Trachea running based on ct image reconstruction and bronchoscope image fusion system

Abstract

The application relates to the technical field of image processing, and discloses a CT image reconstruction and bronchoscope image fusion system based on trachea running, which first extracts the three-dimensional structure of a trachea and bronchi from a chest CT original image, generates a cavity perspective CT reconstruction image consistent with the trachea running based on a trachea center line, further acquires a bronchoscope real-time video stream, realizes spatial registration of the cavity perspective CT image and the real-time video through three-dimensional coordinate mapping, and fuses and renders the registered CT image and the bronchoscope video in a gray base and color overlay mode to obtain a fusion image with mucosa details and structural background; the system can also plan a bronchoscope insertion path based on the cavity CT image, and supports interactive and associated display among the original CT image, a three-dimensional trachea tree model and the cavity perspective CT image. The application can form a planning path before bronchoscope examination, provide real-time structural guidance in the process, and form a multi-modal fusion image after completion.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and more specifically, to a CT image reconstruction and bronchoscopic image fusion system based on the tracheal pathway. Background Technology

[0002] In current clinical interventional respiratory care, bronchoscopy is a crucial tool for obtaining information on airway lesions, performing biopsies, and conducting treatments. However, due to the complex structure of the tracheal tree, its numerous branches, and significant individual differences, physicians often rely on experience to determine turning directions and insertion depths during bronchoscopy, especially in areas with narrow or highly tortuous peripheral bronchial pathways. This increases the clinical difficulty of the procedure and limits lesion access rates. Traditional bronchoscopy only displays color images of the mucosal surface and cannot visualize the anatomical structures of the airway walls and surrounding tissues. This leads to a loss of global spatial reference during deep navigation, increasing the risk of entering the wrong branch or being unable to proceed further.

[0003] With the development of medical imaging technology, chest CT has become an important basic imaging method for assessing lung lesions. Technologies such as 3D reconstruction based on CT data, virtual bronchoscopy, and airway tree analysis are increasingly being used for bronchoscopy navigation assistance. However, traditional CT reconstructions mostly use fixed planes such as axial, coronal, and sagittal views, or achieve a holistic display of airway structures through curved surface reconstruction. These imaging perspectives differ significantly from the intraluminal perspective of an actual bronchoscope, making it difficult to provide an anatomical reference consistent with the endoscopic view. Although virtual bronchoscopy can simulate the intraluminal perspective, its images are derived from CT grayscale data, which differs significantly from the colored mucosal texture of a real bronchoscope. Doctors often need to frequently switch between the two systems, making real-time integrated reference difficult.

[0004] On the other hand, matching the location of real-time bronchoscopy video with CT images is quite challenging. Bronchoscopy imaging is dynamic, involves rapid changes in viewing angle, and is subject to unstable lighting conditions, while CT images are static grayscale images. The two differ fundamentally in spatial coordinates, imaging principles, and information representation, making direct alignment and fusion between them difficult to achieve using traditional methods.

[0005] Some existing navigation systems rely mainly on pre-calculated bronchial tree paths or 3D models to guide doctors, but they cannot provide real-time image display, nor can they dynamically present CT tissue backgrounds that match the current field of view during actual endoscopic procedures. Therefore, they still have limitations when navigating deep into complex bronchial trees.

[0006] In clinical practice, another problem exists: when interpreting raw chest CT images, physicians often cannot intuitively determine the specific anatomical location of a particular slice within the bronchoscopic view. This "perspective tomography" results in a lack of correlation between CT images and endoscopic procedures, affecting the accuracy of diagnosis and navigation. The lack of a system capable of establishing an interactive correspondence between raw CT images, a 3D model of the tracheal tree, and the endoscopic view is also a deficiency in current technology.

[0007] Therefore, there is an urgent need for a CT image reconstruction and bronchoscopic image fusion system based on the tracheal pathway to solve these problems. Summary of the Invention

[0008] The purpose of this invention is to solve the technical problems mentioned in the background art and to provide a CT image reconstruction and bronchoscopic image fusion system based on the tracheal course.

