A stereotactic positioning system for neurosurgery based on multi-modal image fusion

Abstract

The application relates to the field of stereotactic surgery technology, and discloses a neurosurgical stereotactic surgery positioning system based on multi-modal image fusion, which comprises a multi-modal image acquisition module, an intelligent registration and fusion module, a three-dimensional modeling module, a surgery navigation engine, an augmented reality interface and a dynamic calibration module; through the setting of the cross-modal elastic registration module, when the multi-modal image fusion is carried out, the spatial matching degree of functional images and structural images is detected in real time through a nonlinear deformation compensation algorithm; when there is a registration deviation between the white matter fiber bundle and the tumor boundary, the system dynamically generates an elastic deformation compensation field, the positioning accuracy of the functional area is improved, and the risk of functional area cutting is reduced; through the setting of the dynamic drift correction module, when deep target point navigation is carried out, the brain tissue displacement change is sensed in real time based on laser surface scanning, the accurate correction of the position and posture of a surgical instrument is realized, and the reliability of deep target point positioning is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of stereotactic surgery, in particular to a neurosurgical stereotactic surgery positioning system based on multi-modal image fusion. BACKGROUND

[0002] Stereotactic surgery is a minimally invasive technique in the field of neurosurgery that achieves precise positioning by establishing a three-dimensional coordinate system. Its core principle is based on the coordinate system of stereographic geometry, combined with CT, MRI image guidance and electrophysiological auxiliary means, which can complete the biopsy, drainage and treatment of deep brain lesions with sub-millimeter accuracy. This technology is mainly applied to the diagnosis and treatment of functional neurosurgery diseases and complex brain diseases, covering Parkinson's disease, epilepsy, mental illness, cerebral hemorrhage and tumor treatment. The technology has developed various branches including stereotactic radiosurgery and neural nucleus transplantation. The surgical mortality rate has dropped to 0-1%. With the application of multi-modal image fusion and intelligent devices, its positioning accuracy and indication range continue to expand.

[0003] Currently, due to the involvement of multi-modal image acquisition and dynamic anatomical structure changes in neurosurgical stereotactic surgery, when performing multi-modal image fusion registration, the existing system cannot detect the spatial elastic deformation error of functional images and structural images in real time. When the white matter fiber bundle and the tumor boundary are spatially misaligned, it can cause a positioning deviation of the neural function area of up to 2-3 mm, which cannot guarantee the accuracy of surgical positioning. At the same time, when navigating deep target points, it cannot monitor the intraoperative dynamic drift of brain tissue in real time, which can cause a 3-8 mm deviation between the navigation trajectory and the actual anatomical position, and cannot correct the positioning coordinates in real time when the tissue drift exceeds the safety threshold. During the surgical navigation process, due to the need to simultaneously display the blood vessel course, functional area boundary and multi-dimensional electrophysiological information, the existing system cannot achieve dynamic fusion and presentation of multi-source data, resulting in the need for the operator to switch between different display interfaces for observation, further affecting the efficiency of surgical decision-making and lesion complete resection rate.

[0004] Therefore, the present application provides a neurosurgical stereotactic surgery positioning system based on multi-modal image fusion to solve the above problems. SUMMARY

[0005] (I) Technical problems solved

[0006] In view of the deficiencies of the prior art, the present application provides a neurosurgical stereotactic surgery positioning system based on multi-modal image fusion to solve the problems raised in the background art.

[0007] (II) Technical solutions

[0008] To achieve the above purpose, the present application provides the following technical solutions: a neurosurgical stereotactic surgery positioning system based on multi-modal image fusion, comprising:

[0009] Multi-modal image acquisition module: including MRI scanning unit, DTI imaging unit and fMRI acquisition unit;

[0010] Intelligent registration fusion module: performing cross-modality registration and wavelet-convolution fusion algorithm;

[0011] Three-dimensional modeling module: implementing improved marching cubes reconstruction and blood flow parameter mapping;

[0012] Surgical navigation engine module: including spatial registration unit, path planning unit and real-time positioning unit;

[0013] Augmented reality interface module: integrating optical positioning system and holographic display device;

[0014] Dynamic calibration module: realizing instrument pose tracking and real-time path correction.

