Brain surgery image analysis method based on computer vision technology

Abstract

The invention provides a brain surgery image analysis method based on a computer vision technology, and relates to the technical field of computer vision, and the method comprises the steps: collecting a three-dimensional binocular image in a brain surgery; determining visual difference between a left eye and a right eye in the binocular camera, generating a depth map space of a three-dimensional binocular image, and further extracting a visual missing area in the brain surgery; determining an obstacle shielding domain of an obstacle in the binocular camera according to the depth map space, and performing light stripe tracking through the obstacle shielding domain and a relative position between a left eye and a right eye in the binocular camera to obtain light stripe missing information in the brain surgery; and determining a reconstruction missing threshold value of the binocular camera according to the light bar missing information and the visual missing area, performing fusion vacancy filling on the three-dimensional binocular image based on the reconstruction missing threshold value to obtain a three-dimensional fusion image, and taking the three-dimensional fusion image as an analysis image of the brain surgery. And the image integrity of the brain surgery is improved.

Description

Technical Field

[0001] The present application relates to the field of computer vision technology, and more specifically, to a method for analyzing brain surgery images based on computer vision technology. Background Art

[0002] With the rapid development of computer vision technology, applications based on image processing and analysis have gradually expanded to the medical field. During brain surgery, doctors need to rely on high-precision image information to guide surgical operations. Through binocular cameras, depth map generation, image reconstruction and other technologies in computer vision technology, accurate three-dimensional image data can be provided for brain surgery. Computer vision technology can not only capture subtle structural information, but also combine multi-dimensional data during the operation to enhance the visualization of the surgical area, greatly improve the accuracy and safety of the operation, and provide doctors with more comprehensive auxiliary decision support.

[0003] During brain surgery, some images may be missing due to occlusion, uneven lighting, etc., which may affect the doctor's comprehensive understanding of the surgical area. Through fusion and gap filling technology, the missing parts of complex brain surgery images can be completed to improve the completeness and consistency of the images. It can not only solve the data fragmentation problem caused by incomplete depth maps or visual missing areas, but also effectively improve the accuracy of surgical images, ensuring that doctors can obtain accurate three-dimensional images in real time. Therefore, how to achieve fusion and gap filling of complex brain surgery images, thereby improving the completeness of brain surgery images, is a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides a method for analyzing neurosurgery images based on computer vision technology, which can achieve the fusion and filling of complex neurosurgery images, thereby improving the completeness of neurosurgery images.

[0005] The present application provides a method for analyzing brain surgery images based on computer vision technology, the method comprising the following steps:

[0006] Acquire 3D binocular images in brain surgery using binocular cameras in computer vision;

[0007] Determine the visual difference between the left eye and the right eye in the binocular camera based on the parallax matching mechanism, generate a depth map space of the three-dimensional binocular image according to the visual difference, and extract the visual loss area in the brain surgery in the three-dimensional binocular image based on the depth map space;

[0008] Determine an obstacle occlusion domain of an obstacle in a binocular camera according to the depth map space, perform light bar tracking through the obstacle occlusion domain and the relative position between a left eye and a right eye in the binocular camera, and obtain light bar missing information in brain surgery;

[0009] The reconstruction missing threshold of the binocular camera is determined according to the light strip missing information and the visual missing area, and the three-dimensional binocular image is fused and supplemented based on the reconstruction missing threshold to obtain a three-dimensional fused image, which is then used as an analysis image for brain surgery.

[0010] Preferably, determining the visual difference between the left eye and the right eye in the binocular camera based on the disparity matching mechanism specifically includes:

[0011] Determine the baseline distance between the left and right eyes in the binocular camera;

[0012] Determine the disparity value between the left eye and the right eye in the binocular camera for each pixel in the three-dimensional binocular image based on the disparity matching mechanism and the baseline distance;

[0013] The visual difference between the left and right eyes in the binocular camera is determined based on all disparity values.