[0009] The above-mentioned objective of the present invention is achieved as follows:

[0010] A CT image reconstruction and bronchoscopic image fusion system based on the tracheal pathway includes:

[0011] The tracheal tree extraction module is used to extract the three-dimensional structural data of the trachea and bronchi of all levels from the input raw chest CT image data, including but not limited to airway radius, branch angle and branch topology, and to establish a three-dimensional tracheal tree model including the tracheal centerline.

[0012] The CT reconstruction module is connected to the tracheal tree extraction module. Using the tracheal centerline of the three-dimensional tracheal tree model as a reference, the tomographic plane angle is adaptively adjusted along the centerline direction, always perpendicular to the centerline, to generate intracavitary CT images consistent with the tracheal course.

[0013] The bronchoscopy video stream data input module is used to acquire the real-time video stream of the bronchoscopy.

[0014] Grayscale base upload module: used to upload reconstructed CT image data to the bronchoscopy system;

[0015] The image matching module is connected to the CT reconstruction module and the bronchoscopy video stream data input module, and is used to achieve spatial position matching and alignment between the intracavitary CT image and the real-time video stream according to the three-dimensional spatial coordinate mapping method.

[0016] The fusion rendering module is connected to the image matching module and is used to fuse and render the registered intracavitary CT images and the real-time video stream. The grayscale structure of the intracavitary CT images is the base layer, and the color images of the real-time video stream are overlaid as the overlay layer.

[0017] The video output module is connected to the fusion rendering module and is used to output the fused and rendered video content and generate an inspection report.

[0018] Furthermore, the navigation module, connected to the CT reconstruction module, is used to analyze the airway geometry and direction of travel based on the intracavitary CT images, and automatically generate an optimal path diagram or dynamic navigation video for bronchoscope insertion.

[0019] Furthermore, the CT re-display module is connected to the tracheal tree extraction module and the CT reconstruction module, and is used to positionally associate the three-dimensional tracheal tree model and intracavitary view CT images with the original chest CT images, so as to interactively retrieve the corresponding intracavitary view CT images on the original chest CT images.

[0020] Furthermore, the image matching module achieves spatial registration of intracavitary CT images and real-time video streams through a three-dimensional coordinate mapping method based on the airway centerline coordinate system.

[0021] Furthermore, the tracheal tree extraction module uses region growing, morphological manipulation, or a combination of both to extract three-dimensional tracheal structures from raw chest CT image data.

[0022] Furthermore, the CT reconstruction module uses multi-planar reconstruction technology, curved surface reconstruction technology, or centerline-based curved surface unfolding reconstruction technology to generate the intracavitary CT image.

[0023] The present invention also provides a method for CT image reconstruction and bronchoscopic image fusion based on the tracheal course, comprising the following steps:

[0024] S1. Extract a three-dimensional tracheal tree model and obtain the tracheal centerline based on the original chest CT image data;

[0025] S2. Using the tracheal centerline as a reconstruction reference, adaptively adjust the reconstruction plane angle along the direction of the centerline to generate intracavitary CT images.

[0026] S3. Acquire real-time video stream from the bronchoscope;

[0027] S4. Based on the three-dimensional coordinate mapping method, the intracavitary CT image and the real-time video stream are matched for anatomical position.

[0028] S5. The matched intracavitary CT images and real-time video streams are fused and rendered, with the former serving as the grayscale base layer and the latter as the color overlay layer.

[0029] S6. Output the merged and rendered video as an inspection report.

[0030] Furthermore, based on the intracavitary CT images generated in step S2, bronchoscope insertion path planning is performed and navigation information is generated.

[0031] Furthermore, the three-dimensional tracheal tree model and the intracavitary CT image are correlated with the original chest CT image in terms of coordinates, so that the original CT image can be interactively retrieved to retrieve the intracavitary CT image at the corresponding position.

