A data processing method and system for dental three-dimensional models

Abstract

The present application discloses a data processing method and system for a dental three-dimensional model, relating to the technical field of dental three-dimensional modeling, and includes the following steps: obtaining initial data from an oral scan, identifying the tooth type in the initial data, and determining the corresponding tooth surface reflection parameters based on the tooth type; generating scanning parameter combinations for different tooth positions based on the tooth surface reflection parameters and the tooth anatomical structure; collecting tooth point cloud data and identifying enamel reflective areas and shadow-occluded areas; performing de-reflection processing on the enamel reflective areas, completing data on the shadow-occluded areas, and generating corrected tooth point cloud data; and performing registration and fusion on the corrected tooth point cloud data to generate a dental three-dimensional model containing the complete dentition. This application solves the problem of data loss caused by enamel reflection and adjacent tooth occlusion by using a differentiated scanning strategy and region-based data processing tailored to the specific optical properties of teeth, thereby achieving the automated construction of high-quality dental three-dimensional models.

Description

Technical Field

[0001] The present application relates to the technical field of dental three-dimensional modeling, and in particular to a data processing method and system for dental three-dimensional models. Background Art

[0002] Intraoral 3D scanning technology has become a crucial tool in modern digital dental diagnosis and treatment. Current mainstream intraoral scanners, which mostly utilize structured light or confocal laser technology, continuously capture sequences of oral images and perform 3D reconstructions, acquiring precise 3D morphological data of oral tissues, including teeth and gums. This digital data can be directly used in computer-aided design and manufacturing (CAD / CAM) systems to achieve precise fabrication of restorations, significantly improving the accuracy and efficiency of oral restorations.

[0003] However, existing dental 3D scanning technology still faces many technical challenges in practical application. First, the highly reflective nature of tooth enamel leads to local overexposure during scanning, resulting in loss of detailed information on the tooth surface. Second, the mutual occlusion of adjacent teeth in the confined space of the oral cavity creates scanning blind spots, making data collection in the adjacent and cervical areas particularly difficult. Third, the morphology of different teeth varies significantly, making it difficult to use uniform scanning parameters to account for both the thin and transparent nature of anterior teeth and the complex pit and fissure structure of the molar occlusal surface. Finally, existing methods lack targeted compensation strategies for data post-processing. Simple filtering or interpolation cannot effectively restore key anatomical features, affecting the clinical usability of the final 3D model. Summary of the Invention

[0004] In view of the above-mentioned problems, this application is proposed.

[0005] Therefore, the present application provides a data processing method and system for a dental three-dimensional model, which can solve the problems mentioned in the background technology.

[0006] To solve the above technical problems, this application provides the following technical solutions:

[0007] In a first aspect, the present application provides a data processing method for a dental three-dimensional model, comprising: obtaining initial data of an oral scan, identifying the tooth type in the initial data, and determining corresponding tooth surface reflection parameters according to the tooth type; generating scanning parameter combinations for different tooth positions according to the tooth surface reflection parameters and the tooth anatomical structure; collecting tooth point cloud data according to the scanning parameter combination, and identifying enamel reflective areas and shadow-blocked areas in the tooth point cloud data; performing de-reflection processing on the enamel reflective areas, performing data completion on the shadow-blocked areas, and generating corrected tooth point cloud data; performing registration and fusion on the corrected tooth point cloud data based on tooth anatomical landmark points, and generating a dental three-dimensional model containing a complete dentition.

[0008] Preferably, the identifying of the tooth type in the initial data and determining the corresponding tooth surface reflection parameters according to the tooth type includes: identifying the tooth position distribution of anterior teeth, premolars and molars through morphological analysis; determining the corresponding curvature distribution values ​​according to the crown morphology of different tooth positions; and determining the surface reflection parameters of each tooth position in combination with the curvature distribution values ​​and the optical parameters of the enamel.

[0009] Preferably, the method of generating a scanning parameter combination for different tooth positions based on the tooth surface reflection parameters and the tooth anatomical structure includes: calculating the ratio of the occlusal pit and fissure depth to the crown height as a morphological complexity index; determining the scanning angle range and scanning density based on the morphological complexity index; wherein the scanning density is a function of the morphological complexity index and the basic density value;

[0010] The laser power is determined according to the scanning density so that the product of the laser power and the scanning density remains constant, and the corresponding exposure time is determined according to the laser power.

[0011] Preferably, the identification of enamel reflection areas and shadow occlusion areas in the tooth point cloud data includes: analyzing the point cloud density distribution, and identifying areas where the point cloud density is lower than a first density value as potential problem areas; judging whether the potential problem area is data missing caused by enamel reflection or a shadow area caused by occlusion of adjacent teeth based on the relationship between the rate of change of the tooth surface normal vector and the distance between adjacent points; extracting the gum line contour, and identifying the data collection blind area of ​​the tooth neck.

[0012] Preferably, the de-reflection processing is performed on the enamel reflective area, and the data completion is performed on the shadow-blocked area to generate corrected tooth point cloud data, including: for the enamel reflective area, calculating the curvature change rate of the boundary point; if the curvature change rate is less than a predetermined ratio of the curvature change rate of the adjacent normal area, reconstructing the missing surface data using an interpolation algorithm; for the shadow-blocked area, determining the number of angles and angle intervals of the supplementary scans based on the ratio of the area of ​​the shadow-blocked area to the amount of data reconstructed by the interpolation algorithm, performing supplementary scans to obtain supplemented tooth point cloud data; fusing the surface data reconstructed by the interpolation algorithm, the supplemented tooth point cloud data, and the valid data in the tooth point cloud data to generate corrected tooth point cloud data.

