A minimally invasive surgery planning method and system based on image navigation

Abstract

This invention discloses a minimally invasive surgical planning method and system based on image navigation, belonging to the field of medical information technology. Specifically, it includes: acquiring a three-dimensional model of the patient through CT scanning; performing structural analysis on the three-dimensional model and dividing the organ region into meshes; calculating the grayscale differences between meshes; identifying lesion meshes; calculating the shortest distance from any in vitro candidate point to the geometric center of the lesion region based on the three-dimensional coordinates of the skin region and the lesion mesh; calculating the optimal score of any in vitro candidate point based on the three-dimensional Euclidean distance; selecting the optimal puncture point; and generating an in vivo and in vitro candidate point set through orthogonal projection and vertical extension. This process is repeated until any puncture point reaches the lesion region. Finally, all puncture points are connected to obtain the final puncture path. This invention automatically obtains the optimal ablation path through iterative optimization, improving the efficiency of liver tumor ablation.

Description

Technical Field

[0001] This invention relates to the field of medical information technology, specifically to a method and system for planning minimally invasive surgery based on image navigation. Background Technology

[0002] Thermal ablation therapy for liver tumors (including radiofrequency ablation and microwave ablation) has become the third major radical treatment after surgical resection and liver transplantation. Under medical image guidance, an ablation needle (radiofrequency electrode / microwave antenna) is inserted into the tumor as a heat source, heating the tumor tissue to a high temperature, causing in-situ coagulative necrosis of the tumor.

[0003] In the treatment of liver cancer, radiofrequency ablation of liver tumors is a commonly used technique to ablate the tumor target area. In order to improve the ablation accuracy and thus ensure treatment efficiency, it is necessary to locate and analyze the patient's tumor target area in order to pre-design the puncture path. At present, the puncture path planning is based on the treatment doctor's general determination of the required number of ablations, needle entry points and puncture depth according to the patient's examination images and personal experience. This can easily lead to an excessive ablation area, causing irreversible damage to the patient's body. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for planning minimally invasive surgery based on image navigation, and to solve the following technical problems:

[0005] In the treatment of liver cancer, radiofrequency ablation of liver tumors is commonly used to ablate the tumor target area. To improve the accuracy of ablation, precise localization and analysis of the tumor target area are necessary, along with the design of a reasonable puncture path. Currently, puncture path planning relies on physicians determining ablation parameters based on examination images and experience, which may lead to an excessively large ablation area and increase the risk of harm to the patient.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A method for planning minimally invasive surgery based on image navigation, comprising the following steps:

[0008] S1, continuously scan the patient's scan area using dynamic CT technology to obtain several detection images, and construct a three-dimensional model based on the detection images;

[0009] S2, based on the medical image segmentation algorithm, performs structural analysis on the 3D model to obtain the skin, organ, blood vessel and rib regions. The blood vessel and rib regions are marked as danger regions, and the organ regions are divided into grids. If the gray value of any grid region is within the preset gray value threshold range, the grid is marked as a lesion grid. The region composed of all lesion grids is marked as a lesion region.

[0010] S3. Construct an in vitro candidate point set based on the skin region, and use the Euclidean distance transformation algorithm to calculate the shortest distance DF1 from each in vitro candidate point to the danger zone; analyze the vertex coordinates of the lesion region and use the centroid method to calculate the geometric center coordinates of the lesion region to obtain the shortest distance DF2 from each in vitro candidate point to the geometric center; calculate the optimization score of each in vitro candidate point based on DF1 and DF2, and select the in vitro candidate point with the highest optimization score as the initial puncture point;

[0011] S4. Based on the three-dimensional coordinates of any in vitro candidate point set, perform orthogonal projection on any in vitro candidate point to obtain the corresponding surface projection point, and extend the projection point into the body along the normal vector direction of the projection point by a preset vertical interval distance d to generate an in vivo candidate point set that has a spatial correspondence with the in vitro candidate point. Repeat the above process for any in vivo candidate point to obtain the second puncture point. Repeat the above process until the first puncture point reaches the lesion area, then connect all puncture points to obtain the final puncture path.