[0032] The present invention also provides a computer-readable storage medium having a computer program stored thereon, the program implementing the steps of the method when executed by a processor.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] 1. The present invention automatically extracts the three-dimensional structure of the trachea and bronchi from the original chest CT images and obtains the continuous tracheal centerline. The three-dimensional structural data includes airway radius, branch angle, and branch topology, enabling the CT reconstruction plane to be dynamically adjusted according to the actual direction of airway travel, thereby generating CT images consistent with the actual bronchoscope viewpoint. This centerline-based reconstruction method overcomes the limitations of the traditional fixed plane, enabling CT images to present structural information that closely matches the actual airway path, solving the cognitive bias caused by the inconsistency between the CT viewpoint and the bronchoscope viewpoint in the prior art, and improving the accuracy of bronchoscope navigation from the essential level of the image.

[0035] 2. This invention utilizes a three-dimensional coordinate mapping method to achieve spatial alignment between intracavitary CT images and real-time bronchoscopic video streams, and employs an adjustable fusion strategy to dynamically overlay the two for display. When operating the bronchoscope, the physician uploads the reconstructed CT image to the existing bronchoscope system via a grayscale base upload module, enabling simultaneous viewing of the real-color mucosal image and the CT structural background. This preserves mucosal texture details while presenting branching structures and surrounding tissue relationships, facilitating accurate determination of bifurcation direction and insertion depth in complex bronchial trees. Compared to relying solely on traditional navigation path guidance, this invention provides a continuous, interpretable, and real-time image fusion navigation mode, significantly improving the reach rate and examination efficiency of peripheral lesions.

[0036] 3. In this system, by establishing an interactive correspondence between the original chest CT images, the three-dimensional tracheal tree model, and the endobronchial CT images, doctors can freely switch between different perspectives, maintaining positional consistency from external structures to endobronchial structures, thus comprehensively improving their understanding of anatomical relationships. This cross-perspective linkage presentation method helps clinicians plan routes before examination, confirm locations during examination, and trace lesions after examination, improving diagnostic interpretability and operational safety. Overall, this invention has achieved breakthroughs in image reconstruction methods, real-time registration strategies, and multimodal fusion display, providing a more accurate, intuitive, and stable solution for bronchoscopy navigation. Attached Figure Description

[0037] Figure 1 This is a diagram showing the overall system architecture and module interaction of the CT image reconstruction and bronchoscopic image fusion system based on the tracheal pathway in this embodiment of the invention.

[0038] Figure 2 This is a flowchart of the intracavitary CT reconstruction and image fusion system based on tracheal course CT image reconstruction and bronchoscopic image fusion in an embodiment of the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the following description is provided in conjunction with embodiments and appendices. Figure 1-2 The present invention will be further described in detail below. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0040] Example 1: This example provides a CT image reconstruction and bronchoscopic image fusion system based on tracheal pathway. The system is deployed on a hospital imaging workstation or dedicated navigation server and includes hardware resources such as a processor, memory, display, bronchoscopic video acquisition card, and network interface, as well as image processing software modules running on an operating system. These software modules are functionally divided into: a tracheal tree extraction module, a CT reconstruction module, a bronchoscopic video stream data input module, an image matching module, a fusion rendering module, a navigation module, a CT re-display module, a grayscale base upload module, and a fused video output module. Each of these modules can be implemented as an independent process, plugin, or function library, and data exchange can occur via a bus or message queue.

[0041] The tracheal tree extraction module preprocesses and segments the raw chest CT image data. First, it obtains candidate airway regions using a thresholding and region growing method, with CT values ​​used as thresholds. As a boundary, the CT value is set at... Voxels between these points are used as candidate voxels for the airway. Seed points for region growth can be selected from voxels within the airway located near the tracheal inlet. A binary mask is set as follows: Then we have:

[0042] ;

[0043] in, For a binary mask, a value of 1 indicates that the voxel belongs to the airway candidate region, and a value of 0 indicates that it is not an airway candidate region. and These represent the lower and upper thresholds for airway segmentation, respectively. The values ​​can be set empirically or automatically estimated based on specific scanning parameters and patient conditions.

