Point cloud processing method and apparatus, electronic device, and storage medium

Abstract

The present application provides a point cloud processing method and device, electronic equipment and storage medium, and relates to the field of oral scanning data processing. The method comprises the following steps: acquiring an original oral point cloud, and determining an implant rod target point cloud from the original oral point cloud according to the spatial connectivity, color information and normal information of each sampling point in the original oral point cloud; performing directional bounding box normal segmentation on the implant rod target point cloud to obtain an implant rod pattern surface point cloud; and obtaining an implant rod pure pattern feature point cloud based on the implant rod pattern surface point cloud, thereby improving the feature stability, noise resistance and convergence speed of the subsequent registration algorithm, greatly reducing the calculation cost while ensuring the clinical precision, and meeting the rigid requirements of the oral implant navigation system for real-time performance, robustness and no manual intervention.

Description

Technical Field

[0001] This invention relates to the field of oral cavity scan data processing, and more specifically, to a point cloud processing method, apparatus, electronic device, and storage medium. Background Technology

[0002] In digital dental implant navigation and personalized surgical guide design, high-precision extraction of the geometric features of the implant bar is a key prerequisite for achieving implant spatial positioning, occlusal relationship reconstruction, and automatic surgical guide registration. Current mainstream clinical workflows typically rely on dental CBCT or intraoral scanners to acquire three-dimensional point cloud data of the patient's oral cavity. However, the original point clouds generally suffer from problems such as high noise, uneven sampling, close proximity of multiple components (crown, gingiva, implant bar, abutment, adjacent soft tissue, etc.), and lack of semantic labels, making it difficult to accurately segment the target area of ​​the implant bar.

[0003] However, traditional point cloud preprocessing methods often use voxel filtering, random sampling consistency and other methods to simplify point clouds. Although they can reduce the amount of data to a certain extent, these methods are usually based on uniform sampling, which can easily lose key feature information or retain a large number of irrelevant redundant planar point clouds during downsampling, thus affecting the accuracy of subsequent registration.

[0004] Furthermore, some preprocessing methods require manual threshold setting or rely on prior models, which are not adaptable to different sizes or postures of implant poles and are difficult to automate in complex surgical environments. At the same time, when the point cloud data is large, the existing preprocessing workflow has high computational complexity and its processing efficiency is difficult to meet the needs of real-time or near-real-time clinical applications. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a point cloud processing method, apparatus, electronic device and storage medium.

[0006] To achieve the above objectives, the technical solutions adopted in the embodiments of the present invention are as follows: In a first aspect, the present invention provides a point cloud processing method, the method comprising: The original oral cavity point cloud is acquired, and the target point cloud of the implant pole is determined from the original oral cavity point cloud based on the spatial connectivity, color information and normal information of each sampling point in the original oral cavity point cloud. The planting pole target point cloud is segmented by a bounding box normal to obtain the planting pole pattern surface point cloud; Based on the point cloud of the planting pole pattern, a pure pattern feature point cloud of the planting pole is obtained.

[0007] Optionally, the step of determining the target point cloud of the implant pole from the original oral cavity point cloud based on the spatial connectivity, color information, and normal information of each sampling point in the original oral cavity point cloud includes: Based on the color information of each sampling point in the original oral cavity point cloud, candidate point clouds for implant poles are determined from the original oral cavity point cloud; Based on the spatial connectivity, color information, and normal information of each sampling point in the candidate point cloud of the planting pole, the candidate point cloud of the planting pole is subjected to region growing clustering to obtain multiple first point cloud clusters, and the target point cloud of the planting pole is determined from the multiple first point cloud clusters.

[0008] Optionally, the color information includes a red component, a green component, and a blue component, and the step of determining the candidate implant point cloud from the original oral cavity point cloud based on the color information of each sampling point in the original oral cavity point cloud includes: For any target sampling point in the original oral cavity point cloud, determine whether the green component and blue component of the target sampling point are both greater than the red component of the target sampling point; If so, the target sampling point will be retained in the original oral cavity point cloud; If not, the target sampling point is removed from the original oral cavity point cloud; By traversing all sampling points in the original oral cavity point cloud, the candidate point cloud for the implant rod is obtained.