[0012] As a further aspect of the present invention: in S1, the specific construction process of the three-dimensional model is as follows:

[0013] A rigid registration algorithm based on feature point matching is used to register the detection images. The registered detection images are then stacked and reconstructed according to a preset structural order. Isotropic interpolation is then used to form a three-dimensional volume dataset with a unified coordinate system. A three-dimensional view with depth information is generated by simulating the optical penetration path using the ray projection method. Simultaneously, threshold segmentation and the Marching Cubes algorithm are combined to extract isosurfaces with specific gray values, thereby constructing a three-dimensional model.

[0014] As a further aspect of the present invention: in step S3, the calculation process for the shortest distance is as follows:

[0015] For any hazardous area, a spatial accelerated mesh is constructed to obtain all the patches of the hazardous area. The shortest distance from the candidate point outside the body to each patch is calculated. The minimum value of the distances of all patches is selected as the undetermined distance value. The undetermined distance values ​​corresponding to all hazardous areas are obtained. The minimum value among the undetermined distance values ​​is selected as the shortest distance.

[0016] As a further aspect of the present invention: in step S3, the specific calculation process of the preferred fraction is as follows:

[0017] ;

[0018] Where a and b are preset unit coefficients, DF1i is the shortest distance from the i-th in vitro candidate point to the danger zone, MaxDF1 is the maximum value of the shortest distances from all in vitro candidate points within the skin region to the danger zone, DF2i is the shortest distance from the i-th in vitro candidate point to the geometric center, and MaxDF2 is the maximum value of the shortest distances from all in vitro candidate points within the skin region to the geometric center.

[0019] As a further aspect of the present invention: in step S2, the process of setting the preset grayscale threshold is as follows:

[0020] A number of CT image samples of liver tumors are obtained, the tumor regions on the CT image samples are manually labeled and the gray values ​​of the tumor regions are obtained, the maximum and minimum values ​​of all gray values ​​are obtained, and the gray threshold range is obtained based on the maximum and minimum values.

[0021] As a further aspect of the present invention: in step S3, if there are two or more candidate points in vitro with equal and maximum scores, they are all marked as initial puncture points and corresponding puncture paths are generated respectively.

[0022] As a further aspect of the present invention, it also includes connecting the initial puncture point and the second puncture point to obtain a path direction vector v, calculating the angle θ between the path direction vector v and the skin region normal vector n, and if the angle θ is greater than or equal to a preset angle threshold, then the second puncture point is eliminated, and a second puncture point is reselected according to the preferred score and a corresponding puncture path is generated.

[0023] As a further aspect of the present invention: S4 further includes, when there are several puncture paths, obtaining the length of each puncture path and selecting the shortest puncture path as the final puncture path.

[0024] A minimally invasive surgical planning system based on image navigation, comprising:

[0025] The data acquisition module is used to continuously scan the patient's scan area using dynamic CT technology to obtain several detection images, and to construct a three-dimensional model based on the detection images.

[0026] The structural analysis module is used to perform structural analysis on the 3D model based on medical image segmentation algorithms to obtain the skin, organ, blood vessel and rib regions. The blood vessel and rib regions are marked as danger areas, and the organ regions are divided into grids. If the gray value of any grid region is within the preset gray value threshold range, the grid is marked as a lesion grid. The region composed of all lesion grids is marked as a lesion region.

[0027] The initial determination module is used to construct an in vitro candidate point set based on the skin region, and use the Euclidean distance transformation algorithm to calculate the shortest distance DF1 from each in vitro candidate point to the danger zone; it analyzes the vertex coordinates of the lesion region and uses the centroid method to calculate the geometric center coordinates of the lesion region, obtaining the shortest distance DF2 from each in vitro candidate point to the geometric center; it calculates the optimization score of each in vitro candidate point based on DF1 and DF2, and selects the in vitro candidate point with the highest optimization score as the initial puncture point;

[0028] The path planning module is used to perform orthogonal projection on any in vitro candidate point based on the three-dimensional coordinates of any in vitro candidate point set to obtain the corresponding surface projection point, and extend it into the body along the normal vector direction of the projection point by a preset vertical interval distance d, generating an in vivo candidate point set that has a spatial correspondence with the in vitro candidate point. The above process is repeated for any in vivo candidate point to obtain the second puncture point; the above process is repeated until the first puncture point reaches the lesion area, then all puncture points are connected to obtain the final puncture path.