[0044] After obtaining the initial binary mask, morphological opening and closing operations are used to denoise and smooth the mask. Let the structuring element be... The morphological operation results are , can be represented as:

[0045] ;

[0046] in, This represents the erosion operation. This indicates the expansion operation. This represents the closing operation; For structural elements (e.g., a 3×3×3 cube structural element); It is a morphologically processed airway mask used to reduce isolated noise and smooth airway boundaries.

[0047] Subsequently, non-airway regions were removed through connected component analysis, retaining only the largest connected component that connects to the main trachea. A distance transformation was then performed on this connected component to calculate the distance values ​​from each voxel to the airway boundary. This is used for subsequent centerline extraction. The distance transformation can be expressed as:

[0048] ;

[0049] in, voxels The minimum Euclidean distance to the airway boundary; Represents the set of boundary points of the airway region; These are the coordinates of the boundary points. Through... Finding the path with the maximum value yields the set of points along the midline of the trachea and bronchi.

[0050] Based on the distance field, the airway centerline point set is extracted using skeleton extraction or a maximum sphere-based method. Each centerline point To facilitate subsequent continuous calculations, the centerline is parameterized according to arc length, and parameters are defined. The arc length along the centerline of the trachea is used to obtain the parameter curve:

[0051] ;

[0052] in, Indicates the arc length parameter The center line point at; These represent the three coordinate components of the point in the three-dimensional coordinate system. This represents the total arc length of the centerline. It can be determined by analyzing a discrete set of points. Interpolation and cumulative arc length calculations are performed to obtain a continuous representation based on the centerline parameter curve. It can calculate the tangent vector at any point. This is used to define the direction of the trachea's course; the tangent vector can be obtained by differentiating it with respect to the centerline.

[0053] ;

[0054] in, For the arc length parameter The unit tangent vector at that location represents the local direction of the trachea at that position; The center line is respectively at The derivative of the direction. It can be numerically estimated using finite difference or spline curve derivatives, and then normalized to obtain a unit vector.

[0055] In order to construct a path perpendicular to the centerline (or perpendicular to the trachea), it is necessary to... Establish a local orthogonal coordinate system at the location ,in The tangent vector, , Two mutually orthogonal pairs that are both perpendicular to each other. Orthogonal normal vectors. This can be achieved by choosing a reference vector. And using the cross product operation, we get:

[0056] ;

[0057] in, The first normal vector is perpendicular to... And with reference vector Forming a plane; The second normal vector is equal to the tangent vector. and The cross product result; This represents the vector cross product operation. Represents the Euclidean norm of a vector. Reference vector. The orientation can be fixed (e.g., (0, 0, 1)) or dynamically adjusted based on numerical stability. In the aforementioned local coordinate system, the reconstruction plane of the intracavitary CT image can be defined as passing through the centerline point. The normal vector is The plane coordinates with world coordinates The mapping relationship between them can be represented as:

[0058] ;

[0059] in, To reconstruct the coordinates of a point on a plane in three-dimensional space; For local coordinates in the plane, respectively along and The offset in direction; These are the three-dimensional coordinates of the centerline point at that location. Uniform sampling on the plane, and Mapping back to original CT body data By performing trilinear interpolation, sequential intraluminal tomographic images along the tracheal course can be generated.

[0060] Based on the aforementioned mathematical relationships, the CT reconstruction module achieves multiplanar reconstruction or curve reconstruction of intracavitary CT images. In one implementation, multiplanar reconstruction (MPR) technology is employed, using multiple different arc length parameters. A series of tomographic images orthogonal to the trachea are generated; in another embodiment, surface reconstruction or centerline-based surface unfolding reconstruction is used to project voxels within a certain radius near the centerline onto the unfolded surface to obtain a continuous “virtual endoscopic field of view”.

