary feature mapping, and feature analysis is performed jointly in the spatial domain and the frequency domain to obtain feature data of the preprocessed medical image data.

[0008] Preferably, the S1 specifically includes: Based on the feature data of preprocessed medical image data, an adaptive transformation model based on morphological gradient consistency is introduced to establish a morphologically consistent mapping relationship. A regularized loss function is constructed in the registration process. By minimizing the gradient deviation between the modalities, the registration accuracy of multimodal medical images is optimized to obtain the registered medical image data.

[0009] Preferably, the S1 specifically includes: Based on the registered medical image data, a weighted feature fusion strategy is introduced; the fusion weight is calculated by the signal-to-noise ratio and entropy value of the registered medical image data; based on the fusion weight, multimodal fusion is performed to obtain fused image data; high-order morphological processing is performed on the fused image data to obtain morphologically processed fused image data; based on the morphologically processed fused image data, three-dimensional reconstruction is performed through voxel interpolation and morphological restoration methods to obtain a three-dimensional anatomical model.

[0010] Preferably, the S2 specifically includes: In the process of implementing the multi-scale dynamic disturbance optimization reinforcement learning algorithm, a multi-scale state description model is constructed to obtain a state vector; the state vector includes the robot's current posture matrix, force feedback matrix, velocity vector, target distance and path parameters.

[0011] Preferably, the S2 specifically includes: In the process of implementing the multi-scale dynamic perturbation optimization reinforcement learning algorithm, a multi-scale dynamic perturbation optimization mechanism is introduced, and the perturbation function is constructed by combining the global perturbation term and the local perturbation term to obtain the perturbation output result; the specific formula of the perturbation function is as follows: , in, It's at the time The output result of the disturbance; Is the surgical robot at the time point The state vector of is the number of disturbances; is the amplitude coefficient, indicating the The intensity of the disturbance; It is The disturbance frequency factor of the disturbance; It is The phase offset of the disturbance; It is The local perturbation weight of the perturbation; It is The rate of change of disturbance of a disturbance; is the global disturbance term; is a local disturbance term.

[0012] Preferably, the S2 specifically includes: The output result of the disturbance is used as the input of the path decision strategy adjustment, and based on the state vector and the update step of the current path decision strategy, a dynamic strategy adjustment process is constructed to update the strategy and determine the new path selection.

[0013] Preferably, the S2 specifically includes: In the implementation process of the multi-scale dynamic perturbation optimization reinforcement learning algorithm, a reward function is constructed based on the posture matrix, force feedback matrix, velocity vector, target distance and path parameters in the state vector; the specific formula of the reward function is as follows: , in, It's at the time The reward function of is the reward coefficient; is the robot’s current posture matrix; is the target location; is the robot’s current force feedback matrix; is the safety force threshold; is the current speed of the robot; is the spatial distance between the robot and the target; It is path parameters; is the partial derivative of the target distance with respect to the path parameter; is the number of path parameters; is the attenuation factor; is the time limit; the first item Measures the Euclidean distance between the current position and the target position. The second term Measuring the safety of force feedback during surgery, the third The fourth item suppresses excessive speed changes by using a logical decay function. Measures the path smoothness, that is, the second-order derivative of the target distance with respect to the path parameter, the fifth term is the time decay integral.

[0014] Preferably, the S2 specifically includes: Combining the perturbation function and the reward function, the surgical path is iteratively updated to obtain a real-time optimized surgical path. The real-time optimized surgical path is converted into the robot's motion instructions, and the intelligent surgical robot is controlled through an error-based control algorithm.

[0015] The beneficial effects of the technical solution of the present invention are: 1. In the process of fusing pre-processed medical image data by a multimodal medical image registration and segmentation algorithm, the present invention introduces an adaptive transformation model based on morphological gradient consistency to optimize the accuracy of multimodal medical image registration, and solves the registration error between multimodal medical images by minimizing the gradient deviation; the fusion weight is calculated based on the signal-to-noise ratio and entropy value of the registered medical image data to obtain fused image data, and the contrast and boundary sharpness of the fused image data are improved through high-order morphological processing to obtain fused image data after morphological processing; the fused image data after morphological processing is three-dimensionally reconstructed, and a fine three-dimensional anatomical model is generated using Bessel interpolation and morphological restoration methods so as to accurately reflect the morphological characteristics of organs and tissues.

