A space mapping based positioning and navigation system and method

Abstract

This invention discloses a spatial mapping-based positioning and navigation system and method. The system includes a key point acquisition module for acquiring key points on both sides of the lamina of a target vertebral segment within a defined anatomical region in a medical image; a spinous process tangent point acquisition module for determining the spinous process tangent point of the target vertebral segment in the medical image; a registration tool module including at least two coplanar point selection parts and an abutment part, respectively selecting feature points on both sides of the lamina of the target vertebral segment corresponding to the aforementioned two key points and the corresponding points of the aforementioned spinous process tangent point in actual space; a registration module for establishing the transformation relationship between actual space and the medical image; and a gyroscope navigation module for acquiring the real-time attitude of the registration tool module, and, combined with the transformation relationship between actual space and the medical image, locating and performing navigation on a planned channel in the medical image. This invention can achieve planned channel positioning, thereby realizing spatial mapping-based positioning and navigation.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a positioning and navigation system and method based on spatial mapping. Background Technology

[0002] Currently, orthopedic surgical robots combined with navigation technology have significantly improved the precision and safety of surgeries, especially in complex anatomical locations such as the spine, where their high-precision positioning capabilities have significantly increased the success rate of implant placement. The ability to provide real-time feedback on the position of surgical tools effectively reduces the risk of human error and decreases reliance on the surgeon's skill level. However, this technology typically relies on a complex system, including robotic arms, optical positioning devices, and C-arm machines, and existing methods lack detailed theoretical explanations regarding the specific procedures for preoperative key point selection and intraoperative CT image registration. Summary of the Invention

[0003] Purpose of the invention: To address the above-mentioned shortcomings, this invention proposes a positioning and navigation system based on spatial mapping, which only requires a gyroscope navigation system to achieve rapid positioning of a planned channel, thereby realizing positioning and navigation based on spatial mapping.

[0004] Technical solution: This invention provides a positioning and navigation system based on spatial mapping, comprising: The registration tool module includes at least two selection parts and one contact part, wherein the selection parts are symmetrical about the contact part; the selection parts and the contact part are used to select actual key points on the area to be dissected and determine the actual setting plane, wherein the actual setting plane serves as the initial pose of the registration tool module; The image acquisition module is used to acquire the image setting plane of the area to be dissected in the medical image. The image setting plane is determined based on the correspondence between the actual key points and the image key points in the medical image. The registration module is used to calculate the transformation relationship between the registration tool module and the medical image based on the spatial mapping between the actual setting plane and the image setting plane; A gyroscope module, installed on the registration tool module, is used to obtain the angle of rotation of the registration tool module based on the axis formed by the selection part; The navigation module calculates the real-time pose of the registration tool module based on its initial pose and the rotation angle of the registration tool module obtained by the gyroscope module. Based on this, and combined with the transformation relationship between the registration tool module and the medical image obtained by the registration module, the module locates the planned channel in the medical image and performs navigation.

[0005] Specifically, the actual key points on the area to be dissected include a first actual key point and a second actual key point located on the target cone symmetry axis with the spinous process as the axis of symmetry, and a third actual key point located on the spinous process. The end of the selection part of the registration tool module selects the first actual key point and the second actual key point respectively, and its abutting part abuts against the third actual key point on the spinous process.

[0006] More specifically, the image setting plane is based on the mapping of actual key points to image key points in medical images, as follows: The image acquisition module obtains the coordinates of the first image key point, obtains the distance between the first image key point and the second image key point based on the set distance between the ends of the two point selection parts, and obtains the second image key point by combining the symmetry of the two point selection parts about the abutment part. The image acquisition module acquires a group of points at the axis of symmetry of the region to be dissected. Based on the candidate plane formed by the points in the group of points and the first and second image key points, the third image key point is obtained according to the geometric structure.

[0007] Furthermore, the process of obtaining the key points of the third image based on geometric construction specifically involves: The first direction is the direction of the line connecting the first and second image key points, and the second direction is the direction of the human head and feet. A third direction that is perpendicular to both the first and second directions is obtained. Then, a reference plane that passes through the two image key points and is parallel to the third direction is obtained. The angle between the candidate plane and the reference plane is traversed, and the point with the smallest angle among the point groups is selected as the third image key point.

[0008] Furthermore, the image acquisition module obtains the articular process vertex of the target vertebral segment on the lamina of the vertebral arch through the first deep learning network model, and uses it as the first image key point. Based on the set distance between the ends of the two point selection parts, the distance between the first image key point and the second image key point is obtained, and then a point group about the second image key point is obtained. The spinous process line, i.e. the axis of symmetry, of the region to be dissected is obtained through the second deep learning network model. The line connecting the first image key point and the point group about the second image key point is traversed. Combining the symmetry of the two point selection parts about the abutment part, the point with the largest angle between the line connecting the two point selection parts and the axis of symmetry is selected as the second image key point.

