Automatic laser cutting method and system for skull repair

Abstract

The invention discloses an automatic laser cutting method and system for skull repair, and the method comprises the steps: firstly obtaining the medical image data of a skull defect area of a patient, carrying out the three-dimensional modeling, then matching with a preset repair implant standard parameter library, determining the shape, size and edge contour of a repair implant, and carrying out the laser cutting of the repair implant. A laser cutting path is generated through a path planning algorithm and optimized, a laser head is driven to automatically cut the repairing material, and the cutting process is monitored in real time; the system comprises a medical image processing module, a three-dimensional modeling module, a path planning module, a laser cutting execution module, a real-time monitoring module and a postoperative evaluation module. Compared with the prior art, the method has the advantages that the medical image data processing precision is improved; personalized design is realized; optimizing a laser cutting path; and the real-time monitoring and adjusting capability is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of medical devices, and specifically refers to an automated laser cutting method and system for skull repair. Background Art

[0002] Skull defect is one of the common diseases in neurosurgery, usually caused by craniocerebral injury, brain tumor resection surgery, etc. Skull repair is an important means to repair skull defects, aiming to restore the integrity of the skull and protect the normal function of the brain tissue. Traditional skull repair methods include manual shaping titanium mesh implantation, autologous bone transplantation, etc. These methods have disadvantages such as complex operation, insufficient precision, and low personalization degree.

[0003] With the continuous progress of medical technology, laser cutting technology has received extensive attention in the field of skull repair due to its high precision, high efficiency, and repeatability. However, there are still many deficiencies in the application of existing laser cutting methods in skull repair. For example, Patent CN105146671A proposes a method for designing a skull defect repair body based on CT images. Although personalized design is achieved, there is a lack of detailed technical solutions in aspects such as laser cutting path planning, cutting parameter optimization, and real-time monitoring and adjustment. In addition, although Patent CN107411654A provides a laser cutting device, it has not been adaptively improved according to the specific needs of skull repair, resulting in the need to further improve cutting precision and efficiency. Summary of the Invention

[0004] The technical problems to be solved by the present invention are: (1) Insufficient precision in medical image data processing: During the processing of medical image data in the skull defect area, the precision of steps such as noise suppression and edge detection directly affects the accuracy of three-dimensional modeling, and thus affects the precision of subsequent laser cutting; (2) Lack of personalization in the design of repair implants: Existing methods usually design implants based on a standard parameter library, but lack fine adjustment for individual patient differences, resulting in a low matching degree between the implant and the defect area; (3) Unreasonable laser cutting path planning: The planning of the cutting path directly affects cutting efficiency and precision. Existing methods often neglect the optimization of the cutting path, resulting in unnecessary material waste and heat accumulation during the cutting process; (4) Insufficient real-time monitoring and adjustment ability: During the laser cutting process, the real-time monitoring and adjustment of parameters such as temperature, deformation, and positioning precision are crucial for ensuring cutting quality, but existing methods often lack effective real-time monitoring and dynamic adjustment mechanisms.

[0005] To solve the above technical problems, the technical solution provided by the present invention is: An automated laser cutting method for skull repair, including:

[0006] S1. Obtain the medical image data of the patient's skull defect area, perform 3D modeling, and generate a 3D reconstruction model of the defect site;

[0007] S2. According to the 3D reconstruction model, match the preset standard parameter library of repair implants, and determine the shape, size, and edge contour of the repair implants;

[0008] S3. Based on the edge contour, generate a laser cutting path through a path planning algorithm and optimize the cutting parameters;

[0009] S4. Input the optimized cutting path and parameters into the laser cutting device, and drive the laser head to automatically cut the repair material;

[0010] S5. Real-time monitor the temperature, deformation, and positioning accuracy during the cutting process, and dynamically adjust the laser power, cutting speed, and focusing position.