[0016] 2. Based on the three-dimensional anatomical model, combined with the multi-scale dynamic perturbation optimization reinforcement learning algorithm, the perturbation function is introduced, and the local and overall perturbation patterns of the path are combined to fine-tune the path to ensure accuracy and safety during the operation, generate a real-time optimized surgical path, and implement precise control through the robot to ensure that the robot can operate stably in a dynamically changing surgical environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a control method for an intelligent surgical robot described in the present invention. DETAILED DESCRIPTION

[0018] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0020] The specific scheme of the control method of an intelligent surgical robot provided by the present invention is described in detail below with reference to the accompanying drawings.

[0021] See attached Figure 1 , which shows a flow chart of a control method of an intelligent surgical robot provided by an embodiment of the present invention, the method comprising the following steps: S1. Obtain the patient's preoperative multimodal medical imaging data and perform preprocessing to obtain preprocessed medical imaging data; based on the preprocessed medical imaging data, introduce an adaptive transformation model based on morphological gradient consistency through a multimodal medical imaging registration and segmentation algorithm, optimize the registration accuracy of multimodal medical imaging, establish a morphologically consistent mapping relationship, and generate a three-dimensional anatomical model; Obtaining the patient's preoperative multimodal medical image data (such as CT, MRI, ultrasound, etc.) through the hospital PACS (Picture Archiving and Communication System) interface, and performing preprocessing such as format conversion, image coordinate system normalization, denoising, image enhancement, and artifact removal on the patient's preoperative multimodal medical image data to obtain preprocessed medical image data. The preprocessing process adopts technical means well known to those skilled in the art, which will not be described in detail here; Furthermore, the pre-processed medical image data is fused through multimodal medical image registration and segmentation algorithms to generate a three-dimensional anatomical model; The specific implementation process of the multimodal medical image registration and segmentation algorithm is as follows: First, feature extraction is performed on the preprocessed medical image data, and the edge, texture, and density information of the preprocessed medical image data are comprehensively analyzed by using the method of fusing local features with global features. , perform preliminary feature mapping, and perform feature analysis in the spatial domain and frequency domain to obtain the feature data of the preprocessed medical image data. The specific formula is as follows: , in, Indicates The preprocessed medical image data is at location Feature data at is the image grid coordinate, which indicates the spatial position of the image and is used to determine the local area for feature calculation; is the number of feature channels, using multiple channels to extract rich feature information; It is The local feature weight distribution of each feature channel represents the contribution of different features and is used to perform weighted combination of features in the local space to adjust the influence of features such as edges and textures. It is obtained through experimental methods. , , Respectively indicated at the position At, along The degree of grayscale change in the direction, that is, the gradient, is used to reflect edge information to enhance the boundary characteristics of the image. The gradient is calculated by the Roberts operator; It is The preprocessed medical image data is at location The gray level at It is The frequency domain feature weight of each feature channel is used to control the influence of the frequency domain feature on the overall feature, so as to adjust the balance of noise and texture information in the preprocessed medical image data. It is an adjustment parameter set according to the image spectrum characteristics. Represents the Fourier transform result, which is used to capture frequency domain information and extract frequency domain features of preprocessed medical image data to describe the periodicity and global structure of preprocessed medical image data, so as to supplement the deficiency of