[0061] The bronchoscope video stream data input module acquires the high-definition color video stream from the bronchoscope in real time via a video capture card or network interface. After decoding, the video stream is sent to the image matching module, which abstracts the imaging model of the bronchoscope lens's line of sight into a pinhole imaging model. The camera pose of the bronchoscope lens in the CT coordinate system is assumed to be... The imaging intrinsic parameter matrix is Then three-dimensional points pixel coordinates on the image plane It can be represented as:

[0062] ;

[0063] in, These are the pixel coordinates in the bronchoscopy image. It is a scaling factor; This is the intrinsic parameter matrix of the bronchoscopic camera, which includes parameters such as focal length and principal point. The camera rotation matrix describes the pose of the camera coordinate system relative to the CT coordinate system; This is a translation vector that describes the position of the origin of the camera coordinate system in the CT coordinate system. The image matching module aims to align the endobronchial CT image with the bronchoscopic video in anatomical positions, using the 3D point coordinates in the CT coordinate system. Based on the imaging model described above, 3D points in the endobronchial CT image can be projected onto the bronchoscopic image plane, and registration is performed by comparing their brightness, edges, or anatomical structures. To this end, a matching cost function can be constructed. ,For example:

[0064] ;

[0065] in, This is the registration cost function; the smaller the value, the better the registration effect. This represents the set of sampling points participating in the matching; Indicates the position of intracavitary CT images. Feature vectors at the location (e.g., grayscale, gradient, structural features, etc.); Indicating the presence of [something] in the bronchoscopy image Corresponding pixel position The feature vector at the given location is obtained through the projection relationship in the above equation; This represents the Euclidean distance between eigenvectors. Nonlinear optimization methods such as gradient descent and Levenberg-Marquardt are used to... Perform iterative optimization to make the cost function Minimize the process to achieve three-dimensional coordinate mapping and registration between intracavitary CT images and bronchoscopic video streams.

[0066] After registration, the fusion rendering module, based on the matching relationship, overlays the intracavitary CT image as a grayscale base layer and the real-time bronchoscopy video as a color overlay layer. Pseudo-color or transparency adjustments can be made to the intracavitary CT image. Let the color vector of the fused pixels be... The color vector of the bronchoscope pixels is The grayscale value of the intracavitary CT image is Then the simple alpha fusion model can be expressed as:

[0067] ;

[0068] in, The color vector of the fusion result; This is the color vector of the corresponding pixel in the bronchoscopy video; This refers to the grayscale value of the corresponding location in the intracavitary CT image; map This is a function that maps grayscale values ​​to pseudocolor or grayscale three-channel colors. To incorporate the weighting coefficients, the values ​​are set to... Between these, the display weights of bronchoscopy video and CT structural information are balanced by adjusting... and map The function can produce a fused image that takes into account both anatomical structure and realistic mucosal texture.

[0069] Based on the reconstruction and registration described above, the navigation module uses a 3D tracheal tree model and centerline data to plan the bronchoscope insertion path, assuming the set of endpoints of each branch in the tracheal tree is denoted as . Each endpoint corresponds to a potential lesion or examination target; from the tracheal inlet point To the target point The candidate path set is Each path can be represented as a discrete sequence of centerline points. Define the path cost function. for:

[0070] ;

[0071] in, For path The total cost; It represents the arc length increment between adjacent centerline points, reflecting the path length; The curvature at that location reflects the degree of path curvature; This is the airway radius at that location, reflecting the available space. The weighting coefficients for the corresponding indicators are used to balance path length, smoothness, and safety. This is achieved through a weighted shortest path search (such as Dijkstra's or A* algorithm) on the tracheal tree topology. The shortest path is used as the optimal path for bronchoscope insertion, and is displayed as a highlighted curve or arrow in the intracavitary CT images and fused videos to form a clear and intuitive navigation prompt.