[0014] Preferably, generating a depth map space of a three-dimensional binocular image according to the visual difference specifically includes:

[0015] Get the focal length and baseline length of the binocular camera;

[0016] For each pixel in the three-dimensional binocular image, obtaining a disparity value of the pixel from the visual disparity;

[0017] Determine the disparity depth of the pixel point according to the disparity value, the focal length and the baseline length, and then obtain the disparity depth of each pixel point in the three-dimensional binocular image;

[0018] The depth map space of the three-dimensional binocular image is generated by combining all disparity depths.

[0019] Preferably, extracting the visual loss area in brain surgery in the three-dimensional binocular image based on the depth map space specifically includes:

[0020] Performing occlusion detection on the three-dimensional binocular image based on the depth map space to obtain an occlusion area in the surgical operation;

[0021] Performing invalid area detection on the three-dimensional binocular image based on the depth map space to obtain an invalid depth area in the surgical operation;

[0022] A visual loss area in brain surgery is determined according to the occlusion area and the invalid depth area.

[0023] Preferably, determining the obstacle occlusion domain of the obstacle in the binocular camera according to the depth map space specifically includes:

[0024] Performing depth jump detection on the depth map space to obtain all three-dimensional boundary point clouds of obstacles in the binocular camera;

[0025] The obstacle occlusion domain of the obstacles in the stereo camera is determined based on all 3D boundary point clouds.

[0026] Preferably, the light bar tracking is performed through the relative position between the obstacle occlusion region and the left eye and the right eye in the binocular camera to obtain the light bar missing information in the brain surgery, specifically including:

[0027] Determine the projection path of the light bar in brain surgery according to the relative position between the left eye and the right eye in the binocular camera;

[0028] Tracking the projection path of the light strip through the obstacle occlusion domain to obtain all occluded light strips;

[0029] The light bar missing information in brain surgery is determined by all the blocked light bars.

[0030] Preferably, determining the reconstruction missing threshold of the binocular camera according to the light strip missing information and the visual missing area specifically includes:

[0031] Determining the area of ​​the visual loss zone;

[0032] Determine the missing intensity in the visual missing area according to the light bar missing information;

[0033] A reconstruction missing threshold of a binocular camera is determined according to the missing intensity and the area of ​​the region.

[0034] Preferably, fusing and filling the gaps of the three-dimensional binocular images based on the reconstruction missing threshold to obtain a three-dimensional fused image specifically includes:

[0035] Screening out a fusion-filled area in the three-dimensional binocular image by using the reconstruction missing threshold;

[0036] Performing depth filling on the fused gap-filling area to obtain gap-filling depth information;

[0037] The gap-filled depth information is fused with the three-dimensional binocular image to obtain a three-dimensional fused image.

[0038] Preferably, the binocular camera is a binocular stereo camera based on the parallax principle.

[0039] Preferably, the relative positions are fixed parallel physical positions.

[0040] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects:

[0041] The three-dimensional binocular images in brain surgery are collected by a binocular camera in computer vision; the visual difference between the left eye and the right eye in the binocular camera is determined based on a disparity matching mechanism, a depth map space of the three-dimensional binocular image is generated according to the visual difference, and the visual missing area in the brain surgery is extracted from the three-dimensional binocular image based on the depth map space; the obstacle occlusion domain of the obstacle in the binocular camera is determined according to the depth map space, and light bar tracking is performed through the obstacle occlusion domain and the relative position between the left eye and the right eye in the binocular camera to obtain the light bar missing information in the brain surgery; the reconstruction missing threshold of the binocular camera is determined according to the light bar missing information and the visual missing area, and the three-dimensional binocular image is fused and supplemented based on the reconstruction missing threshold to obtain a three-dimensional fused image, and then the three-dimensional fused image is used as an analysis image of the brain surgery.