spatial domain features. The purpose of feature extraction is to provide a consistent representation for subsequent fusion, so that data of different modalities can be registered in the same feature space; Based on the feature data of preprocessed medical image data, an adaptive transformation model based on morphological gradient consistency is introduced to optimize the registration accuracy of multimodal medical images. Its goal is to establish a morphologically consistent mapping relationship by minimizing the gradient deviation between modalities. The regularization loss function in the registration process is defined as follows: , in, It is the regularized loss function in the registration process, which represents the difference between the medical image data after the registration transformation and the reference image data, and also contains a smoothing term (gradient term) to ensure the smoothness and continuity of the registration transformation; It is the domain of the entire image, which represents the location set of all voxels in the preprocessed medical imaging data; is the first Medical image data in voxels (positions) The characteristic data at , Reference image data In voxel The feature data at the position is used as the registration target to guide the pre-processed medical image data at the position Feature data at Transform it to align it with the reference image data; It is a minimization operation, which means finding a solution that can minimize the regularization loss function among all possible registration transformations, that is, finding the optimal registration transformed medical image data to minimize the error between the medical image feature data; It is the reference image data, determined by expert experience method; is a regularization parameter used to suppress artifacts caused by excessive deformation; is the first Medical image data in voxels (positions) The gradient at . The regularized loss function is solved by the existing steepest descent method to obtain the registered medical image data ; Furthermore, after the registration is completed, the multimodal fusion stage is entered, and a weighted feature fusion strategy is adopted to calculate the fusion weight through the signal-to-noise ratio (SNR) and entropy value; based on the fusion weight, multimodal fusion is performed to obtain fused image data, and the formula is as follows: , , in, is in position Fusion image data at It is The registered medical image data is at position The characteristic data of is the total number of registered medical image data; It is The weight of the registered medical image data, that is, the fusion weight; It is The signal-to-noise ratio of the registered medical image data; It is The signal-to-noise ratio of the registered medical image data; It is a regulating factor used to balance the influence of signal-to-noise ratio and entropy value on weight, and is determined according to expert experience; It is The entropy value of the registered medical image data; It is The entropy value of the registered medical image data. The purpose of multimodal fusion is to integrate the advantages of multimodal images and improve the contrast and detail retention of anatomical structures. Based on the fused image data, high-order morphological processing is used to remove noise and improve the edge sharpness of the fused image data to obtain the fused image data after morphological processing; Based on the fused image data after image processing, the existing voxel interpolation and morphological restoration methods are used to perform three-dimensional reconstruction to obtain a three-dimensional anatomical model. The specific implementation process is as follows: first, based on the interpolation weight, the continuous voxel space is interpolated. The interpolation weight is obtained based on the Bessel interpolation formula. The purpose of interpolation is to generate a more refined voxel expression to improve the spatial accuracy of the three-dimensional anatomical model; then, morphological restoration is performed to eliminate artifacts and defects that may occur in the three-dimensional reconstruction process, and the voxel density function of the restored three-dimensional anatomical model is obtained. ; Further generate isosurfaces through MarchingCubes algorithm to form a visual 3D anatomical model : , in, is the tissue density threshold, which is set according to expert experience; finally, a three-dimensional anatomical model is obtained, which reflects the morphological characteristics of organs and tissues.