[0011] Preferably, S1 further includes:

[0012] S1.1. Use a CT scanner to obtain the tomographic scan data of the patient's skull area, and output the image data as a sequence of DICOM format grayscale images;

[0013] S1.2. Use the Perona-Malik diffusion equation to suppress the noise of the original DICOM sequence, separate the skull and soft tissues based on the HU value, select the skull threshold segmentation, and apply the Sobel edge detection operator to the segmented binary image;

[0014] S1.3. Perform radial basis function interpolation on the discrete edge point set P = {p i (x i , y i , z i )} of the defect area, convert the interpolated volume data into a triangular mesh model, and perform Laplacian smoothing on the reconstructed mesh;

[0015] S1.4. Output the 3D reconstruction model and store it in the STL format.

[0016] Preferably, S2 further includes:

[0017] S2.1. Input the 3D reconstruction model, and use the Gaussian filtering algorithm to smooth the 3D model to remove noise and irrelevant details;

[0018] S2.2. Input the smoothed 3D model, and use the Canny edge detection algorithm to extract the edge features of the skull defect area;

[0019] S2.3. Use the DTW dynamic deformation algorithm for the extracted edge features and the standard parameter library of repair implants to match the biomechanical adaptation parameters in the standard parameter library;

[0020] S2.4. Input the matched parameters and edge contours, and use the least squares method to fit and determine the shape and size of the repair implant;

[0021] S2.5. Input the determined shape and size, and use B-spline curve fitting to generate the edge contour of the repair implant.

[0022] Preferably, S3 further includes:

[0023] S3.1. Input the edge contour data of the repair implant, and use the parametric discretization method to convert the edge contour data into a discrete point set;

[0024] S3.2. Input the discretized edge contour point set, and use the path planning algorithm to generate the laser cutting path. The path planning algorithm includes the progressive ant colony optimization algorithm and genetic algorithm;

[0025] S3.3. Input the generated laser cutting path, and optimize the laser cutting parameters. The laser cutting parameters include laser power, cutting speed, slit width, and heat affected zone range;

[0026] S3.4. Output the optimized laser cutting path and cutting parameters, input the path and parameters into the laser cutting equipment, and prepare to execute the cutting operation.

[0027] Preferably, S5 further includes:

[0028] S5.1. Input the infrared temperature sensor data T raw (t), the visual positioning system data Δx(t), Δy(t), and the deformation detector data ∈(t), and use the Kalman filter to remove the sensor noise and improve the signal-to-noise ratio. Among them, T raw (t) represents the temperature of the cutting area, Δx(t), Δy(t) represent the lateral / longitudinal deviation of the cutting trajectory from the preset path, and ∈(t) represents the local strain or curvature change of the material;

[0029] S5.2. Calculate the deviation between the actual temperature and the safety threshold T threshold ; calculate the Euclidean distance deviation between the cutting trajectory and the preset path; calculate the material deformation rate according to the strain data ∈(t);

[0030] S5.3. Synthesize the temperature, positioning, and deformation errors, construct a dynamic weight objective function, and set the parameter adjustment rules for the laser power P(t), cutting speed v(t), and focus position f(t);

[0031] S5.4. Input the adjusted parameters P(t), v(t), f(t) into the laser cutting equipment, and at the same time update the sensor data in real time to form a closed-loop control; dynamically optimize the weight coefficient according to the historical error data to avoid local optimum.

[0032] Preferably, an automated laser cutting system for implementing the method according to any one of claims 1-5 is further provided, including:

[0033] A medical image processing module for acquiring and processing medical image data of the skull defect area;

[0034] A three-dimensional modeling module connected to the image processing module to generate a three-dimensional model of the defect site;

[0035] A path planning module for outputting an optimized laser cutting path and parameters based on the three-dimensional model;

[0036] A laser cutting execution module, including a high-precision laser generator, a multi-axis robotic arm and a material fixing platform, for performing cutting operations; an adaptive fixture is provided at the end of the multi-axis robotic arm for adjusting the laser focal length according to the thickness of the repair material;

[0037] A real-time monitoring module, including a temperature sensor, a vision positioning unit and a deformation detector, for dynamically adjusting the cutting process; the real-time monitoring module is connected to the path planning module through a closed-loop feedback mechanism to realize adaptive iterative optimization of cutting parameters;

[0038] A postoperative evaluation module for performing three-dimensional scanning comparison between the cut implant and the three-dimensional model and outputting a matching degree report.