[0013] Preferably, the correction tooth point cloud data is registered and fused based on tooth anatomical landmarks, including: identifying the cusp point, ridge line, and central pit point as tooth anatomical landmarks; establishing a tooth coordinate system based on the tooth anatomical landmarks to determine the spatial position of each tooth; performing overall registration of the dentition according to the adjacency and occlusal relationship of adjacent teeth; and retaining data of the pit and fissure area and tooth edge area when fusing multiple scan data.

[0014] Preferably, it also includes: measuring the distance between adjacent tooth anatomical landmark points, if the distance between adjacent tooth anatomical landmark points exceeds the width of a single normal tooth, it is determined to be a tooth missing area, and a three-dimensional boundary of the missing space is generated according to the adjacent tooth morphology and the position of the opposing teeth; detecting the area where the tooth surface normal vector is locally reversed, calculating the depth value of the locally reversed area, if the ratio of the depth value to the average curvature radius of the tooth surface represents a concave morphology, marking the locally reversed area as a lesion area and calculating the three-dimensional volume of the lesion area; generating an evaluation index of the tooth health status according to the ratio of the three-dimensional volume of the lesion area to the total volume of the tooth.

[0015] In a second aspect, the present application also provides a data processing system for a dental three-dimensional model, comprising: a tooth recognition module, a scanning strategy module, a data acquisition module, a data correction module and a model construction module, wherein: the tooth recognition module is used to obtain initial data of an oral scan, identify the tooth type in the initial data, and determine the corresponding tooth surface reflection parameters according to the tooth type; the scanning strategy module is used to generate scanning parameter combinations for different tooth positions according to the tooth surface reflection parameters and the tooth anatomical structure; the data acquisition module is used to collect tooth point cloud data according to the scanning parameter combination, and identify the enamel reflective area and the shadow occlusion area in the tooth point cloud data; the data correction module is used to perform de-reflection processing on the enamel reflective area, perform data completion on the shadow occlusion area, and generate corrected tooth point cloud data; the model construction module is used to perform registration and fusion of the corrected tooth point cloud data based on tooth anatomical landmark points to generate a dental three-dimensional model containing a complete tooth arch.

[0016] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0017] Acquire initial data from an oral scan, identify the tooth type in the initial data, and determine corresponding tooth surface reflection parameters based on the tooth type; generate scanning parameter combinations for different tooth positions based on the tooth surface reflection parameters and tooth anatomical structure; collect tooth point cloud data according to the scanning parameter combination, and identify enamel reflective areas and shadow-blocked areas in the tooth point cloud data; perform de-reflection processing on the enamel reflective areas, perform data completion on the shadow-blocked areas, and generate corrected tooth point cloud data; perform registration and fusion on the corrected tooth point cloud data based on tooth anatomical landmarks to generate a dental three-dimensional model containing a complete dentition.

[0018] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0019] Acquire initial data from an oral scan, identify the tooth type in the initial data, and determine corresponding tooth surface reflection parameters based on the tooth type; generate scanning parameter combinations for different tooth positions based on the tooth surface reflection parameters and tooth anatomical structure; collect tooth point cloud data according to the scanning parameter combination, and identify enamel reflective areas and shadow-blocked areas in the tooth point cloud data; perform de-reflection processing on the enamel reflective areas, perform data completion on the shadow-blocked areas, and generate corrected tooth point cloud data; perform registration and fusion on the corrected tooth point cloud data based on tooth anatomical landmarks to generate a dental three-dimensional model containing a complete dentition.

[0020] The implementation of this application has the following beneficial effects: This application provides a data processing method and system for dental three-dimensional models. By constructing a mapping relationship between tooth types and surface reflection parameters, when different tooth positions such as anterior teeth, premolars or molars are identified, the scanning parameter combination corresponding to the morphological characteristics of each tooth position can be retrieved. Even when facing the translucent incisal edge of the anterior teeth or the deep pit and groove structure of the molars, the parameters can be configured in a targeted manner. At the same time, the classification and identification mechanism of the problem area can distinguish between data missing caused by enamel reflection and adjacent tooth occlusion. When a reflective area is detected, the surface morphology is reconstructed through an interpolation algorithm constrained by curvature continuity, and the integrity of the tooth contour can be maintained even under high reflection conditions. When an occluded shadow is identified, the supplementary scanning angle is calculated based on the geometric relationship between the shadow area and the light source, and the missing data is obtained from the physical level, avoiding the morphological deviation caused by relying solely on algorithmic speculation.