[0022] S2. Generate a real-time optimized surgical path based on a three-dimensional anatomical model through multi-scale dynamic perturbation optimization reinforcement learning algorithm; and control the robot based on the real-time optimized surgical path.

[0023] Based on the three-dimensional anatomical model, a multi-scale dynamic perturbation optimization reinforcement learning algorithm is introduced to generate a real-time optimized surgical path, and the robot is controlled based on the real-time optimized surgical path; the specific implementation process of the multi-scale dynamic perturbation optimization reinforcement learning algorithm is as follows: First, the surgical environment is modeled; the surgical environment is composed of multiple complex physiological organizations and constraints. In order to achieve optimal path planning while ensuring surgical safety, it is necessary to construct a multi-scale state description model to model the surgical environment. The multi-scale state description model comprehensively considers factors such as the robot's posture in space, force feedback, speed, and target distance to construct a state vector; the surgical robot is located at any point in time. The state is represented as the state vector , which consists of the robot's current posture matrix , force feedback matrix , velocity vector , Target distance And the path parameters ; The mathematical expression of the state vector is: , In the process of path planning, the non-deterministic characteristics of the surgical environment will affect the robot's motion stability. Therefore, a multi-scale dynamic perturbation optimization mechanism is introduced to combine multiple perturbation modes and construct a perturbation function to fine-tune the path at different scales, ensuring that the robot can make local and global joint adjustments when responding to dynamic changes in physiological tissues. The specific formula of the perturbation function is as follows: , in, It's at the time The perturbation output describes the dynamic perturbation of the surgical robot during the path planning process, which is used to represent the impact of minor adjustments during the surgery and changes in the surgical environment; is the number of disturbances; is the amplitude coefficient, indicating the The intensity of the disturbance is used to control the magnitude of the disturbance. It is obtained through experimental fitting and has a value range of ; It is The disturbance frequency factor of the disturbance determines the oscillation frequency of the disturbance and affects the sensitivity of path planning. It is obtained according to the expert experience method and has a value range of ; It is The phase offset of the disturbance is used to adjust the phase offset of the sine function to ensure the alignment of the disturbance with the state of the intelligent surgical robot. It is obtained according to the expert experience method and has a value range of ; It is The local perturbation weight of the perturbation is used to control the contribution of the local perturbation term to adjust the accuracy of the local refinement of the path. It is obtained by experimental method and has a value range of ; It is The disturbance change rate of a disturbance is used to control the response rate of the disturbance function to the state change of the intelligent surgical robot, affecting the adjustment frequency of the local path. It is determined according to the expert experience method and the response speed requirements in different surgical scenarios. The value range is ; is a global perturbation term, which is used to provide global path perturbation to break through the local optimal trap; is a local disturbance term (logistic regression function) used to adjust the refined path of the intelligent surgical robot in the micro-surgical area to ensure that tissue damage is avoided; Furthermore, the output result of the disturbance is used as the input for the subsequent path decision strategy adjustment, providing dynamic disturbance input for path optimization. In the path search process, the path decision strategy needs to be continuously adjusted to balance the relationship between exploration and utilization. To this end, a dynamic strategy adjustment process is constructed based on the state vector and the update step size of the current path decision strategy. , update the strategy and decide the new path selection. The expression of strategy update is: , in, It indicates that the intelligent surgical robot is in the state Next select action Specifically, is a policy function that gives the Next, select different actions probability; Indicates the intelligent surgical robot at the time point The selected action is a control signal or an operation instruction of the robot, such as how the robot should adjust its position, posture, speed or perform a certain operation step; Indicates at a point in time The strategy update step in is used to control the speed of path decision strategy adjustment. The value will lead to a larger policy update, which will cause the path decision strategy to change quickly, and vice versa, it will be a smaller update, which is determined by experimental method; the left side of the arrow It means that after the strategy is modified, the intelligent surgical robot is in state Next select action The probability of It indicates that the intelligent surgical robot is in the state of Next select action The probability of Indicates that based on the current state and strategy parameters The calculated policy correction function, which is used to adjust the current path decision strategy to make the robot more effective in performing tasks, is derived through