[0039] Preferably, a computer-readable storage medium is further provided, and the storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1-5 is implemented.

[0040] Preferably, an electronic device is further provided, including a memory, a processor and a computer program stored on the memory and operable on the processor, and when the processor executes the program, the method according to any one of claims 1-5 is implemented.

[0041] The advantages of the present invention compared with the prior art are as follows: (1) Improving the accuracy of medical image data processing: By adopting advanced image processing algorithms such as the Perona-Malik diffusion equation and the Sobel edge detection operator, the processing accuracy of medical image data in the skull defect area is improved, providing an accurate data basis for 3D modeling; (2) Realizing personalized design: By matching the preset standard parameter library of repair implants and making fine adjustments according to the individual differences of patients, the personalized design of repair implants is realized, improving the matching degree between the implant and the defect area; (3) Optimizing the laser cutting path: By adopting progressive ant colony optimization algorithms and genetic algorithms for path planning and optimizing cutting parameters, the efficiency and accuracy of laser cutting are improved, reducing material waste and heat accumulation; (4) Enhancing the real-time monitoring and adjustment ability: By real-time monitoring parameters such as temperature, deformation and positioning accuracy during the cutting process and dynamically adjusting cutting parameters, precise control of the cutting process is achieved, improving the cutting quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 FIG. is a schematic diagram of the steps of an automated laser cutting method for skull repair.

[0043] Figure 2 FIG. is a schematic diagram of the steps of S1.

[0044] Figure 3 FIG. is a schematic diagram of the steps of S2.

[0045] Figure 4 FIG. is a schematic diagram of the steps of S3.

[0046] Figure 5 FIG. is a schematic diagram of the steps of S5.

[0047] Figure 6 FIG. is a schematic diagram of the composition of an automated laser cutting system for skull repair.

[0048] As shown in the figure: 1-1, medical image processing module, 1-2, 3D modeling module, 1-3, path planning module, 1-4, laser cutting execution module, 1-4-1. High-precision laser generator, 1-4-2. Multi-axis robotic arm, 1-4-3, material fixing platform, 1-5, real-time monitoring module, 1-5-1, temperature sensor, 1-5-2, visual positioning unit, 1-5-3, deformation detector, 1-6, postoperative evaluation module 1-6. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] The present invention will be further described in detail below with reference to the accompanying drawings.

[0050] Embodiment 1

[0051] As Figure 1 、Figure 2 , Figure 3 , Figure 4 , Figure 5 As shown in Figure 5 , this embodiment provides an automated laser cutting method for skull repair, where:

[0052] S1. Obtain the medical image data of the patient's skull defect area, perform three-dimensional modeling, and generate a three-dimensional reconstruction model of the defect site;

[0053] S1.1. Use a CT scanner to obtain tomographic scan data of the patient's skull area, with the X-ray tube voltage: 120 - 140 kV, slice thickness: 0.5 - 1.0 mm, and reconstruction matrix: 512×512 pixels; the image data is output as a DICOM format gray image sequence, and the Hounsfield unit HU of each pixel satisfies:

[0054] HU = μ tissue - μ water ×1000

[0055] μ tissue is the linear attenuation coefficient of the tissue, and μ water is the linear attenuation coefficient of water.

[0056] S1.2. Use the Perona - Malik diffusion equation to suppress the noise of the original DICOM sequence. The Perona - Malik diffusion equation:

[0057]

[0058] Diffusion coefficient I(x, y, z) is the three-dimensional voxel gray value, and K is the gradient threshold to suppress the noise in the non-edge area.