[0009] Optionally, the step of performing directional bounding box normal segmentation on the target point cloud of the planting pole to obtain the surface point cloud of the planting pole pattern includes: Generate the first oriented bounding box of the target point cloud of the planting pole; The face category of each sampling point in the planting pole target point cloud is determined based on the angle between each sampling point in the planting pole target point cloud and the normal vector of each face of the first oriented bounding box; Based on the surface category of each sampling point in the target point cloud of the planting pole, the surface point cloud of the planting pole pattern is determined from the target point cloud of the planting pole.

[0010] Optionally, the step of obtaining the pure pattern feature point cloud of the planting stalk based on the pattern surface point cloud includes: Based on the spatial connectivity, color information, and normal information of each sampling point in the planting pole pattern dot cloud, the planting pole pattern dot cloud is subjected to region growing clustering to obtain multiple second dot cloud clusters, and the pure pattern dot cloud of the planting pole is determined from the multiple second dot cloud clusters. Based on the point cloud of the pure pattern of the planting pole, the feature point cloud of the pure pattern of the planting pole is obtained.

[0011] Optionally, the step of obtaining the feature point cloud of the pure pattern of the planting stalk based on the pure pattern surface point cloud includes: Generate a second oriented bounding box for the pure pattern surface point cloud of the planting pole; The third oriented bounding box is obtained by symmetrically shrinking the face with the largest area in the second oriented bounding box. Using the third directional bounding box and the target point cloud of the planting pole, the pure pattern feature point cloud of the planting pole is obtained.

[0012] Optionally, the step of obtaining the pure pattern feature point cloud of the planting pole using the third directional bounding box and the target point cloud of the planting pole includes: The portion of the planting pole target point cloud located within the third directional bounding box is used as the planting pole pattern surface clipping point cloud; The target plane is obtained by performing plane fitting based on the point cloud of the planting stem pattern surface cutting; Based on the angle between each sampling point in the cutting point cloud of the planting pole pattern surface and the normal vector of the target plane, the pure pattern feature point cloud of the planting pole is determined from the cutting point cloud of the planting pole pattern surface.

[0013] In a second aspect, the present invention provides a point cloud processing apparatus, the apparatus comprising: The acquisition module is used to acquire the raw oral cavity point cloud; The processing module is used to determine the target point cloud of the implant rod from the original oral cavity point cloud based on the spatial connectivity, color information, and normal information of each sampling point in the original oral cavity point cloud; to perform directional bounding box normal segmentation on the target point cloud of the implant rod to obtain the implant rod pattern surface point cloud; and to obtain the pure pattern feature point cloud of the implant rod based on the implant rod pattern surface point cloud.

[0014] Thirdly, the present invention provides an electronic device including a processor and a memory, the memory storing machine-executable instructions executable by the processor, the processor executing the machine-executable instructions to implement the point cloud processing method described in the first aspect above.

[0015] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the point cloud processing method as described in the first aspect above.

[0016] The point cloud processing method, apparatus, electronic device, and storage medium provided in this invention acquire an original oral cavity point cloud, and determine the implant rod target point cloud from the original oral cavity point cloud based on the spatial connectivity, color information, and normal information of each sampling point in the original oral cavity point cloud; perform directional bounding box normal segmentation on the implant rod target point cloud to obtain the implant rod pattern surface point cloud; and obtain the implant rod pure pattern feature point cloud based on the implant rod pattern surface point cloud. Because this invention achieves robust and automated segmentation of the target point cloud of the implant pole by fusing spatial connectivity, color information, and normal information, it overcomes the problems of strong background interference and insignificant features caused by the regularity of the target structure in edentulous scenarios. On this basis, it uses directional bounding box normal segmentation to locate and extract patterned surface point clouds with high recognizability, avoiding the accidental deletion of key geometric semantics by traditional uniform sampling or curvature thresholding methods. This results in a highly pure, compact, and geometrically clear pure patterned feature point cloud, which improves the feature stability, noise resistance, and convergence speed of the subsequent registration algorithm. While ensuring clinical accuracy, it significantly reduces computational overhead and meets the rigid requirements of dental implant navigation systems for real-time performance, robustness, and no human intervention.