[0029] The beneficial effects of this invention are:

[0030] This invention first utilizes dynamic CT scanning to generate a three-dimensional model of the patient's body. Medical image segmentation technology is then used to analyze areas such as skin, organs, blood vessels, and ribs, identifying danger zones to avoid damage to critical tissues during surgery. Organ areas are divided into grids and analyzed for grayscale. Combined with preset thresholds, lesion areas are automatically identified, reducing subjective errors from manual interpretation. The shortest distance from in vitro candidate points to danger zones and the lesion center is calculated. This can be understood as measuring the spatial distance between in vitro candidate points and danger zones, avoiding critical structures such as blood vessels and ribs; the smaller the spatial distance, the greater the actual risk. The shortest distance from in vitro candidate points to the lesion center reflects the proximity of the candidate points to the lesion center; a smaller spatial distance reflects the efficiency of the puncture point in approaching the target area, meaning higher positioning accuracy and better precision in reaching the lesion core. An optimal score is calculated based on the spatial distance, and the best initial puncture point is dynamically selected based on this score. Furthermore, in vivo candidate points are generated through orthogonal projection and normal vector extension, iteratively generating a puncture path to ensure the path avoids danger zones and directly reaches the lesion target. This invention automatically obtains the optimal ablation path through repeated iterative optimization, improving the efficiency of liver tumor ablation. Attached Figure Description

[0031] The invention will now be further described with reference to the accompanying drawings.

[0032] Figure 1 This is a schematic diagram of a minimally invasive surgical planning method based on image navigation according to the present invention. Detailed Implementation

[0033] 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.

[0034] Please see Figure 1 As shown, this invention is a minimally invasive surgical planning method based on image navigation, comprising the following steps:

[0035] S1, continuously scan the patient's scan area using dynamic CT technology to obtain several detection images, and construct a three-dimensional model based on the detection images;

[0036] S2, based on the medical image segmentation algorithm, performs structural analysis on the 3D model to obtain the skin, organ, blood vessel and rib regions. The blood vessel and rib regions are marked as danger regions, and the organ regions are divided into grids. If the gray value of any grid region is within the preset gray value threshold range, the grid is marked as a lesion grid. The region composed of all lesion grids is marked as a lesion region.

[0037] S3. Construct an in vitro candidate point set based on the skin region, and use the Euclidean distance transformation algorithm to calculate the shortest distance DF1 from each in vitro candidate point to the danger zone; analyze the vertex coordinates of the lesion region and use the centroid method to calculate the geometric center coordinates of the lesion region to obtain the shortest distance DF2 from each in vitro candidate point to the geometric center; calculate the optimization score of each in vitro candidate point based on DF1 and DF2, and select the in vitro candidate point with the highest optimization score as the initial puncture point;

[0038] S4. Based on the three-dimensional coordinates of any in vitro candidate point set, perform orthogonal projection on any in vitro candidate point to obtain the corresponding surface projection point, and extend the projection point into the body along the normal vector direction of the projection point by a preset vertical interval distance d to generate an in vivo candidate point set that has a spatial correspondence with the in vitro candidate point. Repeat the above process for any in vivo candidate point to obtain the second puncture point. Repeat the above process until the first puncture point reaches the lesion area, then connect all puncture points to obtain the final puncture path.