[0015] Preferably, the following steps are included:

[0016] S1, acquiring multi-modal medical image data of the patient's head, including structural image data and functional image data;

[0017] S2, performing cross-modality registration processing on the structural image data and functional image data to generate registered multi-modal image data;

[0018] S3, performing image fusion processing based on the registered multi-modal image data to generate multi-modal fusion image data;

[0019] S4, performing three-dimensional reconstruction processing on the multi-modal fusion image data to construct a three-dimensional anatomical model of the surgical area;

[0020] S5, performing spatial mapping registration between the three-dimensional anatomical model and the actual head position of the patient to generate a registered surgical navigation model;

[0021] S6, planning a surgical path and target point coordinates on the surgical navigation model to generate surgical planning data;

[0022] S7, acquiring real-time spatial pose data of the surgical instrument and mapping the spatial pose data to the surgical navigation model to generate instrument real-time positioning data;

[0023] S8, performing surgical navigation guidance based on the surgical planning data and instrument real-time positioning data to generate augmented reality navigation information;

[0024] S9, performing dynamic path correction according to the spatial deviation between the surgical instrument and the target point coordinates to generate real-time path calibration instructions.

[0025] Preferably, the S1 includes:

[0026] S11. Acquire T1- and T2-weighted structural imaging data of the patient's head using magnetic resonance imaging equipment;

[0027] S12. Acquire functional image data of white matter fiber bundles using a diffusion tensor imaging device;

[0028] S13. Acquire functional imaging data of blood oxygenation level-dependent signals using functional magnetic resonance imaging equipment;

[0029] S14. Construct a multimodal image data matrix ,in For the first Various image modal data, This represents the total number of image modalities. It is a multimodal data matrix.

[0030] Preferably, S2 includes:

[0031] S21. Extract the set of bony landmarks from the structural image data. ,in For structural image point sets, These are structural feature points;

[0032] S22. Extract the feature point set of the functional image data. ,in For functional image point sets, For functional feature points;

[0033] S23. Calculate the transformation matrix using the improved ICP algorithm:

[0034]

[0035] in The optimal spatial transformation matrix. For nonlinear deformation constraint weights, This is a penalty term for elastic deformation. For the optimal solution operator, For structural image feature points, For functional image feature points, The square of the Euclidean distance between the feature points;

[0036] S24, Application Spatial transformation is performed on functional image data to generate registered data. .

[0037] Preferably, S3 includes:

[0038] S31. Construct a wavelet-convolutional fusion model:

[0039]

[0040] in For low-frequency fusion components, For high-frequency fusion components, Low-frequency coefficients These are high-frequency coefficients. It is an adaptive weighting matrix. These are the low-frequency coefficients of the structural image. For functional image low-frequency coefficients, This is a 3D convolution operation. For stitching together high-frequency coefficients of structural and functional images. For the Sigmoid function;

[0041] S32. Generate the fused image through inverse wavelet transform:

[0042]

[0043] in To merge the output images.

[0044] Preferably, S4 includes:

[0045] S41. Extracting isosurfaces using an improved moving cube algorithm:

[0046]

[0047] Among them For the set of grid vertices, The total number of voxels. It is the j-th vertex of the i-th voxel;

[0048] The vertex determination function is defined as follows:

[0049]

[0050] in The gradient direction vector, The image gradient vector, Gradient threshold;

[0051] S42. Construct a three-dimensional model with hemodynamic parameters. ,in These are blood flow parameters.

[0052] Preferably, S5 includes:

[0053] S51. Attach infrared reflective marker balls to the patient's head. ;

[0054] S52. Establish the mapping relationship between the spatial coordinates of the marker sphere and the model marker points:

[0055]

[0056] in Pre-set marker points for the model. Let be the spatial transformation matrix. Let k be the coordinates of the k-th marker sphere in real space. The optimal solution operator;

[0057] S53. Update the registration matrix in real time using extended Kalman filtering. .

[0058] Preferably, S6 includes:

[0059] S61. Establish a path optimization model based on blood vessel avoidance constraints:

[0060]

[0061] in For path curvature, It is a vascular structure. For the surgical path, The path is the differential arc length. This is the minimum vessel distance operator. It is a spatial distance function. Weighting for blood vessels;

[0062] S62, Output the optimal path set ,in For path trajectory function, The coordinates of the skull entry point are: The coordinates of the lesion target point.