[0009] Furthermore, the spinous process line is obtained as follows: The segmented vertebrae are used as input, and local geometric features are extracted through a third deep learning network model. The spatial relationship between the point clouds of the spinous processes is obtained through a graph neural network, thereby obtaining the point set corresponding to the main axis of the spinous processes. The spinous process line is obtained by least squares line fitting.

[0010] Furthermore, the registration module calculates the transformation relationship between the registration tool module and the medical image, as follows: Select the end of the one-point selection part of the registration tool module at the first actual key point in the actual space, and abut the end of the registration tool module at the third actual key point in the actual space. The axis of the abutment part of the registration tool module is parallel to the spinous process of the target vertebra in the actual space. At this time, the contact point between the end of the other selection part of the registration tool module and the lamina of the target vertebra in the actual space is the third actual key point. The actual setting plane is obtained from these three actual key points. Based on this, the registration module, in conjunction with the image acquisition module, obtains the image key points corresponding to the aforementioned three actual key points in the medical image, and obtains the spatial mapping relationship between the actual setting plane and the image setting plane. Thus, the transformation relationship between the registration tool module and the medical image is calculated.

[0011] Furthermore, the image acquisition module obtains two articular process vertebrae on the lamina of the target vertebral segment through a first deep learning network model, takes one of them as the first image key point, obtains the distance between the first image key point and the second image key point based on the set distance between the ends of the two point selection parts, and finds the second image key point on the line connecting the two articular process vertebrae.

[0012] Furthermore, based on the constraint that the selection part of the registration tool module is symmetrical with respect to the contact part and is separated by a set distance, the corresponding position points of the two articular process vertebrae in the medical image are obtained in the actual space. One end of the selection part of the registration tool module is selected at one of the position points, and the other end of the selection part is placed along the direction from the position point to the other position point. The contact part of the registration tool module is then aligned with the spinous process of the target vertebra in the actual space, thereby obtaining the actual set plane. Based on this, the registration module combines the image key points obtained by the image acquisition module to obtain the spatial mapping relationship between the actual set plane and the image set plane, thereby calculating the transformation relationship between the registration tool module and the medical image.

[0013] Specifically, the navigation module locates the planned channels in the medical image, as follows: Based on the planned channel in the medical image, the channel entry point and channel direction are obtained. The navigation module obtains the real-time pose of the registration tool module based on the initial pose of the registration tool module and the rotation angle of the registration tool module obtained by the gyroscope module. Based on this, combined with the transformation relationship between the registration tool module and the medical image obtained by the registration module, the theoretical entry point and theoretical implantation direction of the implant are calculated. The implantation operation is guided accordingly, so that the distance between the actual entry point and the theoretical entry point is less than a set threshold, and the angle between the actual implantation direction and the theoretical implantation direction is less than a set angle.

[0014] More specifically, the implantation operation is guided by the navigation module so that the actual entry point coincides with the theoretical entry point, and the actual implantation direction coincides with the theoretical implantation direction.

[0015] More specifically, the implantation operation is guided by the navigation module, so that the Euclidean distance between the actual entry point and the theoretical entry point is less than a set threshold, and the angle between the actual implantation direction and the theoretical implantation direction is less than a set angle.

[0016] The present invention also provides a positioning and navigation method for a positioning and navigation system based on spatial mapping as described above, comprising: S1. Obtain medical images of the patient's affected area; S2. The selection part and the contact part of the registration tool module select the actual key points on the area to be dissected and determine the actual setting plane. The actual setting plane serves as the initial pose of the registration tool module. S3. The image setting plane of the area to be dissected in the medical image is obtained through the image acquisition module. The image setting plane is determined based on the correspondence between the actual key points and the image key points in the medical image. S4. The registration module calculates the transformation relationship between the registration tool module and the medical image based on the spatial mapping between the actual set plane and the image set plane. S5. The gyroscope module obtains the rotation angle of the registration tool module based on the axis formed by the point selection unit. The navigation module calculates the real-time pose of the registration tool module based on the initial pose of the registration tool module. Based on this, and combined with the transformation relationship between the registration tool module and the medical image obtained by the registration module, the planned channel in the medical image is located and navigation is performed.