[0039] This invention first utilizes dynamic CT scanning to generate a three-dimensional model of the patient's body. Medical image segmentation technology is then used to analyze areas such as skin, organs, blood vessels, and ribs, identifying danger zones to avoid damage to critical tissues during surgery. Organ areas are divided into grids and analyzed for grayscale. Combined with preset thresholds, lesion areas are automatically identified, reducing subjective errors from manual interpretation. The shortest distance from in vitro candidate points to danger zones and the lesion center is calculated. This can be understood as measuring the spatial distance between in vitro candidate points and danger zones, avoiding critical structures such as blood vessels and ribs; the smaller the spatial distance, the greater the actual risk. The shortest distance from in vitro candidate points to the lesion center reflects the proximity of the candidate points to the lesion center; a smaller spatial distance reflects the efficiency of the puncture point in approaching the target area, meaning higher positioning accuracy and better precision in reaching the lesion core. An optimal score is calculated based on the spatial distance, and the best initial puncture point is dynamically selected based on this score. Furthermore, in vivo candidate points are generated through orthogonal projection and normal vector extension, iteratively generating a puncture path to ensure the path avoids danger zones and directly reaches the lesion target. This invention automatically obtains the optimal ablation path through repeated iterative optimization, improving the efficiency of liver tumor ablation.

[0040] In a preferred embodiment of the present invention, the specific construction process of the three-dimensional model in S1 is as follows:

[0041] A rigid registration algorithm based on feature point matching is used to register the detection images. The registered detection images are then stacked and reconstructed according to a preset structural order. Isotropic interpolation is then used to form a three-dimensional volume dataset with a unified coordinate system. A three-dimensional view with depth information is generated by simulating the optical penetration path using the ray projection method. Simultaneously, threshold segmentation and the Marching Cubes algorithm are combined to extract isosurfaces with specific gray values, thereby constructing a three-dimensional model.

[0042] A rigid registration algorithm based on feature point matching is used to register the detection images to ensure accurate alignment of images acquired at different times or angles, avoiding deviations in subsequent model construction due to image misalignment. The registered images are then stacked and reconstructed according to a pre-defined structural order, allowing the two-dimensional images to be combined layer by layer to form a preliminary three-dimensional structural framework, providing a foundation for subsequent processing. Isotropic interpolation eliminates resolution differences in different directions, ensuring consistent accuracy across all dimensions of the three-dimensional volume dataset and improving the spatial accuracy of the model. A 3D view with depth information is generated by simulating optical penetration paths using ray casting, visually presenting the three-dimensional structure of the scanned area and facilitating subsequent observation and analysis of specific tissues or regions. Simultaneously combining threshold segmentation and the Marching Cubes algorithm to extract isosurfaces of specific grayscale values ​​accurately separates the boundaries and contours of the target tissue, removing irrelevant background interference, thereby constructing a clear and accurate 3D model reflecting the anatomical structure of the patient's scanned area. The overall purpose of this series of operations is to provide a high-precision, high-reliability 3D visualization model for minimally invasive surgical planning, enabling surgeons to clearly understand the spatial relationships of the patient's skin, organs, blood vessels, ribs, and other structures before surgery.

[0043] In a preferred embodiment of the present invention, the calculation process for the shortest distance in step S3 is as follows:

[0044] For any hazardous area, a spatial accelerated mesh is constructed to obtain all the patches of the hazardous area. The shortest distance from the candidate point outside the body to each patch is calculated. The minimum value of the distances of all patches is selected as the undetermined distance value. The undetermined distance values ​​corresponding to all hazardous areas are obtained. The minimum value among the undetermined distance values ​​is selected as the shortest distance.

[0045] It is understandable that directly calculating the distance from a point to a complex 3D hazardous area is cumbersome and computationally intensive. However, by constructing a spatial acceleration mesh, the hazardous area can be discretized into multiple patches, transforming the complex 3D distance calculation into the distance calculation from a point to each patch. This significantly reduces computational complexity and workload, improves computational efficiency, calculates the shortest distance from the candidate point outside the body to each patch, and takes the minimum value as the undetermined distance value. This ensures that the distance calculation result for each hazardous area accurately reflects the true shortest distance from the point to the area. Then, the minimum value is selected from the undetermined distance values ​​of all hazardous areas as the final shortest distance DF1, which guarantees that this distance value is the safest distance from the candidate point outside the body to all hazardous areas.