machine learning algorithms (e.g., deep reinforcement learning), such as , is the total number of dimensions of strategy revision; It is in force feedback matrix The weight parameters related to the square root of are determined according to the expert experience method; Indicated in The strategy parameters in the dimension can be weights in neural networks or adjustable parameters of other machine learning models, which are used to control how the path decision strategy is in the current state. Select actions and optimize the robot's behavior by adjusting these parameters; Furthermore, multiple factors such as target distance, force feedback, speed and path smoothness are combined to construct a reward function that adapts to complex physiological environments to ensure the safety and optimality of the path and to ensure that the intelligent surgical robot achieves path optimization in the shortest time; the specific formula of the reward function is as follows: , in, is a reward function, which indicates the quality of the robot's current path selection. It is an objective function used to optimize path selection to ensure that the robot can balance safety, efficiency, and path smoothness during operation. is the reward coefficient, which is used to balance the contribution of different evaluation items to the reward function. The size of different coefficients indicates the influence of different goals on the final path selection, which is determined according to the expert experience method; the first item Weighing current position against target position The Euclidean distance of Measuring the safety of force feedback during surgery, the third The fourth item suppresses excessive speed changes by using a logical decay function. Measures the path smoothness, that is, the second-order derivative of the target distance with respect to the path parameter, the fifth term It is the time decay integral, which is used to ensure that the path planning process gradually becomes stable as the operation time increases; is the robot's current posture matrix, which represents the robot's position and posture (rotation and displacement) in three-dimensional space; is the target position, which indicates the predetermined target position or target area to be reached by the surgical robot and is set by expert experience method; is the current force feedback matrix of the robot, which represents the interaction force between the robot and the surgical target; is the safety force threshold, which indicates the safety force value when the robot contacts the tissue during surgery, and is set according to the safety standard when the robot is designed; is the current speed of the robot, indicating the speed of the robot when it moves; It is the form of the logistic decay function of speed, which is used to control the sensitivity of the reward function to speed through a smooth decay function; is the position or distance on the target path, indicating the spatial distance between the robot and the target; It is a path parameter, which indicates a specific parameter in the path; is the partial derivative of the target distance with respect to the path parameter, which indicates the sensitivity of the target distance to changes in the path parameter; is the number of path parameters, which indicates the number of all parameters that may be involved in the path; is the decay factor, which controls the rate of the time decay function and determines how quickly the reward decays over time; is the time upper limit, which indicates the maximum time range considered in the path planning process; It is the integral of the time decay function, indicating that as time goes by, the influence of time on path optimization in the path planning process gradually weakens; Finally, the perturbation function and reward feedback are combined to achieve iterative update of the surgical path, which is mathematically expressed as follows: , in, is the new path parameter, i.e. the real-time optimized surgical path, which means the path parameter after one optimization step, and refers to the path parameter of the robot, which may include the position, posture, speed, acceleration, etc. of the robot; It represents the current path parameter, that is, the path that the robot is executing at the current moment; is the path update step size, which indicates the update amplitude of the path parameters during the optimization process. It is used to control the speed and stability of the path update and is set according to the expert experience method; is the gradient of the reward function, which means the reward function Relative to the state vector The gradient of reflects the optimization direction and magnitude of the path, and determines how the robot should adjust its path during the optimization process to improve the reward value during execution; is the local derivative of the reward function with respect to the state vector, which represents the reward function At the present moment For the state vector The rate of change of is the perturbation function; is the number of accumulations, indicating the number of accumulation calculations required during the path optimization process, which represents the reward function The total number of items; Through the above process, the intelligent surgical robot can make dynamic adjustments in a multi-scale environment during path planning, and combine multi-factor disturbances and reward feedback to achieve accurate planning of the surgical path. During the path exploration process, it can make full use of global information for large-scale search, and combine real-time feedback for local corrections to form an adaptive path optimization solution to cope with complex surgical environments.