[0042] It can be seen that in the present application, the reconstruction missing threshold of the binocular camera is determined according to the light strip missing information and the visual missing area, and the three-dimensional binocular image is fused and supplemented based on the reconstruction missing threshold to obtain a three-dimensional fused image, and then the three-dimensional fused image is used as the analysis image of the brain surgery; first, the determination of the visual missing area can obtain the image blind area caused by occlusion, viewing angle limitation or invalid area during the operation, which helps to clarify the area where three-dimensional data cannot be obtained by conventional means, and then provide a clear target for subsequent image reconstruction and supplementation. Targeted missing area identification reduces unnecessary reconstruction calculations, optimizes resource allocation, ensures that key image data is supplemented, and improves the integrity of the image; then, the determination of the light strip missing information tracks the light trajectory in the occluded area through the light strip, identifies the part that cannot be imaged due to the light being blocked during the operation, and can more accurately identify the image blind spots caused by light obstruction, provide additional detail information, and combine these light strip missing data to better supplement the image missing part caused by the occlusion, thereby enhancing the continuity and integrity of the image; in summary, based on the above scheme, the fusion and supplementation of complex brain surgery images can be achieved, thereby improving the integrity of brain surgery images. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0044] Figure 1 It is a flowchart of the method for analyzing brain surgery images based on computer vision technology provided in this application.

[0045] Figure 2It is a flow chart for determining the visual loss area provided in the present application. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0047] The embodiment of the present application provides a method for analyzing images of brain surgery based on computer vision technology, the core of which is to collect three-dimensional binocular images in brain surgery through a binocular camera in computer vision; determine the visual difference between the left and right eyes in the binocular camera based on a parallax matching mechanism, generate a depth map space of the three-dimensional binocular image based on the visual difference, and extract the visual missing area in the brain surgery from the three-dimensional binocular image based on the depth map space; determine the obstacle occlusion domain of the obstacle in the binocular camera based on the depth map space, track the light stripes through the obstacle occlusion domain and the relative position between the left and right eyes in the binocular camera, and obtain the light strip missing information in the brain surgery; determine the reconstruction missing threshold of the binocular camera based on the light strip missing information and the visual missing area, fuse and fill the three-dimensional binocular image based on the reconstruction missing threshold, and obtain a three-dimensional fused image, and then use the three-dimensional fused image as the analysis image of the brain surgery. The above scheme can realize the fusion and filling of complex brain surgery images, thereby improving the integrity of the brain surgery images.

[0048] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Figure 1 As shown in FIG. 1 , this figure is an exemplary flow chart of a method for analyzing brain surgery images based on computer vision technology according to this embodiment of the present application. The analysis method includes the following steps:

[0049] In step S1, a three-dimensional binocular image in a brain surgery is acquired by a binocular camera in computer vision.

[0050] It should be noted that in the present application, the binocular camera is a camera composed of two cameras, which is used to shoot the same scene from different perspectives; the three-dimensional binocular image represents the three-dimensional image in brain surgery; in the specific implementation, a high-resolution binocular camera is selected and accurately calibrated to ensure that the relative position and angle between the left and right cameras are known. The calibrated binocular camera can be used to collect the three-dimensional image of the surgical site at the current moment, and the three-dimensional image is used as the three-dimensional binocular image in brain surgery.

[0051] In step S2, the visual difference between the left eye and the right eye in the binocular camera is determined based on the disparity matching mechanism, a depth map space of the three-dimensional binocular image is generated according to the visual difference, and the visual loss area in brain surgery is extracted from the three-dimensional binocular image based on the depth map space.

[0052] In this embodiment, the visual difference between the left eye and the right eye in the binocular camera can be determined based on the parallax matching mechanism by using the following steps:

[0053] Determine the baseline distance between the left and right eyes in the binocular camera;

[0054] Determine the disparity value between the left eye and the right eye in the binocular camera for each pixel in the three-dimensional binocular image based on the disparity matching mechanism and the baseline distance;

[0055] The visual difference between the left and right eyes in the binocular camera is determined based on all disparity values.