[0072] The CT re-display module is used to achieve interactive association between the original chest CT images, the 3D tracheal tree model, and the intracavitary view CT images. In 3D space, the planar position and orientation of each intracavitary view CT image are determined by... The description, through the aforementioned mapping formula This allows the corresponding voxel position to be located within the original CT volume data. When a user selects a point or path on the original CT image (such as axial, sagittal, or coronal views), the nearest point to the centerline can be calculated. The corresponding arc length parameter is obtained. Then, the CT image of the intracavitary view at that location is retrieved and displayed, achieving a "from outside to inside" perspective linkage; conversely, the user can select a location in the intracavitary view image and also... By reverse-engineering its projection onto the original CT scan, a "from the inside out" visual understanding can be achieved.

[0073] The fused video output module encodes and stores the fused and rendered real-time video stream. This stream is displayed in real-time during the examination and can also be output as part of the examination report, along with navigation and lesion marking information. The examination report can be in the form of an electronic document containing video links and keyframe screenshots, or it can be output in a standard medical image archive format (such as DICOM-encapsulated video frames).

[0074] Example 2: This example is based on the method of the system in Example 1 above, and proceeds according to the following steps:

[0075] Step 1: The corresponding tracheal tree extraction module is run to complete the extraction and parameterization of the 3D tracheal tree model and tracheal centerline; this includes airway radius, branch angles, and branch topology. Regarding branch topology: Anatomically, the human tracheal tree exhibits a clear tree-like hierarchical branching structure originating from the trachea, extending progressively from level 0 to level 23. The number, diameter, and functional characteristics of each level of branch exhibit a relatively stable pattern. Grade 0 consists of the trachea itself, roughly a C-shaped tube supported by cartilage; Grade 1 consists of the left and right main bronchi, a total of 2; Grade 2 consists of lobar bronchi, usually 2 on the left and 3 on the right; Grade 3 consists of segmental bronchi, about 10 on the right and about 8 on the left; Grades 4-16 consist of thousands of bronchioles, whose structure gradually transitions from cartilage support to smooth muscle dominance; Grades 17-19 consist of terminal bronchioles, numbering in the tens of thousands, which are the final stage of airway conduction function; Grades 20-22 consist of respiratory bronchioles, numbering in the hundreds of thousands, where alveolar structures begin to appear; Grade 23 consists of alveolar sacs and alveoli, numbering in the hundreds of millions, which are the final nodes of gas exchange.

[0076] Step 2: The corresponding CT reconstruction module is used to generate intracavitary CT images that are consistent with the course of the trachea using the centerline and local orthogonal coordinate system.

[0077] Step 3: Acquire real-time endoscopic video through the bronchoscopy video stream data input module;

[0078] Step four, executed by the image matching module, optimizes the cost function based on 3D coordinate mapping and the imaging model. To achieve anatomical position registration between intracavitary CT images and real-time video streams;

[0079] Step 5: Performed by the blending rendering module, according to the blending model. map Perform image fusion display;

[0080] Step 6: Output the merged video as an inspection report through the merged video output module.

[0081] In this embodiment, the path planning and navigation information generation described above correspond to optimized path search and visualization overlay based on the above; wherein, the association between the three-dimensional tracheal tree model and intracavitary CT images with the original chest CT images is achieved through a mapping relationship. The nearest point search capability allows the operator to freely switch between the raw CT scan and the intracavitary view.

[0082] The following is an example of a specific application: The system of the present invention is used for bronchoscopic biopsy navigation of suspicious pulmonary nodular lesions. First, a thin-slice spiral CT scan is performed on the patient. A three-dimensional tracheal tree model is constructed using the tracheal tree extraction and centerline reconstruction algorithm of the present invention. Then, terminal bronchial branches connected to the suspected lesion area are automatically detected and used as target points for path planning, through a cost function. The optimal insertion path is determined, and intuitive navigation prompts are generated on the intracavitary CT images.