[0024] Finally, the real-time optimized surgical path is converted into the robot's motion instructions, and the intelligent surgical robot is controlled using an error-based control algorithm such as a proportional-integral-derivative (PID) controller or a more complex robust control algorithm.

[0025] In summary, a control method for an intelligent surgical robot is completed.

[0026] The order of the embodiments of the invention is for description only and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0027] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0028] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A control method for an intelligent surgical robot, characterized in that: The following steps are involved: S1. Obtain the patient's preoperative multimodal medical imaging data and perform preprocessing to obtain preprocessed medical imaging data; based on the preprocessed medical imaging data, introduce an adaptive transformation model based on morphological gradient consistency through a multimodal medical imaging registration and segmentation algorithm, optimize the registration accuracy of multimodal medical imaging, establish a morphologically consistent mapping relationship, and generate a three-dimensional anatomical model; S2. Based on the three-dimensional anatomical model, a real-time optimized surgical path is generated through multi-scale dynamic perturbation optimization reinforcement learning algorithm; based on the real-time optimized surgical path, the robot is controlled.

2. The control method of an intelligent surgical robot according to claim 1, characterized in that: The S1 specifically includes: In the process of implementing the multimodal medical image registration and segmentation algorithm, the preprocessed medical image data is subjected to preliminary feature mapping, and feature analysis is performed jointly in the spatial domain and the frequency domain to obtain feature data of the preprocessed medical image data.

3. The control method of an intelligent surgical robot according to claim 2, characterized in that: The S1 specifically includes: Based on the feature data of preprocessed medical image data, an adaptive transformation model based on morphological gradient consistency is introduced to establish a morphologically consistent mapping relationship. A regularized loss function is constructed in the registration process. By minimizing the gradient deviation between the modalities, the registration accuracy of multimodal medical images is optimized to obtain the registered medical image data.

4. The control method of an intelligent surgical robot according to claim 3, characterized in that: The S1 specifically includes: Based on the registered medical image data, a weighted feature fusion strategy is introduced; the fusion weight is calculated by the signal-to-noise ratio and entropy value of the registered medical image data; based on the fusion weight, multimodal fusion is performed to obtain fused image data; high-order morphological processing is performed on the fused image data to obtain morphologically processed fused image data; based on the morphologically processed fused image data, three-dimensional reconstruction is performed through voxel interpolation and morphological restoration methods to obtain a three-dimensional anatomical model.

5. The control method of an intelligent surgical robot according to claim 1, characterized in that: The S2 specifically includes: In the process of implementing the multi-scale dynamic disturbance optimization reinforcement learning algorithm, a multi-scale state description model is constructed to obtain a state vector; the state vector includes the robot's current posture matrix, force feedback matrix, velocity vector, target distance and path parameters.

6. The control method of an intelligent surgical robot according to claim 5, characterized in that: The S2 specifically includes: In the process of implementing the multi-scale dynamic perturbation optimization reinforcement learning algorithm, a multi-scale dynamic perturbation optimization mechanism is introduced, and the perturbation function is constructed by combining the global perturbation term and the local perturbation term to obtain the perturbation output result; the specific formula of the perturbation function is as follows: ， in, It's at the time The output result of the disturbance; Is the surgical robot at the time point The state vector of is the number of disturbances; is the amplitude coefficient, indicating the The intensity of the disturbance; It is The disturbance frequency factor of the disturbance; It is The phase offset of the disturbance; It is The local perturbation weight of the perturbation; It is The rate of change of disturbance of a disturbance; is the global disturbance term; is a local disturbance term.

7. The control method of an intelligent surgical robot according to claim 6, characterized in that: The S2 specifically includes: The output result of the disturbance is used as the input of the path decision strategy adjustment, and based on the state vector and the update step of the current path decision strategy, a dynamic strategy adjustment process is constructed to update the strategy and determine the new path selection.

8. The control method of an intelligent surgical robot according to claim 7, characterized in that: The S2 specifically includes: In the implementation process of the multi-scale dynamic perturbation optimization reinforcement learning algorithm, a reward function is constructed based on the posture matrix, force feedback matrix, velocity vector, target distance and path parameters in the state vector; the specific formula of the reward function is as follows: ， in, It's at the time The reward function of is the reward coefficient; is the robot’s current posture matrix; is the target location; is the robot’s current force feedback matrix; is the safety force threshold; is the current speed of the robot; is the spatial distance between the robot and the target; It is path parameters; is the partial derivative of the target distance with respect to the path parameter; is the number of path parameters; is the attenuation factor; is the time limit; the first item Measures the Euclidean distance between the current position and the target position. The second term Measuring the safety of force feedback during surgery, the third The fourth item suppresses excessive speed changes by using a logical decay function. Measures path smoothness, the fifth term is the time decay integral.

9. The control method of an intelligent surgical robot according to claim 8, characterized in that: The S2 specifically includes: Combining the perturbation function and the reward function, the surgical path is iteratively updated to obtain a real-time optimized surgical path. The real-time optimized surgical path is converted into the robot's motion instructions, and the intelligent surgical robot is controlled through an error-based control algorithm.

Images