[0059] Then, based on the HU value, separate the skull from the soft tissue, select skull threshold segmentation, and select the threshold range:

[0060] Skull area = {(x, y, z)|200 ≤ HU(x, y, z) ≤ 3000}

[0061] The HU value of the high-density bone of the skull is significantly higher than that of the soft tissue HU < 200 and air HU ≈ -1000.

[0062] Then apply the Sobel edge detection operator to the segmented binary image:

[0063]

[0064] Edge intensity: Locate the skull defect boundary through the gradient magnitude

[0065] S1.3. Radially basis function interpolation is performed on the discrete edge point set P = {p i (x i , y i , z i )} of the defect area:

[0066]

[0067] The kernel function φ(r) = r 3 The coefficients α i and β are determined by solving a system of linear equations to fill the gap between the tomograms, and vertex interpolation is performed by running the formula:

[0068]

[0069] where p 1 , p 2 are adjacent vertices on the voxel edge, and f(p 1 ) < τ < f(p 2 ).

[0070] The interpolated volume data is converted into a triangular mesh model, and the isosurface threshold is defined:

[0071] τ = 0.5 × (max(f(p)) + min(f(p)))

[0072] The vertex status is calculated for each voxel unit, triangular patches are generated by looking up the table, and Laplacian smoothing is performed on the reconstructed mesh:

[0073]

[0074] where v i is the vertex coordinate, N(i) is the set of adjacent vertices, and λ ∈ [0.1, 0.5] is the smoothing coefficient.

[0075] S1.4. Output the three-dimensional reconstruction model and store it in STL format, including: the vertex coordinates V = {v k} of the triangular patches, the patch topological relationship F = {(v k1 , v k2 , v k3 )}, and the calculation of the defect area volume is as follows:

[0076]

[0077] where a t , b t , c t are the edge vectors of a single triangular patch, and T is the total number of patches.

[0078] S2. Match the preset standard parameter library of the repair implant according to the three-dimensional reconstruction model, and determine the shape, size and edge contour of the repair implant;

[0079] S2.1. Input the three-dimensional reconstruction model, and use the Gaussian filtering algorithm to smooth the three-dimensional model to remove noise and irrelevant details:

[0080]

[0081] where σ is the standard deviation, which controls the degree of smoothing.

[0082] S2.2. Input the smoothed three-dimensional model, and use the Canny edge detection algorithm to extract the edge features of the skull defect area:

[0083]

[0084] where is the gradient amplitude, which is used to detect edges.

[0085] S2.3. Use the DTW dynamic deformation algorithm to match the biomechanical adaptation parameters in the standard parameter library with the extracted edge features and the standard parameter library of the repair implant:

[0086]

[0087] where d(i, j) is the distance between feature points, which is used to match the edge contour.

[0088] S2.4. Input the matched parameters and edge contour, and use the least squares method to fit and determine the shape and size of the repair implant:

[0089]

[0090] where y i is the actual edge point, f(x i , θ) is the fitting function, and θ is the fitting parameter.

[0091] S2.5. Input the determined shape and size, and use B-spline curve fitting to generate the edge contour of the repair implant:

[0092]

[0093] where N i，p (t) is the B-spline basis function, P i is the control point, and t is the parameter.

[0094] S3. Based on the edge contour, generate a laser cutting path through a path planning algorithm and optimize the cutting parameters;

[0095] S3.1. Input the edge profile data of the repair implant, and use the parametric discretization method to convert the edge profile data into a discrete point set:

[0096]

[0097] where C(t) is the B-spline curve of the edge profile, and t i is the parametric variable, and P i is the discrete point.

[0098] S3.2. Input the discretized edge profile point set, and use the path planning algorithm to generate the laser cutting path. The path planning algorithm includes the progressive ant colony optimization algorithm and the genetic algorithm. Ant colony optimization algorithm: Initialize the ant population, and each ant traverses the discrete point set starting from the starting point. Calculate the transition probability of the ant from point i to point j:

[0099]

[0100] where τ ij is the pheromone concentration on the path i→j, η ij is the heuristic information and is the reciprocal of the distance, and α and β are the parameters that control the weights of the pheromone and the heuristic information. The ant colony optimization algorithm is used to guide the ants to select the path.