[0056] It should be noted that, in the present application, visual difference refers to the degree of depth change between the left eye and the right eye in the binocular camera; in specific implementation, first, after the binocular camera is installed, the physical distance between the two camera lenses is obtained by precise measurement as the baseline distance between the left eye and the right eye in the binocular camera, and the baseline distance refers to the physical distance between the left eye and the right eye in the binocular camera; then, for each pixel point in the three-dimensional binocular image, the baseline distance is used as the matching constant in the disparity matching algorithm, and the disparity matching algorithm (for example: SAD global optimization algorithm) is used to calculate the disparity value of the pixel point as the disparity value between the left eye and the right eye in the binocular camera. The disparity value between the left eye and the right eye in the binocular camera for each pixel point in the three-dimensional binocular image can be obtained in the above manner, and the disparity value represents the difference in the horizontal coordinates of the same object in the left and right images in the three-dimensional space, reflecting the depth information of the object from the camera; finally, the set of all disparity values ​​can be used as the visual difference between the left eye and the right eye in the binocular camera.

[0057] In this embodiment, the depth map space of the three-dimensional binocular image is generated according to the visual difference, which can be achieved by using the following steps:

[0058] Get the focal length and baseline length of the binocular camera;

[0059] For each pixel in the three-dimensional binocular image, obtaining a disparity value of the pixel from the visual disparity;

[0060] Determine the disparity depth of the pixel point according to the disparity value, the focal length and the baseline length, and then obtain the disparity depth of each pixel point in the three-dimensional binocular image;

[0061] The depth map space of the three-dimensional binocular image is generated by combining all disparity depths.

[0062] It should be noted that, in the present application, the depth map space represents the spatial distribution and depth information of objects in a three-dimensional scene; in specific implementation, first, the focal length and baseline length can be obtained during the camera calibration process; secondly, for each pixel in the three-dimensional binocular image, the disparity value of the pixel is obtained from the visual difference; then, the ratio of the product of the focal length and the baseline length to the disparity value can be used as the disparity depth of the pixel. The disparity depth of each pixel in the three-dimensional binocular image can be obtained in the above manner, and the disparity depth is a parameter used to indicate the relative position of objects in three-dimensional space; finally, all the disparity depths are combined according to the positions of the corresponding pixels, and the result of the combination can be used as the depth map space of the three-dimensional binocular image.

[0063] In step S3, an obstacle occlusion domain of obstacles in the binocular camera is determined according to the depth map space, and light bar tracking is performed through the obstacle occlusion domain and the relative position between the left eye and the right eye in the binocular camera to obtain light bar missing information in brain surgery.

[0064] In this embodiment, the visual loss area in brain surgery is extracted from the three-dimensional binocular image based on the depth map space. Figure 2 As described above, the figure is a flow chart of determining the visual loss area provided by the present application. In this embodiment, determining the visual loss area can be achieved by using the following steps:

[0065] In step S31, occlusion detection is performed on the three-dimensional binocular image based on the depth map space to obtain an occlusion area in the surgical operation;

[0066] In step S32, invalid area detection is performed on the three-dimensional binocular image based on the depth map space to obtain an invalid depth area in the surgical operation;

[0067] In step S33, the visual loss area in brain surgery is determined according to the occlusion area and the invalid depth area.

[0068] It should be noted that, in the present application, the visual loss area refers to the area that cannot provide a clear view during the operation; the occlusion area refers to the area that cannot be observed due to the presence of foreground objects during the operation; the invalid depth area refers to the area that cannot provide effective depth information.