[0046] In a preferred embodiment of the present invention, the specific calculation process of the preferred fraction in S3 is as follows:

[0047] ;

[0048] Where a and b are preset unit coefficients, DF1i is the shortest distance from the i-th in vitro candidate point to the danger zone, MaxDF1 is the maximum value of the shortest distances from all in vitro candidate points within the skin region to the danger zone, DF2i is the shortest distance from the i-th in vitro candidate point to the geometric center, and MaxDF2 is the maximum value of the shortest distances from all in vitro candidate points within the skin region to the geometric center.

[0049] In a preferred embodiment of the present invention, the process of setting the preset grayscale threshold in S2 is as follows:

[0050] A number of CT image samples of liver tumors are obtained, the tumor regions on the CT image samples are manually labeled and the gray values ​​of the tumor regions are obtained, the maximum and minimum values ​​of all gray values ​​are obtained, and the gray threshold range is obtained based on the maximum and minimum values.

[0051] In a preferred embodiment of the present invention, in step S3, if two or more in vitro candidate points have the same preferred score and are all the maximum value, they are all marked as initial puncture points and corresponding puncture paths are generated respectively.

[0052] When multiple candidate points in vitro have equal and maximum optimization scores, these points are all designated as initial puncture points, and corresponding puncture paths are generated for each. This is because equal optimization scores indicate that these candidate points are at an equally optimal level in terms of both avoiding dangerous areas (represented by the shortest distance DF1) and approaching the geometric center of the lesion area (represented by the shortest distance DF2), and a single criterion cannot further differentiate their merits. Retaining all equally optimal initial puncture points and generating multiple puncture paths avoids overlooking potentially more flexible or adaptable paths due to rigidly selecting a single puncture point. The advantage of this approach is that it provides multiple alternatives for surgical planning, increasing the flexibility and tolerance for errors in path selection.

[0053] In a preferred embodiment of the present invention, the method further includes connecting the initial puncture point and the second puncture point to obtain a path direction vector v, calculating the angle θ between the path direction vector v and the skin region normal vector n, and if the angle θ is greater than or equal to a preset angle threshold, then the second puncture point is removed, and a second puncture point is reselected according to the preferred score and a corresponding puncture path is generated.

[0054] If the included angle θ is too large, it will cause the puncture direction to be too tilted, increasing the risk of the path deviating from the predetermined trajectory during the puncture. It may even make it difficult to accurately enter the target area or damage tissues along the way due to an unfavorable puncture angle. By setting a preset included angle threshold to screen the path direction, it is possible to ensure that the puncture path forms a reasonable angle with the skin surface, making the puncture operation more in line with the mechanical properties of human anatomy and the convenience of surgical operation. Angle constraints improve the rationality and operability of the puncture path, avoiding puncture difficulties or increased risks caused by abnormal angles. This ensures the feasibility of the puncture path from a directional perspective, ensuring that the direction of each puncture path meets the basic angle requirements of surgical operation. By strictly controlling the rationality of the puncture direction angle, intraoperative adjustments or operational errors caused by improper path direction are reduced, making the entire puncture path both safe and efficient in terms of spatial distance.

[0055] In a preferred embodiment of the present invention, step S3 further includes obtaining the length of each puncture path when there are multiple puncture paths, and selecting the shortest puncture path as the final puncture path.

[0056] By comparing the lengths of various pathways and selecting the shortest one, the efficiency of the puncture path can be optimized, unnecessary tissue penetration and operational steps can be reduced, and the minimally invasiveness and economy of the surgery can be further improved while ensuring the safety and feasibility of the pathway. By quantifying the key indicator of pathway length, the optimal solution can be selected from multiple compliant pathways, giving the surgical planning an additional advantage of operational efficiency in addition to safety and precision. By shortening the puncture path, the probability of complications such as intraoperative bleeding and infection can be reduced.