[0017] Beneficial effects: This invention can transform the preoperative planned channel into the actual space, thereby completing the positioning of the planned channel and realizing positioning and navigation based on spatial mapping. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely embodiments of this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is an isometric view of the gyroscope registration tool of the present invention; Figure 2 An example diagram showing the selection of key points on the lamina of the target vertebral segment in Embodiment 1 of the present invention; Figure 3 This is an example diagram showing the selection of key points on the lamina of the target vertebral segment in Embodiment 2 of the present invention; Figure 4 An example diagram showing the determination of the spinous process tangent point of the target vertebral segment in this invention; Figure 5 Example diagram for planning channel positioning for gyroscope navigation modules; Figure 6 This is a flowchart of the spatial mapping-based positioning and navigation method of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions and advantages of the present invention clearer, the present application will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0021] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect.

[0022] This invention provides a positioning and navigation system based on spatial mapping, comprising: The registration tool module includes at least two selection parts and one contact part, wherein the selection parts are symmetrical about the contact part; the selection parts and the contact part are used to select actual key points on the area to be dissected and determine the actual setting plane, wherein the actual setting plane serves as the initial pose of the registration tool module; The image acquisition module is used to acquire the image setting plane of the area to be dissected in the medical image. The image setting plane is determined based on the correspondence between the actual key points and the image key points in the medical image. The registration module is used to calculate the transformation relationship between the registration tool module and the medical image based on the spatial mapping between the actual setting plane and the image setting plane; A gyroscope module, installed on the registration tool module, is used to obtain the angle of rotation of the registration tool module based on the axis formed by the selection part; The navigation module calculates the real-time pose of the registration tool module based on its initial pose and the rotation angle of the registration tool module obtained by the gyroscope module. Based on this, and combined with the transformation relationship between the registration tool module and the medical image obtained by the registration module, the module locates the planned channel in the medical image and performs navigation.

[0023] In this invention, the medical images acquired by the image acquisition module are three-dimensional medical images, including but not limited to CT images.

[0024] In this invention, the area to be dissected is set as the target cone segment. The actual setting plane is determined by the first and second actual key points selected by the selection unit on the target cone segment, and the third actual key point abutted by the contact unit. Correspondingly, the image key points include a first image key point, a second image key point, and a third image key point corresponding to the first, second, and third actual key points, respectively. The first and second actual key points are selected from points on the target cone segment with defined physiological characteristics, thus making it easier for the image acquisition module to obtain the coordinates of the corresponding first image key point on the medical image. Since the selection unit of the registration tool module is symmetrical about the contact unit as an axis of symmetry, as... Figure 1 As shown, the distances between the ends of at least two selection sections of the registration tool module and the contact section are equal. Therefore, the points with defined physiological characteristics mentioned here are preferably points on the target cone segment that have symmetrical points about the spinous process. In this case, the first and second actual key points are approximately symmetrical about the spinous process. The third actual key point is on the spinous process, which facilitates the coordinate calculation of the second and third image key points. The selection sections of the registration tool module are selected on the first and second actual key points respectively. The registration tool module is then rotated about the line connecting the selection sections until it contacts the spinous process; the contact point is the third actual key point.

[0025] In this invention, the selection unit selects the articular process vertex on either side as the first actual key point. When selecting the size of the registration tool module, it is necessary to consider that the set distance r between the ends of the two selection units is close to the distance between the articular process vertexes selected on the left and right sides of the spinous process. Therefore, the first image key point and the second image key point are approximately symmetrical about the spinous process. The positions of the second image key point and the third image key point calculated by the image acquisition module are closer to the second actual key point and the third actual key point, and the registration result of the registration module is more accurate.

[0026] In this invention, the image acquisition module obtains the coordinates of the first image key point through a first deep learning network model, obtains the distance between the first image key point and the second image key point based on the set distance between the ends of the two point selection parts, and obtains the second image key point by combining the symmetry of the two point selection parts about the abutment part.

[0027] The present invention provides an embodiment 1 in which the image acquisition module obtains the coordinates of the first image key point through a first deep learning network model, and obtains the distance between the first image key point and the second image key point according to the set distance between the ends of the two point selection parts, thereby obtaining a point group about the second image key point. The spinous process line, i.e. the axis of symmetry line, of the region to be dissected is obtained through the second deep learning network model. The line connecting the first image key point and the point group of the second image key point is traversed. Combining the symmetry of the two point selection parts about the abutment part, the point with the largest angle between the line connecting the point group and the axis of symmetry line is selected as the second image key point.