[0101] Genetic algorithm: Initialize the population, and each individual represents a possible cutting path. Calculate the total length of the fitness function path:

[0102]

[0103] where d(P i , P i+1 ) is the Euclidean distance between point P i and P i+1 . Optimize the population through selection, crossover, and mutation operations to generate the optimal path. The genetic algorithm is used to evaluate the quality of the path.

[0104] S3.3. Input the generated laser cutting path, and optimize the laser cutting parameters. The laser cutting parameters include laser power, cutting speed, kerf width, and heat affected zone range. The objective function: minF = w 1 ·Power + w 2 ·Speed + w 3 ·Kerf + w 4 ·HAZ, where Power is the laser power, Speed is the cutting speed, Kerf is the kerf width, HAZ is the heat affected zone range, and w 1 to w 4is the weight coefficient, which is used to balance the importance of the named parameters. In terms of the cutting quality constraint conditions, the slit width and the range of the heat affected zone need to meet the preset thresholds; in terms of the equipment capacity constraint conditions, the laser power and the cutting speed need to be within the allowable range of the equipment.

[0105] S4. Input the optimized cutting path and parameters into the laser cutting equipment, and drive the laser head to automatically cut the repair material;

[0106] S5. Monitor the temperature, deformation and positioning accuracy during the cutting process in real time, and dynamically adjust the laser power, cutting speed and focusing position.

[0107] S5.1. Input the data T raw (t) of the infrared temperature sensor, the data Δx(t), Δy(t) of the vision positioning system, and the data ∈(t) of the deformation detector, where T raw (t) represents the temperature of the cutting area, Δx(t), Δy(t) represent the lateral / longitudinal deviation between the cutting trajectory and the preset path, and ∈(t) represents the local strain or curvature change of the material; then use the Kalman filter to remove the sensor noise and improve the signal-to-noise ratio of the signal:

[0108] Prediction equation:

[0109]

[0110] Update equation:

[0111]

[0112] Among them, is the temperature estimated value at the Kth moment, z k is the sensor measurement value; A and B are the state transition matrices, Q is the covariance of the process noise and the observation noise, H is the observation matrix, and K k is the Kalman gain.

[0113] S5.2. Calculate the deviation between the actual temperature and the safety threshold T threshold :

[0114] E T (t) = max(0, T(t) - T threshold )

[0115] The adjustment is triggered only when the temperature exceeds the threshold.

[0116] Calculate the Euclidean distance deviation between the cutting trajectory and the preset path:

[0117]

[0118] Calculate the material deformation rate according to the strain data ∈(t):

[0119]

[0120] S5.3. Construct a dynamic weight objective function by integrating temperature, positioning, and deformation errors:

[0121] J(t) = w T ·E T (t) + w pos ·E pos (t) + w ∈ ·E ∈ (t)

[0122] where ω T , ω pos , ω ∈ are dynamic weight coefficients, adjusted according to the priority.

[0123] Set the laser power P(t):

[0124] P(t + △t) = P(t) - α T ·E T (t)

[0125] where the power linearly decreases when the temperature is too high, and α T is the power attenuation coefficient.

[0126] Set the cutting speed v(t):

[0127]

[0128] where the cutting speed is increased to reduce heat accumulation when the positioning deviation increases, and β pos is the speed gain coefficient.

[0129] Set the parameter adjustment rule for the focus position f(t):

[0130] f(t + △t) = f(t) + γ ∈ ·sgn(E ∈ (t))

[0131] where γ ∈ is the focus compensation step size, and the sign function controls the direction.

[0132] S5.4. Input the adjusted parameters P(t), v(t), and f(t) into the laser cutting equipment, and at the same time, update the sensor data in real time to form a closed-loop control; dynamically optimize the weight coefficient according to the historical error data to avoid local optimality:

[0133]

[0134] where η is the learning rate, and the weight allocation is optimized by the gradient descent method.