[0083] During the examination, the real-time bronchoscopic video and the endobronchial CT images are spatially registered and fused. Furthermore, the reconstructed CT image is uploaded to the existing bronchoscopic system via a grayscale base upload module. Doctors can simultaneously see the real endoscopic image and the CT structural background on the monitor, which helps to accurately select the turning direction in complex branches, thereby improving the lesion reach rate and biopsy success rate.

[0084] In this embodiment, the functions of each module of the present invention are implemented by computer program instructions, wherein the computer program is stored in a non-volatile computer-readable storage medium, such as a solid-state drive, disk, optical disk, flash memory, or other memory. When the program is loaded and executed by the processor, steps one to six above, as well as the navigation and re-display association steps, are performed to realize the method flow of the above embodiment.

[0085] Furthermore, it is worth noting that those skilled in the art will understand that different modules can be implemented on the same physical device or deployed on different computing nodes via a network. As long as the overall technical solution of CT image reconstruction based on tracheal course and bronchoscopy image fusion is achieved, it falls within the protection scope of this invention.

[0086] As can be seen from the above embodiments, the overall system of the present invention constitutes a multimodal image fusion and navigation platform based on the tracheal course, which is beneficial for clinicians to conduct bronchoscopy examinations intuitively, safely and efficiently.

[0087] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A CT image reconstruction and bronchoscopic image fusion system based on tracheal course, characterized in that, include: The tracheal tree extraction module is used to extract the three-dimensional structural data of the trachea and bronchi of all levels based on the input raw chest CT image data, including but not limited to airway radius, branch angle and branch topology, and to establish a three-dimensional tracheal tree model including the tracheal centerline. The CT reconstruction module, connected to the tracheal tree extraction module, is used to adaptively adjust the tomographic plane angle along the tracheal centerline of the three-dimensional tracheal tree model, always perpendicular to the centerline, to generate intracavitary CT images consistent with the tracheal course; specifically, it is based on the centerline parameter curve. Calculate the tangent vector at any point. This is used to define the direction of the trachea's course; the tangent vector can be obtained by differentiating it with respect to the centerline. ； in, For the arc length parameter The unit tangent vector at that location represents the local direction of the trachea at that position; , , The center line is respectively at , , The derivative of the direction; In each Establish a local orthogonal coordinate system at the location ,in The tangent vector, , Two mutually orthogonal pairs that are both perpendicular to each other. Orthogonal normal vectors; In the aforementioned local orthogonal coordinate system, the reconstruction plane of the intracavitary CT image can be defined as passing through the centerline point. The normal vector is The plane, its plane coordinates with world coordinates The mapping relationship between them can be represented as: ； in, To reconstruct the coordinates of a point on a plane in three-dimensional space; , Local coordinates in the plane along respectively and The offset in direction; The three-dimensional coordinates of the center line point at this location; through... Uniform sampling on the plane, and Mapping back to original CT body data By performing trilinear interpolation, sequential intraluminal tomographic images along the tracheal course can be generated. The bronchoscopy video stream data input module is used to acquire the real-time video stream of the bronchoscopy. The grayscale base upload module is used to upload the reconstructed CT image data to the bronchoscopy system; The image matching module is connected to the CT reconstruction module and the bronchoscopy video stream data input module, and is used to achieve spatial position matching and alignment between the intracavitary CT image and the real-time video stream based on the three-dimensional spatial coordinate mapping method. The fusion rendering module is connected to the image matching module and is used to fuse and render the registered intracavitary CT images and the real-time video stream. The grayscale structure of the intracavitary CT images is the base layer, and the color images of the real-time video stream are overlaid as the overlay layer. The video output module is connected to the fusion rendering module and is used to output the fused and rendered video content and generate an inspection report.