[0063] Preferably, S7 includes:

[0064] S71. Acquire the coordinates of the instrument tip using an optical positioning system. and direction vector ;

[0065] S72. The mapping position of the computing instrument in model space:

[0066]

[0067] in Let be the coordinates of the instrument in the model space at time t. Let be the spatial transformation matrix at time t. Let t be the actual spatial coordinates of the instrument tip;

[0068] S73, Generate real-time pose data of the instrument ,in The instrument direction vector, Let t represent the spatial state of the surgical instruments.

[0069] Preferably, S8 includes:

[0070] S81. Constructing an augmented reality overlay model:

[0071]

[0072] in This represents the perspective superposition operator. To enhance the view of reality, Images of real-world scenes. Transparency coefficient It is a three-dimensional anatomical model;

[0073] Simultaneously define the highlight function:

[0074]

[0075] in For path neighborhood, For pixel coordinates, For pixel RGB values, The optimal surgical approach;

[0076] S82. Generate a danger zone warning signal:

[0077]

[0078] in As a warning signal identifier, Let be the coordinates of the instrument model at time t. These are the coordinates of key anatomical structures.

[0079] (III) Beneficial Effects

[0080] Compared with existing technologies, this invention provides a neurosurgical stereotactic surgical positioning system based on multimodal image fusion, which has the following beneficial effects:

[0081] 1. This invention sets up a cross-modal elastic registration module. When performing multimodal image fusion, it uses a nonlinear deformation compensation algorithm to detect the spatial matching degree between functional images and structural images in real time. When there is a registration deviation between the white matter fiber bundle and the tumor boundary, the system dynamically generates an elastic deformation compensation field to improve the localization accuracy of the nerve functional area and reduce the risk of miscutting the functional area.

[0082] 2. This invention, by setting up a dynamic drift correction module, senses changes in brain tissue displacement in real time based on laser surface scanning during deep target navigation. When the navigation trajectory deviates from the actual anatomical location, the system automatically updates the spatial mapping relationship of the three-dimensional model, thereby achieving precise correction of the surgical instrument pose and ensuring the reliability of deep target positioning.

[0083] 3. By setting up a multi-source fusion channel, this invention uses augmented reality technology to dynamically fuse vascular course, functional area boundaries, and electrophysiological data in three dimensions during surgical navigation. The system automatically generates fused images and projects them onto the surgical field, eliminating the cognitive burden of switching between multi-dimensional information and improving the efficiency of surgical decision-making and the completeness of lesion treatment. Attached Figure Description

[0084] Figure 1 This is a schematic diagram of the overall system architecture of the present invention. Detailed Implementation

[0085] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0086] Please see Figure 1 This is a neurosurgical stereotactic surgical positioning system based on multimodal image fusion. The system includes:

[0087] Multimodal image acquisition module: includes MRI scanning unit, DTI imaging unit and fMRI acquisition unit;

[0088] Intelligent registration and fusion module: performs cross-modal registration and wavelet-convolution fusion algorithms;

[0089] 3D modeling module: Implements improved moving cube reconstruction and blood flow parameter mapping;

[0090] Surgical navigation engine module: includes spatial registration unit, path planning unit and real-time positioning unit;

[0091] Augmented Reality Interface Module: Integrates an optical positioning system and a holographic display device;

[0092] Dynamic calibration module: Enables instrument pose tracking and real-time path correction.

[0093] Includes the following steps:

[0094] S1. Collect multimodal medical imaging data of the patient's head, including structural imaging data and functional imaging data;

[0095] S2. Perform cross-modal registration processing on structural image data and functional image data to generate registered multimodal image data;

[0096] S3. Perform image fusion processing based on the registered multimodal image data to generate multimodal fused image data;

[0097] S4. Perform three-dimensional reconstruction processing on the multimodal fused image data to construct a three-dimensional anatomical model of the surgical area;

[0098] S5. Spatial mapping and registration are performed between the three-dimensional anatomical model and the actual head position of the patient to generate a registered surgical navigation model.

[0099] S6. Plan the surgical path and target coordinates on the surgical navigation model and generate surgical planning data;

[0100] S7. Real-time acquisition of spatial pose data of surgical instruments, mapping spatial pose data to surgical navigation model, and generating real-time instrument positioning data;

[0101] S8. Based on surgical planning data and real-time instrument positioning data, perform surgical navigation guidance and generate augmented reality navigation information;

[0102] S9. Based on the spatial deviation between the surgical instruments and the target coordinates, perform dynamic path correction and generate real-time path calibration instructions.