[0021] Therefore, during the data fusion phase, a registration strategy based on anatomical landmarks fully exploits the stability of biological features such as cusps and ridges. Once these feature points are extracted, a spatial coordinate system based on the anatomical landmarks can be established. This allows the splicing of multiple scan segments to no longer rely on iterative surface matching calculations, but rather to achieve rapid and accurate positioning through the inherent constraints of anatomical structure. Furthermore, the simultaneous detection of missing data and identification of lesions during the registration process ensures that the resulting 3D model not only contains complete morphological information but also carries clinically valuable diagnostic markers, effectively integrating data acquisition with clinical analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0023] Figure 1 This is an overall flow chart of a data processing method for a dental three-dimensional model involved in this application;

[0024] Figure 2 This is a schematic diagram of the overall structure of a dental three-dimensional model data processing system involved in this application;

[0025] Figure 3 This is a computer device diagram of a data processing method for a dental three-dimensional model involved in this application. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0027] In an exemplary embodiment, a method for processing dental three-dimensional model data is provided, comprising the following steps 100 to 500. Figure 1 As shown:

[0028] Step 100: Acquire initial data of an oral scan, identify the tooth type in the initial data, and determine corresponding tooth surface reflection parameters according to the tooth type.

[0029] During oral 3D scanning, different tooth types exhibit distinct reflection characteristics to the scanning light due to differences in their anatomical morphology and surface properties. The incisal edges of anterior teeth are thin and translucent, while the occlusal surfaces of molars have complex pit and fissure structures. These differences make it difficult to obtain high-quality 3D data using uniform scanning parameters. Therefore, it is necessary to first obtain initial scan data, identify the tooth types at different positions, and then determine the surface reflection parameters that match each tooth position.

[0030] In the embodiment of the present application, step 100 includes the following steps 110 and 120:

[0031] Step 110: Acquire initial data of oral scan.

[0032] Initial oral scan data refers to low-resolution 3D data obtained through a pre-scan using an intraoral scanner. This pre-scan uses standard scanning parameters to quickly capture basic morphological information for the entire dentition. This initial data includes rough tooth outlines, their approximate spatial positions, and preliminary depth information.

[0033] For example, when using a structured light scanner for a pre-scan, a fringe pattern is projected into the oral cavity. The camera captures the deformed fringe image, and triangulation is used to calculate the initial point cloud data. The pre-scan speed can be set to 20-30 frames per second and the resolution to 0.5-1.0mm to quickly capture the overall morphology.

[0034] In an alternative embodiment, initial data can be obtained through confocal laser scanning. As the scanning head moves along the tooth, the laser beam is focused at different depths on the tooth surface, and depth information is constructed by detecting changes in the intensity of the reflected light. Initial data obtained through confocal laser scanning is particularly suitable for scanning confined spaces such as adjacent areas.

[0035] In another optional embodiment, the initial data can be obtained through active wavefront sampling technology. This technology reconstructs the three-dimensional shape by projecting structured light of a known pattern and analyzing the wavefront distortion of the reflected light. It has a good scanning effect on the smooth surface of the anterior teeth.

[0036] Step 120: Identify the tooth type in the initial data and determine the corresponding tooth surface reflection parameters according to the tooth type.

[0037] Based on the initial data acquired, different tooth types are identified by analyzing their morphological characteristics. Identifying tooth type is the basis for subsequently determining surface reflectance parameters. Differences in optical properties between different tooth types dictate the need for different scanning strategies.

[0038] Step 121: Identify the tooth position distribution of the anterior teeth, premolars, and molars through morphological analysis.

[0039] Morphological analysis is the process of classifying and identifying teeth based on their geometric features. Specifically, tooth type is identified by analyzing features such as the width and height ratio of each tooth in the initial point cloud data, as well as the crown contour shape.

[0040] Identifying features of anterior teeth (including central incisors, lateral incisors, and canines) include a flat crown, a labiolingual diameter significantly larger than the mesiodistal diameter, and a straight or slightly curved incisal edge. By calculating the length-to-width ratio of the crown, a tooth with a ratio greater than 1.2 can be preliminarily identified as an anterior tooth.

[0041] Identifying features of premolars include a nearly cubic crown with distinct buccal and lingual cusps, and a relatively simple developmental groove on the occlusal surface. Premolars can be identified by examining the occlusal surface for the presence of two primary cusps.

[0042] Identifying features of molars include a large crown, four to five cusps on the occlusal surface, and a complex pit and fissure system. By analyzing the complexity of the occlusal surface, a molar can be identified when three or more recessed areas are detected.

[0043] In an optional embodiment, machine learning methods can also be used to identify tooth types. A classification model is trained using pre-labeled 3D tooth data to extract feature vectors, including crown volume, surface area, and principal curvature distribution. Automatic classification is then performed using a support vector machine or random forest algorithm.

[0044] In another optional embodiment, tooth position identification can be performed based on the dental arch curve. First, the dental arch curve is fitted, and the tooth position is determined based on the position of the teeth on the dental arch. For example, the teeth closest to the dental arch midline are the central incisors, followed by the lateral incisors and canines as they move further away.

[0045] Step 122: Determine corresponding curvature distribution values ​​according to the crown shapes of different tooth positions.

[0046] The curvature distribution value reflects the degree of curvature of the tooth surface and is a key parameter affecting light reflection. For each point in the initial point cloud data, its principal curvatures k1 and k2 are calculated, and then the mean curvature H = (k1 + k2) / 2 and Gaussian curvature K = k1 × k2 are obtained.