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 An example diagram of a planting pole model provided by an embodiment of the present invention is shown; Figure 2 This figure shows a schematic block diagram of an electronic device provided by an embodiment of the present invention; Figure 3 A flowchart illustrating a point cloud processing method provided by an embodiment of the present invention is shown; Figure 4 An example diagram of an original oral cavity point cloud provided by an embodiment of the present invention is shown; Figure 5 An example diagram of a target point cloud of a planting pole provided by an embodiment of the present invention is shown; Figure 6 An example diagram of a planting pole pattern dot cloud provided by an embodiment of the present invention is shown; Figure 7An example diagram of a pure patterned surface dot cloud for a planting pole provided by an embodiment of the present invention is shown; Figure 8 An example diagram of a planting pole pattern surface cut-out point cloud provided by an embodiment of the present invention is shown; Figure 9 An example diagram of a pure pattern feature point cloud of a planting stalk provided by an embodiment of the present invention is shown; Figure 10 A functional block diagram of a point cloud processing device provided in an embodiment of the present invention is shown.

[0020] Icons: 100 - Electronic device; 110 - Memory; 120 - Processor; 130 - Communication module; 200 - Point cloud processing device; 201 - Acquisition module; 202 - Processing module. Detailed Implementation

[0021] 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. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0022] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0023] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0024] In the clinical practice of digital implantology for edentulous jaws, the implant bar is positioned preoperatively / intraoperatively (see...). Figure 1High-precision real-time positioning of the implant's spatial pose is a crucial prerequisite for ensuring the accuracy of the implant's angle, depth, and mechanical distribution. Current mainstream technologies rely on structured light or Time-of-Flight (ToF) depth sensors to perform 3D scanning of the implant pole, acquiring its surface point cloud data (denoted as the original point cloud org_cloud). This point cloud is then aligned with a pre-defined CAD model using point cloud registration algorithms (such as ICP, Go-ICP, FPFH+RANSAC, etc.) to achieve pose calculation.

[0025] However, existing technologies face a triple challenge when processing point clouds of implant rods in edentulous jaws, making it difficult to balance registration robustness, accuracy, and efficiency: (1) The planting rod is a machined part with a high degree of regularity. The main body is composed of a cylindrical surface, an end plane, and engraved patterns (such as cross grooves, digital codes, etc.). The overall curvature changes gently and there are very few local geometric abrupt changes. Traditional feature extraction methods based on curvature, normal change rate, or local descriptors such as FPFH / SHOT are difficult to generate stable, repeatable, and discriminative key points on such low-texture, weak-edge, and highly symmetric structures.

[0026] (2) The existing general point cloud preprocessing process usually performs the following steps in sequence: statistical filtering for noise reduction, voxel mesh downsampling, and planar / cylindrical RANSAC coarse segmentation. Although this paradigm can reduce the amount of data, the voxel filtering uses uniform spatial partitioning and ignores the semantics of the functional area of ​​the planting pole (e.g., only the pattern surface carries positioning information, while the screw hole surface and the side cylindrical surface are both sources of interference); RANSAC planar fitting is easily affected by screw hole depressions, edge burrs, and scanning shadows, resulting in a large deviation of the normal of the fitted plane, which in turn causes the subsequent normal constraint filtering to fail; global downsampling inevitably loses the sharpness of the pattern edge, causing the response intensity of the subsequent feature descriptor based on the edge gradient to be too high.

[0027] (3) In clinical practice, implant rods have multiple models, postures, and working conditions (with bloodstains / saliva reflection, partial occlusion, etc.). Existing methods rely on manually setting hyperparameters such as curvature threshold, sampling voxel size, and RANSAC iteration number, or require pre-registration of template models for each model, which cannot achieve "one-time deployment, full model adaptation", violating the core requirements of oral surgery scenarios for rapid start-up and zero manual intervention.

[0028] To overcome the shortcomings of the prior art, embodiments of the present invention provide a point cloud processing method, apparatus, electronic device, and storage medium, which will be described in detail below.

[0029] Please refer to Figure 2This is a block diagram of an electronic device 100. The electronic device 100 includes a memory 110, a processor 120, and a communication module 130. The memory 110, processor 120, and communication module 130 are electrically connected to each other directly or indirectly to realize data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.

[0030] The memory 110 is used to store programs or data. The memory 110 may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0031] The processor 120 is used to read / write data or programs stored in the memory 110 and to perform corresponding functions.

[0032] The communication module 130 is used to establish a communication connection between the electronic device 100 and other communication terminals through the network, and to send and receive data through the network.

[0033] It should be understood that, Figure 2 The structure shown is only a schematic diagram of the electronic device 100. The electronic device 100 may also include components that are larger than... Figure 2 The more or fewer components shown, or having the same Figure 1 The different configurations shown. Figure 2 The components shown can be implemented using hardware, software, or a combination thereof.