[0028] Specifically, on one side of the lamina of the target vertebral segment in medical imaging, the first deep learning network model, namely the 3D U-Net network, is used to automatically extract the articular process apex A on the lamina of the target vertebral segment. ct As a key image point, doctors can manually adjust it to more closely approximate the articular process apex in the image. The 3D U-Net network structure used in this invention is trained end-to-end on 3D medical images (CT scans) to automatically extract the articular process apex of the spine. The input to the 3D U-Net network is an annotated 3D image of the vertebral segment, labeled with sparse points or small region masks (dot-like annotations or small spherical masks) of the articular process apex. By introducing coordinate guidance or enhanced cropping of key regions, the 3D U-Net network can effectively learn the spatial features of the articular processes in the local morphology of the vertebral segment and automatically predict the position of its articular process apex. The output of the 3D U-Net network is a 3D probability map of the same size as the input volume, where peak points correspond to the articular process apex. Finally, post-processing (such as taking the maximum response point and non-maximum suppression) is used to obtain the coordinates of the structural articular process apex, achieving precise localization.

[0029] Based on this, the image acquisition module uses the set distance r between the ends of the two point selection parts of the registration tool module, with the articular process vertex A as the reference. ct Centered on the target vertebra, the second feature point group B can be obtained by intersecting the lamina on the other side of the target vertebra with a sphere of radius r. i The second deep learning network model, PointNet++, combined with a graph neural network (GCN), automatically identifies and extracts the main axis of the spinous process, obtaining the spinous process line n. The second feature point group B... i With the articular process apex A ct Generate m straight lines L j traverse and calculate the relationship between the spinous process line n and the straight line L. j The angle between the two points (if the calculated angle is greater than 90 degrees, subtract the angle from 180 degrees) is used to obtain the line corresponding to the maximum angle. The point in the second feature point group corresponding to this line is another image keypoint B. ct ,like Figure 2As shown (the vertebral plate selected as the key point in the example image).

[0030] Specifically, the spinous process line is obtained as follows: The segmented vertebrae are used as input, and local geometric features are extracted using PointNet++. The spatial relationship between the point clouds of the spinous processes is obtained through a graph neural network, thereby predicting the point set corresponding to the main axis of the spinous processes. The output of the graph neural network is a set of key points or ordered path points representing the main axis path of the spinous processes. By least squares straight line fitting, a smooth central main axis is obtained, which is used to represent the spatial direction and positioning reference of the spinous processes, that is, the spinous process line is obtained.

[0031] Since the distance r between the ends of the selection section of the registration tool module designed in this invention is a fixed value, when the patient's vertebral arch is small, the above embodiment cannot effectively select the second image key point on the vertebral arch. Therefore, this invention provides an embodiment two, in which the image acquisition module obtains two articular process vertebrae on the vertebral arch of the target vertebral segment through a first deep learning network model, takes one of them as the first image key point, obtains the distance between the first image key point and the second image key point according to the set distance between the ends of the two selection sections, and finds the second image key point on the line connecting the two articular process vertebrae.

[0032] Specifically, in medical imaging, the first deep learning network model, namely the 3D U-Net network, is used to automatically acquire the two articular process vertebral apexes T1 on the lamina of the target vertebral segment. ct and T2 ct The doctor can also manually adjust it to be closer to the articular process apex in the image. The method for extracting the articular process apex in this embodiment is similar to that described above. The line connecting the two articular process apexes is denoted as v, where the first key point A... ct That is, T1 ct The second key point is that on the straight line connecting v, with T1 ct The point with endpoint and length r is B. ct ,like Figure 3 As shown (example key point selected on the vertebral plate).

[0033] The image acquisition module obtains the coordinates of the first and second image key points as described above, and obtains a group of points at the axis of symmetry of the region to be dissected through a third deep learning network model. Based on the candidate plane formed by the points in the point group and the first and second image key points, the third image key point is obtained according to the geometric structure. Specifically, taking the direction of the line connecting the first and second image key points as the first direction and the direction of the human head and feet as the second direction, a third direction that is perpendicular to both the first and second directions can be obtained. Then, a reference plane that passes through the two image key points and is parallel to the third direction can be obtained. The angle between the candidate plane and the reference plane is traversed, and the point with the smallest angle in the point group is selected as the third image key point.

[0034] In this invention, the image acquisition module obtains the third image key point on the target vertebral segment based on the geometric structure. Generally, the third image key point is the spinous process tangent point. It should be noted that the definition of the spinous process tangent point is achieved by simulating the two point selection ends of the registration tool module being selected on two image key points of the pedicle of the target vertebral segment. The registration tool module is then rotated as a whole around the line connecting the two point selection ends. When rotated to a specific position, the bifurcation of the simulated registration tool module can abut against a certain point of the spinous process of the target vertebral segment. This point is the spinous process tangent point, and thus the coordinates of the spinous process tangent point in the medical image are obtained.