[0069] In the specific implementation, first, a pair of adjacent pixels in the three-dimensional binocular image is selected as selected adjacent pixels, and the depth difference between two pixels in the selected adjacent pixels in the depth map space is compared. If the depth difference is greater than a preset adjacent depth threshold, the selected adjacent pixels are used as occluded pixels. All occluded pixels can be obtained by repeating the above steps. The image area composed of all occluded pixels can be used as the occlusion area in surgery. The adjacent depth threshold can be preset through a large number of experiments; then, the effective range of depth values ​​is preset through the performance of the binocular camera, and the image area composed of pixels corresponding to the parallax depth that is not within the effective range in the depth map space can be used as an invalid depth area; finally, the occlusion area and the invalid depth can be combined and used as the visual loss area in brain surgery.

[0070] In this embodiment, determining the obstacle occlusion domain of obstacles in the binocular camera according to the depth map space can be achieved by using the following steps:

[0071] Performing depth jump detection on the depth map space to obtain all three-dimensional boundary point clouds of obstacles in the binocular camera;

[0072] The obstacle occlusion domain of the obstacles in the stereo camera is determined based on all 3D boundary point clouds.

[0073] It should be noted that, in the present application, the obstacle occlusion domain refers to the area in the three-dimensional space that cannot be observed due to the existence of obstacles; in the specific implementation, first, the depth jump detection is performed on the depth map space using the existing jump detection algorithm, and each pixel point obtained by the depth jump detection can be used as the three-dimensional boundary point cloud of the obstacle in the binocular camera, and all the three-dimensional boundary point clouds of the obstacles in the binocular camera can be obtained, and the three-dimensional boundary point cloud represents the boundary contour of a certain object in space; then, all the three-dimensional boundary point clouds can be connected as the obstacle occlusion domain of the obstacle in the binocular camera.

[0074] In this embodiment, light bar tracking is performed by using the relative position between the left eye and the right eye in the binocular camera to obtain the light bar missing information in brain surgery, which can be achieved by the following steps:

[0075] Determine the projection path of the light bar in brain surgery according to the relative position between the left eye and the right eye in the binocular camera;

[0076] Tracking the projection path of the light strip through the obstacle occlusion domain to obtain all occluded light strips;

[0077] The light bar missing information in brain surgery is determined by all the blocked light bars.

[0078] It should be noted that, in the present application, the light strip missing information refers to the light strip information that cannot be obtained due to the obstruction of obstacles during the operation, which reflects the blind spot in the surgical field of view; in the specific implementation, first, a binocular camera field of view is initialized, and the relative position between the left eye and the right eye in the binocular camera is used as the binocular position in the binocular camera field of view, and the binocular camera field of view is used to simulate the projection behavior in brain surgery. The simulation result of the binocular camera field of view can be used as the light strip projection path in brain surgery, and the light strip projection path is the three-dimensional space path from the binocular camera to a specific target area; then, the obstacle occlusion domain is used as the projection breakpoint in the light strip projection path, and the shadow under the projection breakpoint can be used as the blocked light strip, so that all the blocked light strips can be obtained. The blocked light strip is the part of the light strip that cannot be received in certain areas due to the presence of obstacles, which affects the integrity of the surgical field of view; finally, the collection of all blocked light strips can be used as the light strip missing information in brain surgery.

[0079] In step S4, the reconstruction missing threshold of the binocular camera is determined according to the light bar missing information and the visual missing area, and the three-dimensional binocular image is fused and supplemented based on the reconstruction missing threshold to obtain a three-dimensional fused image, which is then used as an analysis image for brain surgery.

[0080] In this embodiment, the reconstruction missing threshold of the binocular camera is determined according to the light strip missing information and the visual missing area, which can be achieved by using the following steps:

[0081] Determining the area of ​​the visual loss zone;

[0082] Determine the missing intensity in the visual missing area according to the light bar missing information;

[0083] A reconstruction missing threshold of a binocular camera is determined according to the missing intensity and the area of ​​the region.