[0034] Please refer to Figure 3 The point cloud processing method provided in this embodiment of the invention includes steps S101 to S103.

[0035] S101, acquire the original oral cavity point cloud, and determine the target point cloud of the implant pole from the original oral cavity point cloud based on the spatial connectivity, color information and normal information of each sampling point in the original oral cavity point cloud.

[0036] In a possible implementation, the process of "determining the target point cloud of the implant pole from the original oral point cloud based on the spatial connectivity, color information, and normal information of each sampling point in the original oral point cloud" can be as follows: determine the candidate point cloud of the implant pole from the original oral point cloud based on the color information of each sampling point in the original oral point cloud; perform region growth clustering on the candidate point cloud of the implant pole based on the spatial connectivity, color information, and normal information of each sampling point in the candidate point cloud of the implant pole to obtain multiple first point cloud clusters, and determine the target point cloud of the implant pole from the multiple first point cloud clusters.

[0037] Furthermore, the color information includes red, green, and blue components. The process of "determining the candidate point cloud for implant rods from the original oral point cloud based on the color information of each sampling point in the original oral point cloud" can be as follows: for any target sampling point in the original oral point cloud, determine whether the green and blue components of the target sampling point are both greater than the red component of the target sampling point; if so, retain the target sampling point in the original oral point cloud; if not, remove the target sampling point from the original oral point cloud; traverse all sampling points in the original oral point cloud to obtain the candidate point cloud for implant rods.

[0038] In other words, to obtain such Figure 4 The original oral cavity point cloud org_cloud shown has each sampling point p_i=(x_i, y_i, z_i, r_i, g_i, b_i) containing three-dimensional spatial coordinates and normalized RGB color components (value range [0, 1]).

[0039] For any target sampling point p_i in org_cloud, the following judgment is performed: if g_i>r_i and b_i>r_i, then the point is determined to belong to the high reflectivity non-red area of ​​the implant rod (the surface of the titanium alloy implant rod is cold gray after clinical disinfection, and the R component in its RGB channel is significantly lower than the G / B component; while oral soft tissue, gingiva, and salivary membrane generally show a red-yellow main color under white light illumination, and the R component accounts for a larger proportion), and the point is retained; otherwise, it is discarded. After traversing all points, the candidate point cloud of the implant rod, candidate_cloud, is obtained.

[0040] Using candidate_cloud as input, perform region growing clustering: randomly select a point from candidate_cloud as a seed; a seed is formed if and only if the Euclidean distance between p_i and p_j is less than a pre-defined distance threshold (this threshold is set based on the minimum structural feature size of the planting stalk and the average density of the point cloud), the included normal angle is less than a pre-defined angle, and the color difference |r_i| r_j|+|g_i g_j|+|b_i When b_j is less than the preset color difference, p_i and p_j are considered as adjacent points and assigned to the same point cloud cluster; repeat the above process until all points in candidate_cloud are assigned to a cluster, and a total of K first point cloud clusters {C_1,C_2, ..., C_K} are obtained.

[0041] Count the number of points in each cluster, select the cluster C_max with the most points, and obtain the following result: Figure 5 The target point cloud of the planting pole is shown as rod_cloud.

[0042] Understandably, embodiments of the present invention adapt color shifts in intraoral scan images under different lighting conditions by using RGB color component ratio criteria; and improve robustness to local occlusion of implant rods and slight motion artifacts by superimposing triple constraints of spatial connectivity (distance), geometric consistency (normal), and color stability (RGB difference) on region growth, thereby avoiding the failure problem of traditional single normal clustering on regular cylindrical surfaces and effectively eliminating interference such as saliva reflection points and gingival margin noise.

[0043] S102, perform directional bounding box normal segmentation on the target point cloud of the planting pole to obtain the surface point cloud of the planting pole pattern.

[0044] In a possible implementation, the process of implementing step S102 may include sub-steps S102-1 to S102-3.

[0045] S102-1, Generate the first oriented bounding box of the planting pole target point cloud.

[0046] Perform principal component analysis (PCA) on rod_cloud to obtain its three eigenvectors v1, v2, and v3 (arranged in descending order of eigenvalues) of the covariance matrix, forming an orthogonal rotation matrix R = [v1v2v3] ∈ ³ x ³; Calculate the centroid c of the point cloud; Translate the coordinates of all points in rod_cloud to a point with c as the origin, and then... Rotate to the PCA coordinate system to obtain the extreme points P_min and P_max of the axis-aligned bounding box.