[0047] The curvature distribution characteristics of the front teeth are: the labial surface is convex, the curvature is relatively uniform, and the average curvature range is 0.2-0.5 mm -1 The central part of the tongue is concave, with a negative curvature area; the curvature at the incisal edge changes dramatically, up to 2-3 mm -1 .

[0048] Curvature distribution characteristics of premolars: buccal and lingual curvature is moderate, about 0.3-0.6 mm -1The occlusal tooth tip is a high curvature area, which can reach 1-2 mm. -1 ; The developmental groove is a negative curvature area.

[0049] The curvature distribution characteristics of molars: the curvature of the occlusal surface changes complexly, and the negative curvature of the bottom of the socket can reach -2 mm -1 , the positive curvature of the cusp can reach 3 mm -1 ; This drastic change in curvature makes it easy to produce shadows and reflections when scanning.

[0050] For example, for a maxillary first molar, the curvature of the mesiobuccal cusp is calculated to be 2.8 mm by local quadratic surface fitting. -1 , the foveal curvature is -1.5 mm -1 ,This difference in curvature indicates the need to adjust the scanning angle to avoid occlusion.

[0051] Step 123: Determine the surface reflection parameters of each tooth position by combining the curvature distribution value and the optical parameters of the tooth enamel.

[0052] The optical parameters of tooth enamel include refractive index, absorption coefficient, and scattering coefficient. The refractive index of natural tooth enamel is approximately 1.62-1.65, and it exhibits anisotropy. Combined with the curvature distribution, the reflectivity at different angles of incidence can be calculated.

[0053] According to Fresnel's law of reflection, when light is incident on a surface with a curvature of κ at an angle θ, the reflectivity R can be expressed as:

[0054] ;

[0055] in, is the refractive index of air (about 1.0), is the refractive index of tooth enamel, is the angle of incidence, is the refraction angle, is the reflectivity.

[0056] Therefore, when the scanning light is irradiated onto the tooth surface, the accurate reflectivity can be calculated according to the Fresnel formula by measuring the curvature value and the incident angle; when the incident angle at the incisal edge of the anterior tooth exceeds 45°, the calculated reflectivity drops below 0.3, indicating that the expected reflection intensity in this area needs to be reduced; and in the deep molar pits and fissures, the incident angle varies from 0° to 60° due to the change in curvature, and the corresponding reflectivity fluctuates violently between 0.04 and 0.25, and therefore it is necessary to set varying reflection parameters to adapt to the scanning requirements at different depths, so that complete three-dimensional data can be obtained even in complex oral environments.

[0057] For anterior teeth, due to the translucency of the incisal edge, when the incident angle is greater than 45°, the transmitted light is enhanced and the reflected light is weakened. Therefore, transmission compensation needs to be considered when determining its surface reflection parameters, and the reflection coefficient is set to 0.3-0.4.

[0058] For the molar pit and fissure area, due to the irregularity and high curvature of the surface, specular reflection and diffuse reflection coexist, and a mixed reflection model needs to be set with a specular reflection coefficient of 0.2 and a diffuse reflection coefficient of 0.6.

[0059] In an optional embodiment, the reflectance parameters can be obtained by actual measurement. A spectrophotometer is used to measure the reflectance of the ex vivo tooth at multiple angles, and an angle-reflectance lookup table is established. During scanning, the corresponding reflectance parameters are obtained by looking up the table according to the real-time incident angle.

[0060] For example, in an actual scan of a patient's maxillary dentition, position 16 is identified as the first molar, and the average curvature of its occlusal surface is 1.2 mm. -1 , the standard deviation of curvature is 0.8 mm -1 , indicating significant surface undulations. Combined with the enamel refractive index of 1.63, we calculated that at a standard scanning angle (45° to the tooth's long axis), the average reflectivity was 0.35. However, this value dropped to 0.15 within the deep pits and fissures. Therefore, surface reflectivity parameters were set for this tooth: a base reflectivity of 0.35 and a reflectivity of 0.15 in the pit and fissure areas. Furthermore, a marker was set for additional scanning at multiple angles.

[0061] Through the above steps, this application can accurately identify different types of teeth in the mouth and determine personalized surface reflection parameters for each tooth position, laying the foundation for the subsequent development of targeted scanning strategies. This tooth type-based parameterization method improves the scanning quality of different tooth positions compared to traditional unified parameter scanning, and its advantages are more clearly demonstrated in areas with large differences in optical properties such as anterior teeth and molars.

[0062] Step 200: Generate scanning parameter combinations for different tooth positions based on tooth surface reflection parameters and tooth anatomical structure.

[0063] After obtaining the surface reflectance parameters for each tooth, these optical properties need to be combined with the anatomical features of the tooth to generate parameter combinations that can meet different scanning challenges. The complexity of the tooth's anatomy directly affects the difficulty of scanning. For example, the deep pits and grooves of molars are prone to shadowing, and the thin incisal edges of anterior teeth are prone to excessive transmission. Therefore, it is necessary to establish a mapping relationship between morphological features and scanning parameters.

[0064] In the embodiment of the present application, step 200 includes steps 210 to 230:

[0065] Step 210: Calculate the ratio of the occlusal pit and fissure depth to the crown height as a morphological complexity index.

[0066] Morphological complexity is a key parameter for quantifying the degree of surface roughness in teeth. Pit and fissure depth reflects the degree of depression on the tooth surface, while crown height represents the overall size of the tooth. The ratio of the two can be used to normalize the morphological complexity of teeth.