[0103] S1 includes:

[0104] S11. Acquire T1- and T2-weighted structural imaging data of the patient's head using magnetic resonance imaging equipment;

[0105] S12. Acquire functional image data of white matter fiber bundles using a diffusion tensor imaging device;

[0106] S13. Acquire functional imaging data of blood oxygenation level-dependent signals using functional magnetic resonance imaging equipment;

[0107] S14. Construct a multimodal image data matrix ,in For the first Various image modal data, This represents the total number of image modalities. It is a multimodal data matrix.

[0108] S2 includes:

[0109] S21. Extract the set of bony landmarks from structural imaging data. ,in For structural image point sets, These are structural feature points;

[0110] S22. Extract the feature point set from the functional image data. ,in For functional image point sets, For functional feature points;

[0111] S23. Calculate the transformation matrix using the improved ICP algorithm:

[0112]

[0113] in The optimal spatial transformation matrix. For nonlinear deformation constraint weights, This is a penalty term for elastic deformation. For the optimal solution operator, For structural image feature points, For functional image feature points, The square of the Euclidean distance between the feature points;

[0114] S24, Application Spatial transformation is performed on functional image data to generate registered data. .

[0115] S3 includes:

[0116] S31. Construct a wavelet-convolutional fusion model:

[0117]

[0118] in For low-frequency fusion components, For high-frequency fusion components, Low-frequency coefficients These are high-frequency coefficients. It is an adaptive weighting matrix. These are the low-frequency coefficients of the structural image. For functional image low-frequency coefficients, This is a 3D convolution operation. For stitching together high-frequency coefficients of structural and functional images. For the Sigmoid function;

[0119] S32. Generate the fused image through inverse wavelet transform:

[0120]

[0121] in To merge the output images.

[0122] S4 includes:

[0123] S41. Extracting isosurfaces using an improved moving cube algorithm:

[0124]

[0125] Among them For the set of grid vertices, The total number of voxels. It is the j-th vertex of the i-th voxel;

[0126] The vertex determination function is defined as follows:

[0127]

[0128] in The gradient direction vector, The image gradient vector, Gradient threshold;

[0129] S42. Construct a three-dimensional model with hemodynamic parameters. ,in These are blood flow parameters.

[0130] S5 includes:

[0131] S51. Attach infrared reflective marker balls to the patient's head. ;

[0132] S52. Establish the mapping relationship between the spatial coordinates of the marker sphere and the model marker points:

[0133]

[0134] in Pre-set marker points for the model. Let be the spatial transformation matrix. Let k be the coordinates of the k-th marker sphere in real space. The optimal solution operator;

[0135] S53. Update the registration matrix in real time using extended Kalman filtering. .

[0136] S6 includes:

[0137] S61. Establish a path optimization model based on blood vessel avoidance constraints:

[0138]

[0139] in For path curvature, It is a vascular structure. For the surgical path, The path is the differential arc length. This is the minimum vessel distance operator. It is a spatial distance function. Weighting for blood vessels;

[0140] S62, Output the optimal path set ,in For path trajectory function, The coordinates of the skull entry point are: The coordinates of the lesion target point.

[0141] S7 includes:

[0142] S71. Acquire the coordinates of the instrument tip using an optical positioning system. and direction vector ;

[0143] S72. The mapping position of the computing instrument in model space:

[0144]

[0145] in Let be the coordinates of the instrument in the model space at time t. Let be the spatial transformation matrix at time t. Let t be the actual spatial coordinates of the instrument tip;

[0146] S73, Generate real-time pose data of the instrument ,in The instrument direction vector, Let t represent the spatial state of the surgical instruments.

[0147] S8 includes:

[0148] S81. Constructing an augmented reality overlay model:

[0149]

[0150] in This represents the perspective superposition operator. To enhance the view of reality, Images of real-world scenes. Transparency coefficient It is a three-dimensional anatomical model;

[0151] Simultaneously define the highlight function:

[0152]

[0153] in For path neighborhood, For pixel coordinates, For pixel RGB values, The optimal surgical approach;

[0154] S82. Generate a danger zone warning signal:

[0155]

[0156] in As a warning signal identifier, Let be the coordinates of the instrument model at time t. These are the coordinates of key anatomical structures.