[0047] Understandably, the parameters of the first oriented bounding box obb1 include: center c, rotation matrix R, extreme points P_min and P_max.

[0048] Define the PCA axis alignment normal vector, representing the normal vectors of the six faces of the first oriented bounding box obb1, with the corresponding formula as follows:

[0049] For each initial PCA normal vector via obb1 rotation matrix RPerform a linear transformation to obtain the normal vector of the i-th face of the obb1 bounding box. The formula is as follows:

[0050]

[0051]

[0052] S102-2, determine the face category of each sampling point in the planting pole target point cloud based on the angle between each sampling point in the planting pole target point cloud and the normal vector of each face of the first oriented bounding box.

[0053] For any point p_i in rod_cloud, its normal vector Normal vectors of each face of the obb1 bounding box Calculate the included angle and take the angle corresponding to the smallest included angle. The formula for classifying this point is as follows:

[0054]

[0055] S102-3, Based on the surface category of each sampling point in the target point cloud of the planting pole, determine the surface point cloud of the planting pole pattern from the target point cloud of the planting pole.

[0056] In this embodiment of the invention, the number of points corresponding to each face category is counted, the category with the most points k_max is selected, and all points with label(p_i)=k_max are extracted to form a structure as follows. Figure 6 The planting pole pattern shown is a cloud of dots.

[0057] Understandably, the embodiments of the present invention use OBB normal segmentation to avoid the feature disappearance problem caused by the curvature of the smooth cylindrical surface of the planting pole approaching zero, without relying on local curvature calculation; and use global bounding box attitude-guided normal clustering to group the point cloud according to macroscopic orientation, so that the subsequent registration algorithm focuses on the most discriminative pattern plane, rather than the redundant side cylindrical surface.

[0058] S103, based on the planting pole pattern point cloud, obtain the pure pattern feature point cloud of the planting pole.

[0059] In a possible implementation, the process of implementing step S103 may include sub-steps S103-1 to S103-3.

[0060] S103-1, Based on the spatial connectivity, color information and normal information of each sampling point in the planting pole pattern point cloud, perform region growth clustering on the planting pole pattern point cloud to obtain multiple second point cloud clusters, and determine the pure pattern point cloud of the planting pole from the multiple second point cloud clusters.

[0061] Using face_cloud as input, perform region growing clustering: randomly select a point from face_cloud as a seed; the seed is selected if and only if the Euclidean distance between p_i and p_j is less than a pre-defined distance threshold (this threshold is set based on the minimum structural feature size of the planting stalk and the average density of the point cloud), the angle between their normals is less than a pre-defined angle, and the color difference |r_i| r_j|+|g_i g_j|+|b_i When b_j is less than the preset color difference, p_i and p_j are considered as adjacent points and assigned to the same point cloud cluster; repeat the above process until all points in face_cloud are assigned to a cluster, and a total of N first point cloud clusters {D_1, D_2, ..., D_N} are obtained.

[0062] The second point cloud cluster D_MAX with the most points is selected as the small_face_cloud of the pure pattern face of the planting stem (see...). Figure 7 ).

[0063] Understandably, because the point cloud in the spiral hole region is sparse and the normal dispersion is high, it is impossible to form a large-scale connected cluster. By performing regional growth clustering on the face_cloud, the spiral hole plane can be automatically removed.

[0064] S103-2, obtain the feature point cloud of the pure pattern of the planting stalk based on the point cloud of the pure pattern surface.

[0065] Furthermore, the implementation process of step S103-2 can be as follows: S103-2-1, generates the second oriented bounding box for the pure pattern surface point cloud of the planting pole.

[0066] Perform PCA on small_face_cloud to generate a second oriented bounding box plane_obb, whose parameters include the center c_p, rotation matrix R_p, and the coordinates of the top-left point of the bounding box in space. P min and the coordinates of the bottom right point P max ,in, P min and P max Satisfy the following formula:

[0067] S103-2-2, symmetrically shrink the face with the largest area in the second oriented bounding box to obtain the third oriented bounding box.

[0068] Determine the face with the largest area among the six faces of plane_obb, and denot it as A. max It satisfies the following formula:

[0069]

[0070]

[0071] Let u1 and u2 be Two orthogonal unit vectors in the plane, u3 is The normal vector; shrink along the u1 direction by N=1.0 mm, and along the u2 direction by M=0.5 mm, to obtain the third oriented bounding box shrink_obb, whose center is still c_p.