[0067] Specifically, the highest point of the crown and the lowest point of the occlusal surface are first extracted from the initial scan data, and the vertical distance is calculated to obtain the crown height (unit: mm). Next, all recessed areas on the occlusal surface are identified. For each recessed area, the vertical distance of its deepest point relative to the surrounding cusp plane is calculated, and the maximum value is taken as the pit and fissure depth (unit: mm). The morphological complexity index is the pit and fissure depth divided by the crown height. It is a dimensionless value, typically ranging from 0 to 0.5.

[0068] For the front teeth, due to the relatively flat occlusal surface and small pit and fissure depth, the morphological complexity index is usually less than 0.1; the premolars have developmental grooves of medium depth, and the index value is between 0.1-0.2; the deep pit and fissure system of the molars makes the index value reach 0.2-0.4, and some molars with deep pits and grooves can even reach 0.5.

[0069] In an alternative implementation, a multi-point measurement method can be used to calculate morphological complexity. Nine measurement points (a 3×3 grid) are evenly selected on the occlusal surface. The distance from each point to the reference plane is measured, and the standard deviation of the distance is calculated. The ratio of the standard deviation to the crown height is used as the morphological complexity index. This method can more comprehensively reflect the overall surface undulations.

[0070] In another alternative embodiment, morphological complexity can be calculated based on surface area ratio. The difference between the actual occlusal surface area and its horizontal projection, divided by the projected area, can be used as a proxy for morphological complexity. A larger surface area ratio indicates a more undulating surface.

[0071] Step 220: Determine the scanning angle range and scanning density based on the morphological complexity index.

[0072] The scanning density value is a function of the morphological complexity index and the basic density value.

[0073] The scanning angle range determines the motion trajectory of the scan head relative to the tooth, while the scanning density determines the number of data points collected per unit area. More complex areas require a wider angle range and higher scanning density to avoid occlusion and data loss.

[0074] The scanning angle range is determined by adding a supplementary angle proportional to the morphological complexity index to the base scanning angle (typically 30°). Specifically, the scanning angle range is equal to the base angle plus the morphological complexity index multiplied by the angle adjustment factor (ranged between 2 and 3 radians). Higher morphological complexity increases the scanning angle range.

[0075] The calculation of scanning density follows the principle of linear growth: scanning density is equal to the base density (usually 100 dots / mm²) multiplied by (1 plus the product of the density adjustment factor and the morphological complexity index), where the density adjustment factor ranges from 3 to 5.

[0076] Therefore, when the morphological complexity index increases, the scanning angle range and scanning density both increase accordingly; for deep pit and fissure molars with a morphological complexity of 0.4, the scanning angle range can reach 30° plus a supplementary angle of about 60°, totaling about 90°, and the scanning density can reach 2.6 times the basic density, thereby ensuring the data integrity deep in the pit and fissure even under complex occlusal surface morphology; and then by establishing a functional relationship between morphological complexity and scanning parameters, the scanning strategy is adjusted according to the tooth morphology, avoiding local data loss or over-scanning caused by fixed parameter scanning.

[0077] In an optional embodiment, a piecewise function can be used to determine the scanning parameters. When the morphological complexity is less than 0.15, the basic parameters are maintained; when it is between 0.15 and 0.3, the scanning parameters are linearly increased; when it is greater than 0.3, a parameter growth strategy, such as a quadratic function relationship, is adopted.

[0078] Step 230: Determine the laser power according to the scanning density so that the product of the laser power and the scanning density remains constant, and determine the corresponding exposure time according to the laser power.

[0079] Laser power and exposure time together determine the energy input during the scanning process. To ensure consistent signal intensity at different scanning densities, an inverse relationship between power and density must be established.

[0080] Laser power is determined according to the principle of constant energy density: the product of laser power and scan density is maintained at a constant value (typically 10,000 mW·mm² / spot). Therefore, laser power is equal to this constant value divided by the scan density. As scan density increases, laser power decreases accordingly to maintain a constant total energy input per unit area.

[0081] The relationship between exposure time and laser power is: exposure time equals the target energy value divided by the product of laser power and photoelectric conversion efficiency. The target energy value is usually 0.1mJ, and the photoelectric conversion efficiency is between 0.6-0.8.

[0082] Therefore, when the scanning density increases, the number of scanning points per unit area increases. To avoid overexposure, the laser power is reduced accordingly. At the same time, lower laser power requires a longer exposure time to accumulate sufficient signal strength. By maintaining the product of power and density constant, stable image quality can be obtained even in high-density scanning mode, and tooth tissue will not be damaged due to excessive local energy concentration, thereby ensuring the safety of the entire scanning process and the consistency of data quality.

[0083] In an optional embodiment, temperature feedback control can be introduced to monitor the tooth surface temperature in real time during the scanning process. When the temperature rises above a threshold, the laser power is automatically reduced or the scanning interval is increased to prevent thermal damage.

[0084] In another optional embodiment, a pulsed laser mode can be used. By switching from a continuous laser to a high-frequency pulse (e.g., 1 kHz), the instantaneous peak power can be increased while maintaining the average power, facilitating penetration of surface contaminants and obtaining a clear signal.