[0135] Example Two

[0136] As Figure 1 and Figure 6 shown, this embodiment provides an implementation manner of an automated laser cutting method for skull repair. The specific steps and related components are as follows:

[0137] (1) Obtain the medical image data of the patient's skull defect area.

[0138] Connect a CT scanner to the medical image processing module 1-1 to obtain the tomographic scan data of the patient's skull area, and output it as a sequence of grayscale images in DICOM format.

[0139] (2) Three-dimensional modeling.

[0140] The Perona-Malik diffusion equation is used to implement noise suppression on the original DICOM sequence in the medical image processing module 1-1. Separating the skull from the soft tissue based on the HU value, selecting the skull threshold segmentation, and applying the Sobel edge detection operator to the segmented binary image are also implemented in the medical image processing module 1-1. Radial basis function interpolation for the discrete edge point set of the defect area is implemented in the three-dimensional modeling module 1-2, and the interpolated volume data is converted into a triangular mesh model. Laplacian smoothing of the reconstructed mesh is implemented in the three-dimensional modeling module 1-2, and the three-dimensional reconstruction model is output and stored in STL format.

[0141] (3) Determine the shape, size, and edge contour of the repair implant.

[0142] Input the three-dimensional reconstruction model into the path planning module 1-3, and use the Gaussian filtering algorithm to smooth the three-dimensional model. Extracting the edge features of the skull defect area using the Canny edge detection algorithm is implemented in the path planning module 1-3. Matching the biomechanical adaptation parameters in the standard parameter library using the DTW dynamic deformation algorithm is implemented in the path planning module 1-3 to determine the shape and size of the repair implant. Generating the edge contour of the repair implant using B-spline curve fitting is implemented in the path planning module 1-3.

[0143] (4) Generate and optimize the laser cutting path and parameters.

[0144] Input the edge contour data of the repair implant into the path planning module 1-3, and use the parametric discretization method to convert the edge contour data into a discrete point set. Generating the laser cutting path using the progressive ant colony optimization algorithm and genetic algorithm is implemented in the path planning module 1-3. Optimizing the laser cutting parameters, including laser power, cutting speed, slit width, and heat affected zone range, is implemented in the path planning module 1-3.

[0145] (5) Perform the laser cutting operation.

[0146] Input the optimized cutting path and parameters into the laser cutting execution module 1-4, which includes a high-precision laser generator 1-4-1, a multi-axis robotic arm 1-4-2, and a material fixing platform 1-4-3. An adaptive fixture is provided at the end of the multi-axis robotic arm 1-4-2 to adjust the laser focal length according to the thickness of the repair material and perform the cutting operation.

[0147] (6) Monitor and adjust the cutting process in real time.

[0148] Use the temperature sensor 1-5-1, vision positioning unit 1-5-2, and deformation detector 1-5-3 in the real-time monitoring module 1-5 to monitor the temperature, deformation, and positioning accuracy during the cutting process. The removal of sensor noise using Kalman filtering is implemented within the real-time monitoring module 1-5 to calculate the deviation between the actual temperature and the safety threshold, the Euclidean distance deviation between the cutting trajectory and the preset path, and the material deformation rate. The construction of a dynamic weight objective function is implemented within the real-time monitoring module 1-5 to dynamically adjust the laser power, cutting speed, and focusing position to form a closed-loop control.

[0149] Embodiment III

[0150] As Figure 1 and Figure 6 shown, based on Embodiment II, this embodiment further details the specific implementation method of laser cutting path planning, including the refinement of step S3:

[0151] (1) Within the path planning module 1-3, use the parametric discretization method to convert the edge contour data of the repair implant into a discrete point set. Specifically, the edge contour data can be represented as a series of ordered points, each point containing three-dimensional coordinate information.