[0157] Example 1: Epilepsy lesion resection surgery:

[0158] In a surgical resection of an epileptic lesion, the patient's right temporal lobe epileptic lesion was located adjacent to the language center. Preoperative fMRI showed a 32% overlap between the activation area and the lesion boundary. Traditional navigation carries the risk of mis-resection due to intraoperative image drift errors. The new technology utilizes multimodal fusion registration of DTI fiber bundles and enhanced MRI, employing an elastic compensation mechanism to eliminate spatial distortion of the optic radiation fibers. In real-time navigation, the AR interface projects fused images and marks the language area in red, while path planning enhances vascular avoidance. The dynamic correction mechanism activates a drift compensation algorithm to update the surgical path when a 0.6mm brain displacement is detected. Ultimately, this resulted in a 100% lesion resection rate, a 93.5% language function preservation rate, a white matter fiber registration error reduced to 0.4mm, and a 38-minute reduction in surgical time.

[0159] Example 2: Brainstem glioma biopsy surgery:

[0160] For a 2.1cm glioma in the left medulla oblongata encasing perforating vessels of the basilar artery, CT angiography revealed three perforating vessels, each 0.6-0.8mm in diameter, traversing the tumor. The technique employed multi-source fusion channels to enhance microvascular contrast; hazard avoidance navigation activated a dynamic avoidance mechanism via directional vectors, calculating the distance between the device and the vessel in real time; when the device was less than 1.5mm from the vessel, it triggered 200Hz high-frequency vibration tactile feedback and AR pulsed red light alarm. The results successfully avoided 11 perforating vessels, with the biopsy path 1.2mm from the nearest vessel. The safety system's circuit breaker response was less than 0.8 seconds during two episodes of respiratory arrest induced by brainstem stimulation. Postoperative pathological diagnosis was 100% with no vascular complications.

[0161] Example 3: Deep brain stimulation for Parkinson's disease:

[0162] Implantation of bilateral subthalamic nuclei electrodes in Parkinson's patients faces the challenge of achieving a target distance of only 1.8 mm between the posterior limb fiber tract of the internal capsule and the target point. The procedure involved activating the DBS golden path database and the chord measurement conversion algorithm; spatial mapping was used to increase the optical tracking sampling rate to 45 Hz, ensuring a mapping delay of less than 3 ms; and the pose control module dynamically corrected the electrode direction vector to an angular deviation of 3.2 degrees. Ultimately, the left electrode achieved a safe distance of 2.3 mm from the internal capsule, with implantation three-dimensional coordinate deviations of 0.3 mm on the X-axis, 0.1 mm on the Y-axis, and 0.1 mm on the Z-axis. Postoperative UPDRS-III score improvement rate was 71.4%, and electrode contact full activation rate was 100%.

[0163] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0164] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A neurosurgical stereotactic surgical positioning system based on multimodal image fusion, characterized in that: The system includes: Multimodal image acquisition module: Includes MRI scanning unit, DTI imaging unit, and fMRI acquisition unit, used to acquire multimodal medical imaging data of the patient's head, including structural and functional imaging data, including: T1- and T2-weighted structural imaging data of the patient's head were acquired using magnetic resonance imaging equipment. Functional image data of white matter fiber bundles were acquired using a diffusion tensor imaging device. Functional imaging data of blood oxygenation level-dependent signals were acquired using functional magnetic resonance imaging (fMRI). Constructing a multimodal image data matrix ,in For the first Various image modal data, This represents the total number of image modalities. It is a multimodal data matrix; Intelligent registration and fusion module: Executes cross-modal registration and wavelet-convolution fusion algorithms to perform cross-modal registration processing on the structural image data and functional image data, generating registered multimodal image data, and performs image fusion processing based on the registered multimodal image data to generate multimodal fused image data, wherein the cross-modal registration includes: Extract the set of bony landmarks from the structural image data. ,in For structural image point sets, These are structural feature points; Extract the feature point set from the functional image data ,in For functional image point sets, For functional feature points; The transformation matrix is ​​calculated using an improved ICP algorithm: in The optimal spatial transformation matrix. For nonlinear deformation constraint weights, This is a penalty term for elastic deformation. For the optimal solution operator, For structural image feature points, For functional image feature points, The square of the Euclidean distance between the feature points; application Spatial transformation is performed on functional image data to generate registered data. ; 3D modeling module: Implements improved moving cube reconstruction and blood flow parameter mapping to perform 3D reconstruction processing on the multimodal fused image data and construct a three-dimensional anatomical model of the surgical area; Surgical navigation engine module: includes a spatial registration unit, a path planning unit, and a real-time positioning unit. It is used to perform spatial mapping and registration between the three-dimensional anatomical model and the actual head position of the patient, generate a registered surgical navigation model, plan the surgical path and target coordinates on the surgical navigation model, generate surgical planning data, and collect the spatial pose data of surgical instruments in real time, map the spatial pose data to the surgical navigation model, and generate real-time instrument positioning data. Augmented Reality Interface Module: Integrates an optical positioning system and a holographic display device, used for surgical navigation guidance based on the surgical planning data and real-time instrument positioning data, and generating augmented reality navigation information; Dynamic calibration module: Enables instrument pose tracking and real-time path correction. It is used to dynamically correct the path based on the spatial deviation between the surgical instrument and the target coordinates, and generate real-time path calibration instructions.