2. The CT image reconstruction and bronchoscopic image fusion system based on tracheal course according to claim 1, characterized in that, Also includes: The navigation module, connected to the CT reconstruction module, is used to analyze the airway geometry and direction of travel based on the intracavitary CT images, and automatically generate an optimal path diagram or dynamic navigation video for bronchoscope insertion.

3. The CT image reconstruction and bronchoscopic image fusion system based on tracheal course according to claim 1, characterized in that, Also includes: The CT re-display module, connected to the tracheal tree extraction module and the CT reconstruction module, is used to positionally associate the three-dimensional tracheal tree model and intracavitary view CT images with the original chest CT images, enabling interactive retrieval of intracavitary view CT images at corresponding positions on the original chest CT images.

4. The CT image reconstruction and bronchoscopic image fusion system based on tracheal course according to claim 1, characterized in that, The image matching module achieves spatial registration of intracavitary CT images and real-time video streams through a three-dimensional coordinate mapping method based on the airway centerline coordinate system.

5. The CT image reconstruction and bronchoscopic image fusion system based on tracheal course according to claim 1, characterized in that, The tracheal tree extraction module uses region growing, morphological manipulation, or a combination of both to extract three-dimensional tracheal structures from raw chest CT image data.

6. The CT image reconstruction and bronchoscopic image fusion system based on tracheal course according to claim 1, characterized in that, The CT reconstruction module uses multi-planar reconstruction technology, curved surface reconstruction technology, or centerline-based curved surface unfolding reconstruction technology to generate the intracavitary CT images.

7. A method for fusing CT image reconstruction and bronchoscopic images based on the tracheal pathway, characterized in that, Includes the following steps: S1. Extract a three-dimensional tracheal tree model and obtain the tracheal centerline based on the original chest CT image data; S2. Using the tracheal centerline as a reconstruction reference, the angle of the reconstruction plane is adaptively adjusted along the direction of the centerline to generate intracavitary CT images; specifically: Based on centerline parameter curve Calculate the tangent vector at any point. This is used to define the direction of the trachea's course; the tangent vector can be obtained by differentiating it with respect to the centerline. ； in, For the arc length parameter The unit tangent vector at that location represents the local direction of the trachea at that position; , , The center line is respectively at , , The derivative of the direction; In each Establish a local orthogonal coordinate system at the location ,in The tangent vector, , Two mutually orthogonal pairs that are both perpendicular to each other. Orthogonal normal vectors; In the aforementioned local orthogonal coordinate system, the reconstruction plane of the intracavitary CT image can be defined as passing through the centerline point. The normal vector is The plane, its plane coordinates with world coordinates The mapping relationship between them can be represented as: ； in, To reconstruct the coordinates of a point on a plane in three-dimensional space; , Local coordinates in the plane along respectively and The offset in direction; The three-dimensional coordinates of the center line point at this location; through... Uniform sampling on the plane, and Mapping back to original CT body data By performing trilinear interpolation, sequential intraluminal tomographic images along the tracheal course can be generated. S3. Acquire real-time video stream from the bronchoscope; S4. Based on the three-dimensional coordinate mapping method, the intracavitary CT image and the real-time video stream are matched for anatomical position. S5. The matched intracavitary CT images and real-time video streams are fused and rendered, with the former serving as the grayscale base layer and the latter as the color overlay layer. S6. Output the merged and rendered video as an inspection report.

8. The method according to claim 7, characterized in that, Also includes: Based on the intracavitary CT images generated in step S2, bronchoscope insertion path planning is performed and navigation information is generated.

9. The method according to claim 7, characterized in that, Also includes: The three-dimensional tracheal tree model and the intracavitary CT image are correlated with the original chest CT image in terms of coordinates, so that the original CT image can be interactively retrieved to retrieve the corresponding intracavitary CT image.

10. A computer-readable storage medium, characterized in that, It contains a computer program that, when executed by a processor, implements the steps of the method described in any one of claims 7 to 9.

Images