[0152] (2) Use the progressive ant colony optimization algorithm to generate a preliminary cutting path. The ant colony algorithm searches for the optimal path in the solution space by simulating the behavior of ants foraging. In this embodiment, the discrete point set of the edge contour is used as the search space of the ant colony, and the pheromone concentration is iteratively updated to guide the ant colony to gradually converge to the optimal cutting path.

[0153] (3) On the basis of the preliminary cutting path, use the genetic algorithm for further optimization. The genetic algorithm encodes, crosses, mutates, etc. the preliminary cutting path by simulating natural selection and genetic mechanisms to generate a series of new cutting path candidate solutions. By evaluating the fitness values of each candidate solution, such as the length and smoothness of the cutting path, the optimal solution is selected as the final cutting path.

[0154] (4) After determining the final cutting path, further optimize the laser cutting parameters. According to the characteristics of the cutting path such as curvature, length, etc., adjust parameters such as laser power, cutting speed, slit width, and the range of the heat affected zone to ensure the stability and precision of the cutting process.

[0155] Example 4

[0156] As Figure 1 and Figure 6 shown,

[0157] Based on Examples 2 and 3, this example further describes the specific implementation details of real-time monitoring and adjustment of the cutting process, including the refinement of step S5:

[0158] (1) Use the temperature sensor 1-5-1 in the real-time monitoring module 1-5 to monitor the temperature of the cutting area. The temperature sensor can measure the temperature value of the cutting area in real time and transmit it to the real-time monitoring module for processing.

[0159] (2) Use the vision positioning unit 1-5-2 to monitor the deviation between the cutting trajectory and the preset path. The vision positioning unit takes real-time images during the cutting process, compares them with the preset path, and calculates the lateral and longitudinal deviations of the cutting trajectory.

[0160] (3) Use the deformation detector 1-5-3 to monitor the deformation of the material. The deformation detector can measure the local strain or curvature change of the material in real time and transmit it to the real-time monitoring module for processing.

[0161] (4) In the real-time monitoring module 1-5, use the Kalman filtering algorithm to denoise the sensor data and improve the signal-to-noise ratio. Then, calculate indicators such as the deviation between the actual temperature and the safety threshold, the Euclidean distance deviation between the cutting trajectory and the preset path, and the material deformation rate.

[0162] (5) Construct a dynamic weight objective function, comprehensively considering the influence of temperature, positioning, and deformation errors on the cutting quality. Dynamically adjust the weight coefficients according to historical error data to avoid the occurrence of local optimal solutions.

[0163] (6) According to the calculation results of the dynamic weight objective function, dynamically adjust cutting parameters such as laser power, cutting speed, and focusing position. The adjusted parameters are transmitted to the laser cutting execution module 1-4 in real time through a closed-loop feedback mechanism to achieve precise control of the cutting process.

[0164] The basic principles, main features and advantages of the present invention have been shown and described above. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. An automated laser cutting method for skull repair, characterized in that include: S1. Obtain medical imaging data of the patient's skull defect area, perform three-dimensional modeling, and generate a three-dimensional reconstruction model of the defect area; S2, according to the three-dimensional reconstructed model, matching a preset repair implant standard parameter library to determine the shape, size and edge contour of the repair implant; S3, based on the edge profile, generating a laser cutting path through a path planning algorithm, and optimizing cutting parameters; S4, inputting the optimized cutting path and parameters into the laser cutting equipment, and driving the laser head to automatically cut the repair material; S5. Real-time monitoring of temperature, deformation and positioning accuracy during the cutting process, and dynamic adjustment of laser power, cutting speed and focus position.

2. The automated laser cutting method for skull repair according to claim 1, characterized in that The S1 further comprises: S1.

1. Use a CT scanner to obtain tomographic data of the patient's skull area, and output the image data as a grayscale image sequence in DICOM format; S1.2, the Perona-Malik diffusion equation is used to suppress the noise of the original DICOM sequence, the skull and soft tissue are separated based on the HU value, the skull threshold is selected for segmentation, and the Sobel edge detection operator is applied to the segmented binary image. S1.3, for the discrete edge point set P of the defect area = {p i (x i ,y i , z i )} perform radial basis function interpolation, convert the interpolated volume data into a triangular mesh model, and perform Laplace smoothing on the reconstructed mesh; S1.