[0057] A minimally invasive surgical planning system based on image navigation, comprising:

[0058] The data acquisition module is used to continuously scan the patient's scan area using dynamic CT technology to obtain several detection images, and to construct a three-dimensional model based on the detection images.

[0059] The structural analysis module is used to perform structural analysis on the 3D model based on medical image segmentation algorithms to obtain the skin, organ, blood vessel and rib regions. The blood vessel and rib regions are marked as danger areas, and the organ regions are divided into grids. If the gray value of any grid region is within the preset gray value threshold range, the grid is marked as a lesion grid. The region composed of all lesion grids is marked as a lesion region.

[0060] The initial determination module is used to construct an in vitro candidate point set based on the skin region, and use the Euclidean distance transformation algorithm to calculate the shortest distance DF1 from each in vitro candidate point to the danger zone; it analyzes the vertex coordinates of the lesion region and uses the centroid method to calculate the geometric center coordinates of the lesion region, obtaining the shortest distance DF2 from each in vitro candidate point to the geometric center; it calculates the optimization score of each in vitro candidate point based on DF1 and DF2, and selects the in vitro candidate point with the highest optimization score as the initial puncture point;

[0061] The path planning module is used to perform orthogonal projection on any in vitro candidate point based on the three-dimensional coordinates of any in vitro candidate point set to obtain the corresponding surface projection point, and extend it into the body along the normal vector direction of the projection point by a preset vertical interval distance d, generating an in vivo candidate point set that has a spatial correspondence with the in vitro candidate point. The above process is repeated for any in vivo candidate point to obtain the second puncture point; the above process is repeated until the first puncture point reaches the lesion area, then all puncture points are connected to obtain the final puncture path.

[0062] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for planning minimally invasive surgery based on image navigation, characterized in that, Includes the following steps: S1, continuously scan the patient's scan area using dynamic CT technology to obtain several detection images, and construct a three-dimensional model based on the detection images; S2, based on the medical image segmentation algorithm, performs structural analysis on the 3D model to obtain the skin, organ, blood vessel and rib regions. The blood vessel and rib regions are marked as danger regions, and the organ regions are divided into grids. If the gray value of any grid region is within the preset gray value threshold range, the grid is marked as a lesion grid. The region composed of all lesion grids is marked as a lesion region. S3. Construct an in vitro candidate point set based on the skin region, and use the Euclidean distance transformation algorithm to calculate the shortest distance DF1 from each in vitro candidate point to the danger zone; analyze the vertex coordinates of the lesion region and use the centroid method to calculate the geometric center coordinates of the lesion region to obtain the shortest distance DF2 from each in vitro candidate point to the geometric center; calculate the optimization score of each in vitro candidate point based on DF1 and DF2, and select the in vitro candidate point with the highest optimization score as the initial puncture point; S4. Based on the three-dimensional coordinates of any in vitro candidate point set, perform orthogonal projection on any in vitro candidate point to obtain the corresponding surface projection point, and extend the projection point into the body along the normal vector direction of the projection point by a preset vertical interval distance d to generate an in vivo candidate point set that has a spatial correspondence with the in vitro candidate point. Repeat the above process for any in vivo candidate point to obtain the second puncture point. Repeat the above process until the first puncture point reaches the lesion area, then connect all puncture points to obtain the final puncture path.

2. The image-guided minimally invasive surgical planning method according to claim 1, characterized in that, In S1, the specific construction process of the three-dimensional model is as follows: A rigid registration algorithm based on feature point matching is used to register the detection images. The registered detection images are then stacked and reconstructed according to a preset structural order. Isotropic interpolation is then used to form a three-dimensional volume dataset with a unified coordinate system. A three-dimensional view with depth information is generated by simulating the optical penetration path using the ray projection method. The isosurfaces with specific gray values ​​are extracted by combining threshold segmentation and the Marching Cubes algorithm, thereby constructing a three-dimensional model.