[0035] Specifically, the image acquisition module obtains the coordinates of the spinous process incision point in medical images as follows: The image acquisition module uses a third-party deep learning network model, PointNet++, in medical images to automatically segment and extract feature point groups at the axis of symmetry of the anatomical region, i.e., the spinous process region of the vertebral body, thus obtaining the spinous process feature point groups, such as... Figure 2 , 3 As shown, a point cloud on the surface of the spinous process of the target vertebra is selected. Any point in this point cloud, along with the first and second image key points obtained by the image acquisition module, forms a candidate plane, as shown below. Figure 4 In plane2, it can be imagined that each point in the point cloud corresponds to a candidate plane. In this invention, PointNet++ can effectively capture the local structure and geometric changes of the point cloud, and autonomously identify the spatial morphological features of the top of the spinous process through a multi-scale feature learning mechanism. Specifically, the input of the PointNet++ network is the three-dimensional point cloud data of the vertebral body, and the output is a point-level classification label of the same dimension as the input point set, marking the feature point group belonging to the top region of the spinous process. Taking the direction of the line connecting the first image key point and the second image key point as the first direction, and the direction of the human head and feet as the second direction, a third direction that is perpendicular to both the first and second directions can be obtained. Then, a reference plane that passes through the first image key point and the second image key point and is parallel to the third direction can be obtained, such as... Figure 4In plane1, the angle between the candidate plane and the reference plane is calculated. The point in the point cloud corresponding to the smallest angle is selected as the spinous process tangent point of the target vertebra, defined as C. ct ,like Figure 4 As shown.

[0036] The image acquisition module of this invention only needs to utilize a simple single conical segment's solid geometry to calculate the corresponding planar angle for each point in the point cloud once, thus filtering out the point corresponding to the smallest angle and quickly identifying the spinous process tangent point of the target vertebra. Compared to existing global iterative search methods, the computational overhead of this invention is generally lower, making it particularly suitable for processing large-scale point cloud data.

[0037] It should be noted that the gyroscope module is bound to the registration tool module. By obtaining the rotation angle of the registration tool module based on the axis formed by the selection part, the navigation module calculates the real-time pose of the registration tool module based on the initial pose of the registration tool module and the rotation angle of the registration tool module obtained by the gyroscope module. Based on the transformation relationship between the registration tool module and the medical image obtained by the registration module, the planned channel in the medical image can be transformed to the gyroscope module to complete the positioning of the planned channel.

[0038] The transformation relationship between the registration module and the medical image is calculated as follows: For Embodiment 1, the end of the point selection unit of the registration tool module is selected on the first actual key point on the lamina of the target vertebral segment in actual space, namely the articular process apex A. ct The contact part of the registration tool module is placed against the third actual key point on the target vertebral segment in actual space, namely the spinous process tangent point C. ct Furthermore, ensure that the axis of the contact part of the registration tool module is parallel to the spinous process of the target vertebra in actual space. At this point, the contact point between the end of the other selection part of the registration tool module and the lamina of the target vertebra in actual space is the third actual key point B. ct The actual setting plane can be obtained from these three actual key points. Based on this, the registration module, in conjunction with the image acquisition module, obtains the image key points corresponding to the aforementioned three actual key points in the medical image, and obtains the correspondence between the actual setting plane and the image setting plane, that is, the spatial mapping relationship between the two, thereby realizing the spatial mapping between the actual setting plane and the image setting plane, and thus calculating the transformation relationship between the registration tool module and the medical image. For Embodiment 2, the doctor needs to obtain the corresponding position points of the two articular process vertices in the medical image in the actual space based on the constraint that the selection part of the registration tool module is symmetrical with respect to the abutment part and is separated by a set distance. The doctor selects one position point with the end of one selection part of the registration tool module, and the other selection part needs to be placed in the direction pointing from the first position point to the second position point. Then, the abutment part of the registration tool module is aligned with the spinous process of the target vertebra in the actual space. This yields the actual set plane. Based on this, the registration module combines the image key points obtained by the image acquisition module to obtain the correspondence between the actual set plane and the image set plane, that is, the spatial mapping relationship between the two, thereby realizing the spatial mapping between the actual set plane and the image set plane. The transformation relationship between the registration tool module and the medical image is then calculated.