2. The neurosurgical stereotactic surgical positioning system based on multimodal image fusion according to claim 1, characterized in that: The intelligent registration and fusion module performs image fusion processing based on the registered multimodal image data, including: Constructing a wavelet-convolution fusion model: in For low-frequency fusion components, For high-frequency fusion components, Low-frequency coefficients These are high-frequency coefficients. It is an adaptive weighting matrix. These are the low-frequency coefficients of the structural image. For functional image low-frequency coefficients, This is a 3D convolution operation. For stitching together high-frequency coefficients of structural and functional images. For the Sigmoid function; Generate a fused image using inverse wavelet transform: in To merge the output images.

3. The neurosurgical stereotactic surgical positioning system based on multimodal image fusion according to claim 1, characterized in that: The 3D modeling module performs 3D reconstruction processing on the multimodal fused image data, including: An improved moving cube algorithm is used to extract isosurfaces: in For the set of grid vertices, The total number of voxels. It is the j-th vertex of the i-th voxel; The vertex determination function is defined as follows: in The gradient direction vector, The image gradient vector, Gradient threshold; Constructing a three-dimensional model with hemodynamic parameters ,in These are blood flow parameters.

4. The neurosurgical stereotactic surgical positioning system based on multimodal image fusion according to claim 1, characterized in that: The surgical navigation engine module performs spatial mapping and registration between the three-dimensional anatomical model and the patient's actual head position, including: Infrared reflective marker balls were attached to the patient's head. ; Establish the mapping relationship between the spatial coordinates of the marker sphere and the marker points of the model: in Pre-set marker points for the model. Let be the spatial transformation matrix. Let k be the coordinates of the k-th marker sphere in real space. The optimal solution operator; The registration matrix is ​​updated in real time using extended Kalman filtering. .

5. The neurosurgical stereotactic surgical positioning system based on multimodal image fusion according to claim 1, characterized in that: The surgical navigation engine module plans the surgical path and target coordinates on the surgical navigation model, including: A path optimization model based on blood vessel avoidance constraints was established. in For path curvature, It is a vascular structure. For the surgical path, The path is the differential arc length. This is the minimum vessel distance operator. It is a spatial distance function. Weighting for blood vessels; Output the optimal path set ,in For path trajectory function, The coordinates of the skull entry point are: The coordinates of the lesion target point.

6. The neurosurgical stereotactic surgical positioning system based on multimodal image fusion according to claim 1, characterized in that: The surgical navigation engine module maps the spatial pose data to the surgical navigation model, including: The coordinates of the instrument tip are acquired using an optical positioning system. and direction vector ; The mapping position of the computational instrument in model space: in Let be the coordinates of the instrument in the model space at time t. Let be the spatial transformation matrix at time t. Let t be the actual spatial coordinates of the instrument tip; Generate real-time pose data of the instrument ,in The instrument direction vector, Let t represent the spatial state of the surgical instruments.

7. The neurosurgical stereotactic surgical positioning system based on multimodal image fusion according to claim 1, characterized in that: The surgical navigation guidance based on the surgical planning data and real-time instrument positioning data in the augmented reality interface module includes: Constructing augmented reality overlay models: in This represents the perspective superposition operator. To enhance the view of reality, Images of real-world scenes. Transparency coefficient It is a three-dimensional anatomical model; Simultaneously define the highlight function: in For path neighborhood, For pixel coordinates, For pixel RGB values, The optimal surgical approach; Generate a danger zone warning signal: in As a warning signal identifier, Let be the coordinates of the instrument model at time t. These are the coordinates of key anatomical structures.

Images