4. Output the 3D reconstructed model and store it in STL format.

3. The automated laser cutting method for skull repair according to claim 1, characterized in that The S2 further includes: S2.1, input the 3D reconstruction model, use the Gaussian filtering algorithm to smooth the 3D model and remove noise and irrelevant details; S2.2, input the smoothed 3D model and use the Canny edge detection algorithm to extract the edge features of the skull defect area; S2.3, using the DTW dynamic deformation algorithm to match the extracted edge features and the standard parameter library of the repair implant with the biomechanical adaptation parameters in the standard parameter library; S2.4, input the matched parameters and edge contour, and use the least squares fitting method to determine the shape and size of the repair implant; S2.

5. Input the determined shape and size, and use B-spline curve fitting to generate the edge contour of the repair implant.

4. The automated laser cutting method for skull repair according to claim 1, characterized in that The S3 further includes: S3.1, inputting edge contour data of the repair implant, and converting the edge contour data into a discrete point set using a parametric discretization method; S3.2, inputting a discretized edge contour point set, and using a path planning algorithm to generate a laser cutting path, the path planning algorithm includes a progressive ant colony optimization algorithm and a genetic algorithm; S3.

3. Input the generated laser cutting path and optimize the laser cutting parameters. The laser cutting parameters include laser power, cutting speed, slit width, and heat affected zone range.

5. The automated laser cutting method for skull repair according to claim 1, characterized in that The S5 further includes: S5.

1. Input infrared temperature sensor data T raw (t), visual positioning system data Δx(t), Δy(t), deformation detector data ∈(t), Kalman filtering is used to remove sensor noise and improve the signal-to-noise ratio, where T raw (t) represents the temperature of the cutting area, Δx(t), Δy(t) represent the lateral / longitudinal deviation of the cutting trajectory from the preset path, and ∈(t) represents the local strain or curvature change of the material; S5.

2. Calculate the actual temperature and safety threshold T threshold Deviation; Calculate the Euclidean distance deviation between the cutting trajectory and the preset path; Calculate the material deformation rate according to the strain data ∈(t); S5.3, comprehensively consider temperature, positioning and deformation errors, construct dynamic weight objective function, and set parameter adjustment rules of laser power P(t), cutting speed υ(t), and focus position f(t); S5.

4. Input the adjusted parameters P(t), υ(t), and f(t) into the laser cutting equipment, and update the sensor data in real time to form a closed-loop control; dynamically optimize the weight coefficient according to the historical error data to avoid local optimality.

6. An automated laser cutting system for implementing the method according to any one of claims 1 to 5, characterized in that include: A medical image processing module (1-1), used for acquiring and processing medical image data of a skull defect area; A three-dimensional modeling module (1-2), connected to the image processing module, to generate a three-dimensional model of the defective part; Path planning module (1-3), outputs optimized laser cutting path and parameters based on the 3D model; A laser cutting execution module (1-4) includes a high-precision laser generator (1-4-1), a multi-axis mechanical arm (1-4-2) and a material fixing platform (1-4-3), and is used to perform a cutting operation; an adaptive clamp is provided at the end of the multi-axis mechanical arm (1-4-2) for adjusting the laser focal length according to the thickness of the repair material; A real-time monitoring module (1-5), including a temperature sensor (1-5-1), a visual positioning unit (1-5-2) and a deformation detector (1-5-3), is used to dynamically adjust the cutting process; the real-time monitoring module (1-5) is connected to the path planning module (1-3) through a closed-loop feedback mechanism to achieve adaptive iterative optimization of cutting parameters; The postoperative evaluation module (1-6) is used to perform a three-dimensional scanning comparison between the cut implant and the three-dimensional model and output a matching report.

7. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method described in any one of claims 1 to 5 is implemented.

Images