[0072] S103-2-3, using the third directional bounding box and the target point cloud of the planting pole, a pure pattern feature point cloud of the planting pole is obtained.

[0073] In this embodiment of the invention, the implementation process of step S103-2-3 may be as follows: the part of the planting pole target point cloud located within the third directional bounding box is taken as the planting pole pattern surface clipping point cloud; a plane fitting is performed based on the planting pole pattern surface clipping point cloud to obtain the target plane; and the planting pole pure pattern feature point cloud is determined from the planting pole pattern surface clipping point cloud according to the angle between each sampling point in the planting pole pattern surface clipping point cloud and the normal vector of the target plane.

[0074] In other words, by using shrink_obb to spatially select rod_cloud and retaining the points within the box, we obtain the cropped point cloud of the planting pole pattern, denoted as plane_cloud (see...). Figure 8 ).

[0075] Specifically, when the spatial coordinates of any sampling point in the point cloud rod_cloud satisfy the following formula, it can be determined that it falls within shrink_obb:

[0076]

[0077]

[0078] In the formula, P This represents the spatial coordinates of any sampling point in the point cloud rod_cloud. C This represents the center point of shrink_obb. uIndicates the directions of the three orthogonal unit axes.

[0079] The plane function is obtained by fitting the plane_cloud to the RANSAC algorithm:

[0080] Accordingly, the plane normal is:

[0081] Iterate through all sampling points in plane_cloud and calculate the normal vector of each sampling point according to the following formula. and The included angle Filter out Sampling points smaller than D=0.9 (flat background without texture) constitute the feature cloud of the pure pattern of the planting stem, which is composed of the final retained sampling points (located in pattern grooves, etched edges, or tiny protrusions). Figure 9 ).

[0082]

[0083] Understandably, the embodiments of the present invention precisely extract the core area of ​​the pattern through shrinkage operations, avoiding the chamfered edges of the planting rod and the transition area of ​​the screw holes; ensure the noise resistance of the planar model through RANSAC fitting; and retain the sampling points of high-discrimination geometric features such as lines, text, and grids of the corresponding pattern through normal threshold filtering, thereby improving the discriminative power of subsequent FPFH / SHOT feature descriptors.

[0084] In summary, this invention addresses the technical challenges of weak features, excessive redundancy, and strong interference in point clouds for edentulous implant poles by performing a five-level progressive processing method: color screening, multidimensional clustering, OBB normal segmentation, shrinkage trimming, and normal sharpening. The resulting feature_cloud maintains clinical reproducibility while possessing three major advantages: high feature density, low data volume, and strong robustness, providing a reliable data foundation for real-time, high-precision surgical navigation.

[0085] To perform the corresponding steps in the above embodiments and various possible methods, an implementation of the point cloud processing device 200 is given below. Further, please refer to... Figure 10 , Figure 10 This is a functional block diagram of a point cloud processing device 200 provided in an embodiment of the present invention. It should be noted that the point cloud processing device 200 provided in this embodiment has the same basic principle and technical effects as those in the above embodiments. For the sake of brevity, any parts not mentioned in this embodiment can be referred to the corresponding content in the above embodiments. The point cloud processing device 200 includes: The acquisition module 201 is used to acquire the original oral cavity point cloud.

[0086] Processing module 202 is used to determine the implant rod target point cloud from the original oral cavity point cloud based on the spatial connectivity, color information and normal information of each sampling point in the original oral cavity point cloud; to perform directional bounding box normal segmentation on the implant rod target point cloud to obtain the implant rod pattern surface point cloud; and to obtain the implant rod pure pattern feature point cloud based on the implant rod pattern surface point cloud.

[0087] Optionally, the above modules can be stored in the form of software or firmware. Figure 2 The memory shown is either stored in or embedded in the operating system (OS) of the electronic device 100, and can be used by... Figure 2 The processor 120 executes the program. Meanwhile, the data and program code required to execute the above modules can be stored in the memory 110.

[0088] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0089] In addition, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0090] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0091] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A point cloud processing method, characterized in that, The method includes: The original oral cavity point cloud is acquired, and the target point cloud of the implant pole is determined from the original oral cavity point cloud based on the spatial connectivity, color information and normal information of each sampling point in the original oral cavity point cloud. The planting pole target point cloud is segmented by a bounding box normal to obtain the planting pole pattern surface point cloud; Based on the point cloud of the planting pole pattern, a pure pattern feature point cloud of the planting pole is obtained.