[0039] Specifically, a key image point of the target vertebral segment extracted by the image acquisition module, such as A. ct With the origin as the origin, the direction of the line connecting the first and second image key points is the y-axis, the normal of the plane formed by the three image key points is the x-axis, and the z-axis is determined by the right-hand rule to establish the image registration coordinate system. Let the transformation relationship between the image registration coordinate system and the image coordinate system be: For a point P in the image registration coordinate system... reg Transforming it to the image coordinate system, we get: ; Among them, P ct Transform a point in the image registration coordinate system to the point obtained in the image coordinate system; Similarly, the two selection sections A at the end of the registration tool module tool B tool Point the first and second actual key points on both sides of the lamina of the target vertebral segment in actual space, respectively. Then, place the end of the registration tool module against the third actual key point of the target vertebral segment, namely the spinous process tangent point. Figure 3 As shown, at this time, the end of the point selection section A of the registration tool module is selected. tool With the origin as the origin, the y-axis is the direction of the line connecting the ends of the two point selection parts of the registration tool module, the x-axis is the normal of the plane determined by the ends of the two point selection parts and the end of the contact part of the registration tool module, and the z-axis is determined by the right-hand rule to establish the gyroscope registration coordinate system. Let the transformation relationship between the gyroscope registered coordinate system and the base coordinate system corresponding to the gyroscope module be: In this context, the base coordinate system corresponding to the gyroscope module is the location of the gyroscope calibration device, which is fixed on the operating table and kept relatively constant relative to the human body during the surgery. Therefore, for a point P in the gyroscope registration coordinate system, which is also a point P in the coordinate system of the registration tool module...reg Transforming it to the image coordinate system, we get: ; Among them, P base Transform a point in the gyroscope's registered coordinate system to the point obtained in the base coordinate system corresponding to the gyroscope module; Therefore, based on the above, we can conclude that: ; ; make ,but This involves registering the gyroscope module with the medical image, which means solving the transformation relationship T between the two. x ; The specific solution is as follows: The unit vectors for the x-axis, y-axis, and z-axis of the image registration coordinate system are as follows: ; ; ; Where v1 is the normal vector of the plane formed by the three image key points of the target vertebral segment in the medical image; Then the rotation transformation of the image registration coordinate system relative to the image coordinate system can be obtained. The translation transformation of the image registration coordinate system relative to the image coordinate system is a translation transformation between their origins, that is, the origin A of the image registration coordinate system. ct Given coordinates in the image coordinate system, the transformation relationship between the image registration coordinate system and the image coordinate system can be obtained as follows: ,as follows: ; Similarly, the unit vectors of the x-axis, y-axis, and z-axis of the gyroscope's registered coordinate system are as follows: ; ; ; Where v2 is the normal vector of the plane determined by the two-point selection end and the abutment end of the registration tool module; Then we can obtain the rotational transformation of the gyroscope's registered coordinate system relative to the base coordinate system corresponding to the gyroscope module. The translation transformation of the gyroscope registered coordinate system relative to the base coordinate system corresponding to the gyroscope module is a translation transformation between their origins, that is, the origin A of the gyroscope registered coordinate system. toolGiven the coordinates in the base coordinate system corresponding to the gyroscope module, the transformation relationship between the gyroscope registered coordinate system and the base coordinate system corresponding to the gyroscope module can be obtained as follows: ,as follows: ; Therefore, the transformation relationship T between the gyroscope module and the medical image can be obtained. x This means obtaining the transformation relationship between the registration tool module and medical images.

[0040] In this invention, the navigation module locates the planned channels in medical images, as follows: Based on the planned channel in the medical image, the channel entry point and channel direction are obtained. The navigation module obtains the real-time pose of the registration tool module based on the initial pose of the registration tool module and the rotation angle of the registration tool module obtained by the gyroscope module. Based on this, combined with the transformation relationship between the registration tool module and the medical image obtained by the registration module, the theoretical entry point and theoretical implantation direction of the implant are calculated. The implantation operation is guided accordingly, so that the distance between the actual entry point and the theoretical entry point is less than a set threshold, and the angle between the actual implantation direction and the theoretical implantation direction is less than a set angle.

[0041] In this embodiment, preferably, the implantation operation is guided by a navigation module so that the actual entry point coincides with the theoretical entry point and the actual implantation direction coincides with the theoretical implantation direction.

[0042] In this invention, reference can be made to Figure 5 In medical imaging, the planned channel obtains the channel entry point and channel direction, such as... Figure 5 Green dot P in ct_in and the green straight line V ct The actual entry point and actual implantation direction of the implant are as follows: Figure 5 Red dot P in base_in And the red straight line V base ; Based on the transformation relationship between the gyroscope module and the medical image, the theoretical entry point P of the implant can be obtained. the_in and theoretical implantation direction V the ,as follows: ; ; Among them, R x This is for rotational transformation between the gyroscope module and medical images. The transformation relationship between the gyroscope module and the medical image can be calculated, where t x This is for translation transformation between the gyroscope module and the medical image; The actual entry point P of the implant can then be calculated.base_in Its theoretical entry point P the_in The Euclidean distance d between them and the actual implantation direction V of the implant base Its theoretical implantation direction V the The angle θ between them is guided by the navigation module to ensure that the Euclidean distance d is less than a set threshold and the angle θ is less than a set angle.