[0085] For example, when scanning a mandibular second molar, the mesial pit depth was measured to be 3.2 mm, the crown height was 8.5 mm, and the calculated morphological complexity index was 0.376. Based on the above method, the scanning angle range was determined to be approximately 84°, and the scanning density was 250 points / mm². Following the principle of constant energy density, the laser power was determined to be 40 mW, corresponding to an exposure time of approximately 3.6 ms. This parameter combination ensures complete scanning of deep pits and fissures while avoiding excessive energy input.

[0086] Through the above steps, when the morphological characteristics and optical properties of each tooth position are obtained, a matching scanning parameter combination can be generated based on quantitative analysis; compared with the traditional empirical parameter setting, the mapping relationship between morphological complexity and scanning parameters established in this application can not only be differentiated for the different characteristics of anterior teeth, premolars, and molars, but also ensure the consistency of data quality under different scanning conditions through the constant product relationship of power-density; furthermore, even in the face of unfavorable scanning conditions such as deep pits and grooves and high reflections, complete and accurate scanning data can be obtained through the coordinated adjustment of parameters.

[0087] Step 300: Collect tooth point cloud data according to the scanning parameter combination, and identify enamel reflective areas and shadow-blocked areas in the tooth point cloud data.

[0088] After determining the optimal scanning parameter combinations for different tooth positions, the actual data acquisition process must be carried out. Due to the complexity of the oral environment, even with optimized scanning parameters, data loss due to high enamel reflectivity and obstruction by adjacent teeth may still occur. Therefore, while acquiring point cloud data, it is necessary to identify these problematic areas to provide a basis for subsequent targeted processing.

[0089] In the embodiment of the present application, step 300 includes step 310 and step 320:

[0090] Step 310: Collect tooth point cloud data according to the scanning parameter combination.

[0091] Based on the scanning parameter combination determined in step 200, including the scanning angle range, scanning density, laser power, and exposure time, three-dimensional data acquisition of the tooth surface is performed. During the acquisition process, the scanning head moves around the tooth within a predetermined angle range. Laser or structured light is projected onto the tooth surface, and the reflected light is captured by the camera and converted into depth information.

[0092] Specifically, the scanning head starts from the buccal side of the tooth and performs a sector scan within a certain angle range. For anterior teeth, the scanning trajectory is relatively simple, mainly covering the labial surface, lingual surface and incisal edge; for molars, it is necessary to add multi-angle scanning from the occlusal surface to ensure data collection deep in the pits and fissures. The raw data collected at each scanning position includes a two-dimensional image and the corresponding depth value, and the three-dimensional coordinates are calculated by triangulation or time-of-flight method.

[0093] The collected point cloud data is stored as (x, y, z) coordinates, where x and y represent the horizontal position in the scanning coordinate system, and z represents the depth value. Each point also contains reflection intensity information, which reflects the surface's ability to reflect the scanning light and provides important information for subsequent identification of reflective areas.

[0094] In an optional embodiment, a continuous scanning mode can be used. The scan head moves continuously along a predetermined path, capturing data as a video stream. Real-time data splicing is achieved through inter-frame registration. This mode offers rapid scanning speeds and is suitable for cooperative patients.

[0095] In another optional implementation, a segmented scanning mode can be used. The dentition is divided into multiple segments, each scanned individually and then joined. This allows the patient to rest between scans, reducing discomfort caused by prolonged mouth opening.

[0096] Step 320: Identify enamel reflective areas and shadow occlusion areas in the tooth point cloud data.

[0097] Collected point cloud data inevitably contains quality issues, primarily manifesting as missing or anomalies in localized areas. These problem areas must be accurately identified and classified so that appropriate treatment strategies can be implemented. Enamel glare and adjacent tooth occlusion are two of the most common causes of these issues, each manifesting differently in point cloud data.

[0098] Step 320 includes steps 321 to 323:

[0099] Step 321: Analyze the point cloud density distribution and identify areas where the point cloud density is lower than a first density value as potential problem areas.

[0100] Point cloud density refers to the number of points per unit area. A properly scanned area should have a relatively uniform point cloud density. A significantly lower point cloud density in a particular area indicates a data collection issue.

[0101] Specifically, the scan area is first divided into a regular grid, typically 0.5mm x 0.5mm in size. The number of point clouds within each grid is counted, and the local point cloud density is calculated. A first density value is set at 50% of the average density of a normal area. When the point cloud density of a grid falls below this threshold, it is marked as a potential problem area.

[0102] Reasons for reduced point cloud density include: camera saturation due to excessive surface reflection, signal loss due to weak surface reflection, or physical obstruction that prevents the scanning light from reaching the target. These different reasons require further analysis to distinguish.

[0103] In an optional embodiment, different density threshold settings can be used. Based on the expected density distribution of different tooth positions, different density thresholds can be set for each area. For example, the density threshold for the smooth surface of anterior teeth can be set higher, while the threshold for the pit and fissure areas of molars can be appropriately lowered.

[0104] Step 322: Based on the relationship between the rate of change of the tooth surface normal vector and the distance between adjacent points, determine whether the potential problem area is data loss caused by enamel reflection or a shadow area caused by occlusion of adjacent teeth.