2. The point cloud processing method as described in claim 1, characterized in that, The step of determining the target point cloud of the implant pole from the original oral cavity point cloud based on the spatial connectivity, color information, and normal information of each sampling point in the original oral cavity point cloud includes: Based on the color information of each sampling point in the original oral cavity point cloud, candidate point clouds for implant poles are determined from the original oral cavity point cloud; Based on the spatial connectivity, color information, and normal information of each sampling point in the candidate point cloud of the planting pole, the candidate point cloud of the planting pole is subjected to region growing clustering to obtain multiple first point cloud clusters, and the target point cloud of the planting pole is determined from the multiple first point cloud clusters.

3. The point cloud processing method as described in claim 2, characterized in that, The color information includes red, green, and blue components. The step of determining the candidate implant point cloud from the original oral cavity point cloud based on the color information of each sampling point in the original oral cavity point cloud includes: For any target sampling point in the original oral cavity point cloud, determine whether the green component and blue component of the target sampling point are both greater than the red component of the target sampling point; If so, the target sampling point will be retained in the original oral cavity point cloud; If not, the target sampling point is removed from the original oral cavity point cloud; By traversing all sampling points in the original oral cavity point cloud, the candidate point cloud for the implant rod is obtained.

4. The point cloud processing method as described in claim 1, characterized in that, The step of performing directional bounding box normal segmentation on the target point cloud of the planting pole to obtain the patterned surface point cloud of the planting pole includes: Generate the first oriented bounding box of the target point cloud of the planting pole; The face category of each sampling point in the planting pole target point cloud is determined based on the angle between each sampling point in the planting pole target point cloud and the normal vector of each face of the first oriented bounding box; Based on the surface category of each sampling point in the target point cloud of the planting pole, the surface point cloud of the planting pole pattern is determined from the target point cloud of the planting pole.

5. The point cloud processing method as described in claim 1, characterized in that, The step of obtaining the pure pattern feature point cloud of the planting pole based on the pattern surface point cloud includes: Based on the spatial connectivity, color information, and normal information of each sampling point in the planting pole pattern dot cloud, the planting pole pattern dot cloud is subjected to region growing clustering to obtain multiple second dot cloud clusters, and the pure pattern dot cloud of the planting pole is determined from the multiple second dot cloud clusters. Based on the point cloud of the pure pattern of the planting pole, the feature point cloud of the pure pattern of the planting pole is obtained.

6. The point cloud processing method as described in claim 5, characterized in that, The step of obtaining the feature point cloud of the pure pattern of the planting stalk based on the pure pattern surface point cloud includes: Generate a second oriented bounding box for the pure pattern surface point cloud of the planting pole; The third oriented bounding box is obtained by symmetrically shrinking the face with the largest area in the second oriented bounding box. Using the third directional bounding box and the target point cloud of the planting pole, the pure pattern feature point cloud of the planting pole is obtained.

7. The point cloud processing method as described in claim 6, characterized in that, The step of obtaining the pure pattern feature point cloud of the planting pole using the third directional bounding box and the target point cloud of the planting pole includes: The portion of the planting pole target point cloud located within the third directional bounding box is used as the planting pole pattern surface clipping point cloud; The target plane is obtained by performing plane fitting based on the point cloud of the planting stem pattern surface cutting; Based on the angle between each sampling point in the cutting point cloud of the planting pole pattern surface and the normal vector of the target plane, the pure pattern feature point cloud of the planting pole is determined from the cutting point cloud of the planting pole pattern surface.

8. A point cloud processing device, characterized in that, The device includes: The acquisition module is used to acquire the raw oral cavity point cloud; The processing module is used to determine the target point cloud of the implant rod from the original oral cavity point cloud based on the spatial connectivity, color information, and normal information of each sampling point in the original oral cavity point cloud; to perform directional bounding box normal segmentation on the target point cloud of the implant rod to obtain the implant rod pattern surface point cloud; and to obtain the pure pattern feature point cloud of the implant rod based on the implant rod pattern surface point cloud.

9. An electronic device, characterized in that, It includes a processor and a memory, the memory storing machine-executable instructions that can be executed by the processor, the processor executing the machine-executable instructions to implement the point cloud processing method according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the point cloud processing method as described in any one of claims 1-7.

Images