[0043] The present invention also provides a positioning and navigation method for the aforementioned positioning and navigation system based on spatial mapping, such as... Figure 6 As shown, the steps include: S1. Obtain medical images of the patient's affected area; S2. The selection part and the contact part of the registration tool module select the actual key points on the area to be dissected and determine the actual setting plane. The actual setting plane serves as the initial pose of the registration tool module. S3. The image setting plane of the area to be dissected in the medical image is obtained through the image acquisition module. The image setting plane is determined based on the correspondence between the actual key points and the image key points in the medical image. S4. The registration module calculates the transformation relationship between the registration tool module and the medical image based on the spatial mapping between the actual set plane and the image set plane. S5. The gyroscope module obtains the rotation angle of the registration tool module based on the axis formed by the point selection unit. The navigation module calculates the real-time pose of the registration tool module based on the initial pose of the registration tool module. Based on this, and combined with the transformation relationship between the registration tool module and the medical image obtained by the registration module, the planned channel in the medical image is located and navigation is performed.

[0044] This invention acquires key points from medical images and uses a registration tool module designed in this invention to select corresponding key points in actual space, thereby establishing a transformation relationship between the actual space and the medical images. This transformation relationship is then established using gyroscope navigation, allowing the pre-operative planned path to be transformed into the actual space, thus completing the positioning of the planned path and achieving spatial mapping-based positioning and navigation. This invention achieves rapid positioning of the planned path using only gyroscope navigation, resulting in simple and efficient positioning and navigation.

[0045] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples; within the framework of the invention, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of the invention as described above, which are not provided in the details for the sake of brevity.

[0046] The embodiments of this invention are intended to cover all such substitutions, modifications, and variations falling within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this invention should be included within the protection scope of this invention.

Claims

1. A positioning and navigation system based on spatial mapping, characterized in that, include: The registration tool module includes at least two selection parts and one contact part, wherein the selection parts are symmetrical about the contact part; the selection parts and the contact part are used to select actual key points on the area to be dissected and determine the actual setting plane, wherein the actual setting plane serves as the initial pose of the registration tool module; The image acquisition module is used to acquire the image setting plane of the area to be dissected in the medical image. The image setting plane is determined based on the correspondence between the actual key points and the image key points in the medical image. The registration module is used to calculate the transformation relationship between the registration tool module and the medical image based on the spatial mapping between the actual setting plane and the image setting plane; A gyroscope module, installed on the registration tool module, is used to obtain the angle of rotation of the registration tool module based on the axis formed by the selection part; The navigation module calculates the real-time pose of the registration tool module based on its initial pose and the rotation angle of the registration tool module obtained by the gyroscope module. Based on this, and combined with the transformation relationship between the registration tool module and the medical image obtained by the registration module, the module locates the planned channel in the medical image and performs navigation.

2. The positioning and navigation system according to claim 1, characterized in that, The actual key points on the area to be dissected include a first actual key point and a second actual key point located on the target cone symmetrical about the spinous process as an axis of symmetry, and a third actual key point located on the spinous process. The end of the selection part of the registration tool module selects the first actual key point and the second actual key point respectively, and its abutting part abuts against the third actual key point on the spinous process.

3. The positioning and navigation system according to claim 2, characterized in that, The image setting plane is based on the mapping of actual key points to image key points in medical images, as detailed below: The image acquisition module obtains the coordinates of the first image key point, obtains the distance between the first image key point and the second image key point based on the set distance between the ends of the two point selection parts, and obtains the second image key point by combining the symmetry of the two point selection parts about the abutment part. The image acquisition module acquires a group of points at the axis of symmetry of the region to be dissected. Based on the candidate plane formed by the points in the group of points and the first and second image key points, the third image key point is obtained according to the geometric structure.

4. The positioning and navigation system according to claim 3, characterized in that, The key points of the third image obtained based on geometric construction are as follows: The first direction is the direction of the line connecting the first and second image key points, and the second direction is the direction of the human head and feet. A third direction that is perpendicular to both the first and second directions is obtained. Then, a reference plane that passes through the two image key points and is parallel to the third direction is obtained. The angle between the candidate plane and the reference plane is traversed, and the point with the smallest angle among the point groups is selected as the third image key point.

5. The positioning and navigation system according to claim 3, characterized in that, The image acquisition module obtains the articular process vertex of the vertebral arch of the target vertebral segment through a first deep learning network model, and uses it as the first image key point. The distance between the first image key point and the second image key point is obtained according to the set distance between the ends of the two point selection parts, and then a point group about the second image key point is obtained. The spinous process line, i.e. the axis of symmetry, of the region to be dissected is obtained through a second deep learning network model. The line connecting the first image key point and the point group about the second image key point is traversed. Combining the symmetry of the two point selection parts about the abutment part, the point with the largest angle between the line connecting the two point groups and the axis of symmetry is selected as the second image key point.