[0105] While both enamel reflections and adjacent tooth occlusions can reduce point cloud density, their geometric features differ significantly. By analyzing the geometric features surrounding potential problem areas, we can accurately determine the cause of the problem.

[0106] For the point cloud at the boundary of the potential problem area, calculate its surface normal vector. This normal vector is obtained by fitting a tangent plane to the local point cloud. Then, calculate the angle between the normal vectors of adjacent points, i.e., the rate of change of the normal vector. Also, measure the Euclidean distance between adjacent points.

[0107] When the normal vector change rate is small (angle less than 15°) and the distance between adjacent points is normal, it indicates that the surface of the area is relatively smooth, and the missing data is more likely due to overexposure caused by specular reflection of the enamel. This is common in areas with less curvature, such as the center of the labial surface of anterior teeth.

[0108] When the normal vector changes rapidly (angles greater than 30°) or the distances between adjacent points change suddenly, this indicates a geometric discontinuity in the area, and data loss is more likely due to physical obstruction by adjacent teeth or other structures. This is common in areas that are difficult to scan directly, such as adjacent areas and the lingual side.

[0109] Therefore, by comprehensively analyzing the density distribution and geometric characteristics of the point cloud, we can accurately distinguish data missing caused by different reasons; when it is identified as an enamel reflective area, it can be improved by reducing the local scanning power or adjusting the incident angle; when it is identified as an occlusion shadow, additional scanning is required from other angles.

[0110] In an optional embodiment, reflection intensity information can be introduced to assist in the judgment. Points around the enamel reflective area usually have abnormally high reflection intensity, while the reflection intensity at the edge of the occluded area is gradually distributed.

[0111] Step 323: Extract the gum line contour and identify the data collection blind area of ​​the tooth neck.

[0112] The tooth neck is the junction between the crown and the root. Due to its unique location, it is often a blind spot in scanning. Accurately identifying the gum line contour is crucial to obtaining complete crown data.

[0113] Gum line extraction relies on a comprehensive analysis of color and geometric features. In scanned data containing texture information, gum tissue appears pink, creating a stark contrast against the white teeth. Through color space conversion and threshold segmentation, a preliminary distinction can be made between the teeth and gum areas.

[0114] Geometrically, the gum line appears as a sudden change in curvature. Curvature along the long axis of the tooth is analyzed. The gum line is defined as the point where the curvature changes from a positive value (convex) to a negative value (concave). The extracted gum line forms a closed curve that encircles the tooth.

[0115] The blind area for cervical ridge data collection is defined as an area within 2 mm of the gum line where the point cloud density is less than 30% of the normal value. Complete data acquisition is difficult in these areas due to scanning angle limitations and soft tissue obstruction.

[0116] In an optional embodiment, multispectral imaging can be used to assist in gum line identification. Light sources of different wavelengths are used for separate illumination, and the difference in reflection of different wavelengths by teeth and gums is exploited to improve the accuracy of gum line extraction.

[0117] In another optional embodiment, the gum line can be identified based on a machine learning method, and a segmentation model can be trained using a large amount of labeled gum line data to achieve automated gum line extraction.

[0118] For example, when scanning a maxillary central incisor, about 50,000 point cloud data were collected according to the scanning parameter combination. Through density analysis, it was found that there was a low-density area of ​​about 3mm×2mm in the central area of ​​the labial surface, and the point cloud density of this area was only 20% of the surrounding area. Further analysis showed that the normal vector change rate of the boundary of this area was less than 10°, which was determined to be caused by enamel reflection. At the same time, another low-density area was detected in the mesial adjacent area, and the normal vector change rate of its boundary was 45°, which was determined to be blocked by adjacent teeth. Gum line extraction showed that there was a data collection blind area of ​​about 1.5mm wide on the labial and lingual sides of the tooth neck. The accurate identification of these problem areas provides a clear target for subsequent targeted treatment.

[0119] Through the above steps, when the scan is completed according to the optimized parameter combination, not only the basic point cloud data of the tooth surface is obtained, but also the type and location of data quality problems are accurately identified through density analysis, geometric feature judgment and gum line extraction; furthermore, even in the complex oral environment with adverse factors such as reflection and occlusion, the classification and identification of problem areas can provide a reliable basis for subsequent precise processing, thereby ensuring the integrity of the final three-dimensional model.

[0120] Step 400: De-glare the enamel reflective area, complete the data of the shadow-occluded area, and generate corrected tooth point cloud data.

[0121] Based on the problem areas identified in step 300, targeted treatment strategies are required. While both reflective enamel areas and shadowed areas exhibit missing data, their causes differ and require different treatment methods. For reflective areas, mathematical reconstruction is possible due to the relatively complete surrounding data. For shadowed areas, physical means are required to retrieve the missing data. Ultimately, data from various sources is fused to generate a complete tooth point cloud.

[0122] In the embodiment of the present application, step 400 includes steps 410 to 430:

[0123] Step 410: For the enamel reflective area, calculate the curvature change rate of the boundary point; if the curvature change rate is less than a predetermined ratio of the curvature change rate of the adjacent normal area, use an interpolation algorithm to reconstruct the missing surface data.

[0124] Enamel reflective areas typically appear on relatively smooth surfaces, such as the center of the labial surface of anterior teeth or the buccal surface of molars. These areas are characterized by gently changing curvatures, allowing for reasonable inference and reconstruction based on surrounding valid data.