[0084] It should be noted that in the present application, the reconstruction missing threshold is a threshold used to determine the area in the three-dimensional image that needs image reconstruction and gap filling, and the reconstruction missing threshold reflects the severity of the visual loss; in specific implementation, first, the contour detection algorithm can be used to adjust the contour of the visual loss area, and the total area of ​​the contour in the visual loss area is calculated as the regional area of ​​the visual loss area. The regional area refers to the area occupied by the visual loss area in the image, which reflects the severity of the information loss in the area; then, the number of light strips in a preset time period (the most recent 30s) can be obtained, the mean of all the light strips can be used as the standard number of light strips, and the ratio of the difference between the standard number of light strips and the number of blocked light strips in the light strip missing information to the standard number of light strips can be used as the loss intensity in the visual loss area; finally, the product of the loss intensity and the regional area can be used as the reconstruction missing threshold of the binocular camera.

[0085] In this embodiment, the three-dimensional binocular images are fused and supplemented based on the reconstruction missing threshold to obtain a three-dimensional fused image, which can be achieved by the following steps:

[0086] Screening out a fusion-filled area in the three-dimensional binocular image by using the reconstruction missing threshold;

[0087] Performing depth filling on the fused gap-filling area to obtain gap-filling depth information;

[0088] The gap-filled depth information is fused with the three-dimensional binocular image to obtain a three-dimensional fused image.

[0089] It should be noted that, in the present application, the three-dimensional fused image is a complete three-dimensional image generated by supplementing and fusing the depth and texture information of the missing areas in the binocular vision image; in the specific implementation, first, all pixels in the three-dimensional binocular image are traversed, and the area with missing intensity above the reconstruction missing threshold is screened out as the fused missing area; then, the interpolation method (for example: bilinear interpolation) can be used to perform depth filling on the fused missing area, and the result of the depth filling can be used as the filling depth information, and the filling depth information is used to fill the three-dimensional depth value of the fused missing area; finally, the depth map optimization fusion algorithm can be used to fuse the filling depth information and the three-dimensional binocular image, and the fusion result can be used as the three-dimensional fused image.

[0090] It can be seen that in the present application, the reconstruction missing threshold of the binocular camera is determined according to the light strip missing information and the visual missing area, and the three-dimensional binocular image is fused and supplemented based on the reconstruction missing threshold to obtain a three-dimensional fused image, and then the three-dimensional fused image is used as the analysis image of the brain surgery; first, the determination of the visual missing area can obtain the image blind area caused by occlusion, viewing angle limitation or invalid area during the operation, which helps to clarify the area where three-dimensional data cannot be obtained by conventional means, and then provide a clear target for subsequent image reconstruction and supplementation. Targeted missing area identification reduces unnecessary reconstruction calculations, optimizes resource allocation, ensures that key image data is supplemented, and improves the integrity of the image; then, the determination of the light strip missing information tracks the light trajectory in the occluded area through the light strip, identifies the part that cannot be imaged due to the light being blocked during the operation, and can more accurately identify the image blind spots caused by light obstruction, provide additional detail information, and combine these light strip missing data to better supplement the image missing part caused by the occlusion, thereby enhancing the continuity and integrity of the image; in summary, based on the above scheme, the fusion and supplementation of complex brain surgery images can be achieved, thereby improving the integrity of brain surgery images.

[0091] The present application is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0092] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, the storage medium including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically-erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0093] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A method for analyzing images of brain surgery based on computer vision technology, characterized in that: The analytical method comprises the following steps: Acquire 3D binocular images in brain surgery using binocular cameras in computer vision; Determine the visual difference between the left eye and the right eye in the binocular camera based on the parallax matching mechanism, generate a depth map space of the three-dimensional binocular image according to the visual difference, and extract the visual loss area in the brain surgery in the three-dimensional binocular image based on the depth map space; Determine an obstacle occlusion domain of an obstacle in a binocular camera according to the depth map space, perform light bar tracking through the obstacle occlusion domain and the relative position between a left eye and a right eye in the binocular camera, and obtain light bar missing information in brain surgery; The reconstruction missing threshold of the binocular camera is determined according to the light strip missing information and the visual missing area, and the three-dimensional binocular image is fused and supplemented based on the reconstruction missing threshold to obtain a three-dimensional fused image, which is then used as an analysis image for brain surgery.