6. The positioning and navigation system according to claim 5, characterized in that, The spinous process line is obtained as follows: The segmented vertebrae are used as input, and local geometric features are extracted through a third deep learning network model. The spatial relationship between the point clouds of the spinous processes is obtained through a graph neural network, thereby obtaining the point set corresponding to the main axis of the spinous processes. The spinous process line is obtained by least squares line fitting.

7. The positioning and navigation system according to claim 5, characterized in that, The registration module calculates the transformation relationship between the registration tool module and the medical image, as follows: Select the end of the one-point selection part of the registration tool module at the first actual key point in the actual space, and abut the end of the registration tool module at the third actual key point in the actual space. The axis of the abutment part of the registration tool module is parallel to the spinous process of the target vertebra in the actual space. At this time, the contact point between the end of the other selection part of the registration tool module and the lamina of the target vertebra in the actual space is the third actual key point. The actual setting plane is obtained from these three actual key points. Based on this, the registration module, in conjunction with the image acquisition module, obtains the image key points corresponding to the aforementioned three actual key points in the medical image, and obtains the spatial mapping relationship between the actual setting plane and the image setting plane. Thus, the transformation relationship between the registration tool module and the medical image is calculated.

8. The positioning and navigation system according to claim 3, characterized in that, The image acquisition module obtains two articular process vertebrae on the lamina of the target vertebral segment through a first deep learning network model, takes one of them as the first image key point, obtains the distance between the first image key point and the second image key point based on the set distance between the ends of the two point selection parts, and finds the second image key point on the line connecting the two articular process vertebrae.

9. The positioning and navigation system according to claim 8, characterized in that, Based on the constraint that the selection part of the registration tool module is symmetrical with respect to the contact part and is separated by a set distance, the corresponding position points of the two articular process vertebrae in the medical image are obtained in the actual space. One end of the selection part of the registration tool module is selected at one of the position points, and the other end of the selection part is placed in the direction pointing from the first position point to the second position point. The contact part of the registration tool module is then aligned with the spinous process of the target vertebra in the actual space, thereby obtaining the actual set plane. Based on this, the registration module combines the image key points obtained by the image acquisition module to obtain the spatial mapping relationship between the actual set plane and the image set plane, thereby calculating the transformation relationship between the registration tool module and the medical image.

10. The positioning and navigation system according to claim 1, characterized in that, The navigation module locates the planned channels in the medical images, as follows: Based on the planned channel in the medical image, the channel entry point and channel direction are obtained. The navigation module obtains the real-time pose of the registration tool module based on the initial pose of the registration tool module and the rotation angle of the registration tool module obtained by the gyroscope module. Based on this, combined with the transformation relationship between the registration tool module and the medical image obtained by the registration module, the theoretical entry point and theoretical implantation direction of the implant are calculated. The implantation operation is guided accordingly, so that the distance between the actual entry point and the theoretical entry point is less than a set threshold, and the angle between the actual implantation direction and the theoretical implantation direction is less than a set angle.

11. The positioning and navigation system according to claim 10, characterized in that, The implantation process is guided by a navigation module, ensuring that the actual entry point and the theoretical entry point coincide, and that the actual implantation direction and the theoretical implantation direction coincide.

12. The positioning and navigation system according to claim 10, characterized in that, The implantation process is guided by a navigation module, ensuring that the Euclidean distance between the actual and theoretical entry points is less than a set threshold, and the angle between the actual and theoretical implantation directions is less than a set angle.

13. A positioning and navigation method for a positioning and navigation system based on spatial mapping as described in any one of claims 1-12, characterized in that, include: S1. Obtain medical images of the patient's affected area; S2. The selection part and the contact part of the registration tool module select the actual key points on the area to be dissected and determine the actual setting plane. The actual setting plane serves as the initial pose of the registration tool module. S3. The image setting plane of the area to be dissected in the medical image is obtained through the image acquisition module. The image setting plane is determined based on the correspondence between the actual key points and the image key points in the medical image. S4. The registration module calculates the transformation relationship between the registration tool module and the medical image based on the spatial mapping between the actual set plane and the image set plane. S5. The gyroscope module obtains the rotation angle of the registration tool module based on the axis formed by the point selection unit. The navigation module calculates the real-time pose of the registration tool module based on the initial pose of the registration tool module. Based on this, and combined with the transformation relationship between the registration tool module and the medical image obtained by the registration module, the planned channel in the medical image is located and navigation is performed.

Images