3. The image-guided minimally invasive surgical planning method according to claim 1, characterized in that, In S3, the calculation process for the shortest distance is as follows: For any hazardous area, a spatial accelerated mesh is constructed to obtain all the patches of the hazardous area. The shortest distance from the candidate point outside the body to each patch is calculated. The minimum value of the distances of all patches is selected as the undetermined distance value. The undetermined distance values ​​corresponding to all hazardous areas are obtained. The minimum value among the undetermined distance values ​​is selected as the shortest distance.

4. The image-guided minimally invasive surgical planning method according to claim 1, characterized in that, In S3, the specific calculation process for the preferred score is as follows: ； Where a and b are preset unit coefficients, DF1i is the shortest distance from the i-th in vitro candidate point to the danger zone, MaxDF1 is the maximum value of the shortest distances from all in vitro candidate points within the skin region to the danger zone, DF2i is the shortest distance from the i-th in vitro candidate point to the geometric center, and MaxDF2 is the maximum value of the shortest distances from all in vitro candidate points within the skin region to the geometric center.

5. The image-guided minimally invasive surgical planning method according to claim 1, characterized in that, In step S2, the process of setting the preset grayscale threshold is as follows: A number of CT image samples of liver tumors are obtained, the tumor regions on the CT image samples are manually labeled and the gray values ​​of the tumor regions are obtained, the maximum and minimum values ​​of all gray values ​​are obtained, and the gray threshold range is obtained based on the maximum and minimum values.

6. The image-guided minimally invasive surgical planning method according to claim 1, characterized in that, In step S3, if two or more candidate in vitro points have the same optimal score and are all the maximum value, they are all marked as initial puncture points and corresponding puncture paths are generated respectively.

7. The image-guided minimally invasive surgical planning method according to claim 6, characterized in that, It also includes connecting the initial puncture point and the second puncture point to obtain the path direction vector v, calculating the angle θ between the path direction vector v and the skin region normal vector n, and if the angle θ is greater than or equal to a preset angle threshold, then the second puncture point is eliminated, and a second puncture point is reselected according to the preferred score and a corresponding puncture path is generated.

8. The image-guided minimally invasive surgical planning method according to claim 1, characterized in that, S4 further includes, when there are several puncture paths, obtaining the length of each puncture path and selecting the shortest puncture path as the final puncture path.

9. A minimally invasive surgical planning system based on image navigation, characterized in that, include: The data acquisition module is used to continuously scan the patient's scan area using dynamic CT technology to obtain several detection images, and to construct a three-dimensional model based on the detection images. The structural analysis module is used to perform structural analysis on the 3D model based on medical image segmentation algorithms to obtain the skin, organ, blood vessel and rib regions. The blood vessel and rib regions are marked as danger areas, and the organ regions are divided into grids. If the gray value of any grid region is within the preset gray value threshold range, the grid is marked as a lesion grid. The region composed of all lesion grids is marked as a lesion region. The initial determination module is used to construct an in vitro candidate point set based on the skin region, and use the Euclidean distance transformation algorithm to calculate the shortest distance DF1 from each in vitro candidate point to the danger zone; it analyzes the vertex coordinates of the lesion region and uses the centroid method to calculate the geometric center coordinates of the lesion region, obtaining the shortest distance DF2 from each in vitro candidate point to the geometric center; it calculates the optimization score of each in vitro candidate point based on DF1 and DF2, and selects the in vitro candidate point with the highest optimization score as the initial puncture point; The path planning module is used to perform orthogonal projection on any in vitro candidate point based on the three-dimensional coordinates of any in vitro candidate point set to obtain the corresponding surface projection point, and extend it into the body along the normal vector direction of the projection point by a preset vertical interval distance d, generating an in vivo candidate point set that has a spatial correspondence with the in vitro candidate point. The above process is repeated for any in vivo candidate point to obtain the second puncture point; the above process is repeated until the first puncture point reaches the lesion area, then all puncture points are connected to obtain the final puncture path.

Images