2. The method for analyzing brain surgery images based on computer vision technology according to claim 1, characterized in that: Determining the visual difference between the left and right eyes in a binocular camera based on the disparity matching mechanism specifically includes: Determine the baseline distance between the left and right eyes in the binocular camera; Determine the disparity value between the left eye and the right eye in the binocular camera for each pixel in the three-dimensional binocular image based on the disparity matching mechanism and the baseline distance; The visual difference between the left and right eyes in the binocular camera is determined based on all disparity values.

3. The method for analyzing brain surgery images based on computer vision technology according to claim 1, characterized in that: Generating a depth map space of a three-dimensional binocular image based on the visual difference specifically includes: Get the focal length and baseline length of the binocular camera; For each pixel in the three-dimensional binocular image, obtaining a disparity value of the pixel from the visual disparity; Determine the disparity depth of the pixel point according to the disparity value, the focal length and the baseline length, and then obtain the disparity depth of each pixel point in the three-dimensional binocular image; The depth map space of the three-dimensional binocular image is generated by combining all disparity depths.

4. The method for analyzing brain surgery images based on computer vision technology according to claim 1, characterized in that: Extracting the visual loss area in brain surgery in the three-dimensional binocular image based on the depth map space specifically includes: Performing occlusion detection on the three-dimensional binocular image based on the depth map space to obtain an occlusion area in the surgical operation; Performing invalid area detection on the three-dimensional binocular image based on the depth map space to obtain an invalid depth area in the surgical operation; A visual loss area in brain surgery is determined according to the occlusion area and the invalid depth area.

5. The method for analyzing brain surgery images based on computer vision technology according to claim 1, characterized in that: Determining the obstacle occlusion domain of the obstacle in the binocular camera according to the depth map space specifically includes: Performing depth jump detection on the depth map space to obtain all three-dimensional boundary point clouds of obstacles in the binocular camera; The obstacle occlusion domain of the obstacles in the stereo camera is determined based on all 3D boundary point clouds.

6. The method for analyzing brain surgery images based on computer vision technology according to claim 1, characterized in that: The light bar missing information in brain surgery is obtained by performing light bar tracking through the obstacle occlusion domain and the relative position between the left eye and the right eye in the binocular camera, specifically including: Determine the projection path of the light bar in brain surgery according to the relative position between the left eye and the right eye in the binocular camera; Tracking the projection path of the light strip through the obstacle occlusion domain to obtain all occluded light strips; The light bar missing information in brain surgery is determined by all the blocked light bars.

7. The method for analyzing brain surgery images based on computer vision technology according to claim 1, characterized in that: Determining the reconstruction missing threshold of the binocular camera according to the light strip missing information and the visual missing area specifically includes: Determining the area of ​​the visual loss zone; Determine the missing intensity in the visual missing area according to the light bar missing information; A reconstruction missing threshold of a binocular camera is determined according to the missing intensity and the area of ​​the region.

8. The method for analyzing brain surgery images based on computer vision technology according to claim 1, characterized in that: Based on the reconstruction missing threshold, the three-dimensional binocular images are fused and supplemented to obtain a three-dimensional fused image, which specifically includes: Screening out a fusion-filled area in the three-dimensional binocular image by using the reconstruction missing threshold; Performing depth filling on the fused gap-filling area to obtain gap-filling depth information; The gap-filled depth information is fused with the three-dimensional binocular image to obtain a three-dimensional fused image.

9. The method for analyzing brain surgery images based on computer vision technology according to claim 1, characterized in that: The binocular camera is a binocular stereo camera based on the parallax principle.

10. The method for analyzing brain surgery images based on computer vision technology according to claim 1, characterized in that: The relative positions are fixed parallel physical positions.

Images