[0125] First, the boundary points of the reflective area are identified. Boundary points are defined as points adjacent to the missing area with valid data. For each boundary point, a local surface is fitted using the point cloud within its neighborhood, and the principal curvatures k1 and k2 at that point are calculated. The curvature change rate is defined as the ratio of the difference in principal curvature between adjacent boundary points to the distance between the points.

[0126] The predetermined ratio is set to 0.3. This means that when the curvature change rate at a boundary point is less than 30% of the curvature change rate of the adjacent normal area (the area beyond 5 mm from the boundary), the reflective area is considered sufficiently smooth to be reconstructed using interpolation. This predetermined ratio is based on statistical analysis of extensive clinical data and ensures reconstruction accuracy while avoiding erroneous interpolation of complex surfaces.

[0127] The choice of interpolation algorithm depends on the shape and size of the missing area. For small missing areas (less than 2 mm in diameter), cubic spline interpolation is used to ensure second-order continuity of the reconstructed surface. For larger missing areas (2-5 mm in diameter), radial basis function interpolation is used to solve a system of linear equations to obtain a smooth reconstructed surface. During the interpolation process, boundary points are used as control points to generate a surface that passes through all control points with minimal curvature change.

[0128] In an optional implementation, a Poisson reconstruction method can be used to transform the surface reconstruction problem into solving the Poisson equation. The vector field is constructed through the normal vector information of the boundary points, the implicit surface is obtained by solution, and the reconstructed point cloud is obtained by isosurface extraction.

[0129] In another optional embodiment, a deep learning-based reconstruction method can be used. A large amount of complete dental scan data is used to train a generative model, learn the statistical laws of the tooth surface, and intelligently fill in the missing areas.

[0130] Step 420: For the shadow-blocked area, determine the number of angles and angle intervals for supplementary scanning based on the ratio of the area of ​​the shadow-blocked area to the amount of data reconstructed by the interpolation algorithm, and perform supplementary scanning to obtain supplementary tooth point cloud data.

[0131] Shadowed areas cannot be accurately reconstructed mathematically due to physical obstructions, and supplementary scanning from different angles is necessary. The supplementary scanning strategy needs to balance data integrity and scanning efficiency.

[0132] First, the area of ​​the shadowed region is calculated by projecting the missing area onto a standard plane and counting the number of grid cells. The region's suitability for interpolation reconstruction is then assessed, and the amount of data required for reconstruction using the interpolation algorithm is calculated. If the ratio of the shadowed region's area to the amount of data required for interpolation reconstruction is greater than 0.7, the region is too large or complex, resulting in significant interpolation errors and requiring additional scanning.

[0133] The determination of supplemental scan angles is based on occlusion analysis. Using the original scan's viewing angle information and the location of the occluded area, the feasible angle range within which the area can be observed is calculated. The number of angles is determined by the formula: the area of ​​the occluded area (mm²) divided by 5, rounded up, but not exceeding 5. The angles are evenly spaced within the feasible angle range.

[0134] When performing a supplemental scan, the scan head is adjusted to the calculated angular position, using the same power and exposure parameters as the original scan. However, the scan range is limited to the occluded area and its surrounding 5mm range to improve scanning efficiency. The data collected at each supplemental angle is transformed and unified to the original coordinate system.

[0135] Step 430: Fusing the surface data reconstructed by the interpolation algorithm, the supplemented tooth point cloud data, and the valid data in the tooth point cloud data to generate corrected tooth point cloud data.

[0136] Data fusion is a key step in generating a complete point cloud, requiring the seamless integration of data from different sources. The fusion process includes coordinate alignment, overlapping area processing, and density homogenization.

[0137] First, coordinate alignment is performed. The supplementary scanned data is aligned with the original data through feature point matching. Significant feature points in the overlapping area (such as tooth cusps and pit and fissure intersections) are selected as matching benchmarks, and the alignment parameters are optimized using the Iterative Closest Point (ICP) algorithm.

[0138] For overlapping areas, where multiple data sources exist for the same location, a weighted average is used for fusion. Weights are determined based on data quality: original valid data has a weight of 1.0, supplementary scan data has a weight of 0.9, and interpolated reconstructed data has a weight of 0.7. This weighted average yields the final coordinates and attributes for that location.

[0139] Density homogenization ensures that the fused point cloud has a consistent density distribution. Areas with excessively high density are downsampled, while areas with excessively low density are interpolated and encrypted, ultimately maintaining the overall point cloud density within the range of 100-150 points / mm².

[0140] Therefore, by distinguishing and processing data missing due to different causes, when a smooth reflective area is detected, the missing data can be reconstructed through a curvature-constrained interpolation algorithm to maintain the geometric continuity of the surface; when encountering a shadow area caused by physical occlusion, a supplementary scanning strategy is calculated to obtain real data from a feasible angle; in addition, multi-source data is generated through weighted fusion to generate a complete point cloud, which not only fills the gaps in the original scan but also ensures the accuracy and consistency of the data.

[0141] In an optional implementation, a confidence assessment mechanism can be introduced. Each point is assigned a confidence value, with the original data having the highest confidence, the supplementary scan data having the second highest, and the interpolated data having the lowest. Subsequent processing can then be selectively processed based on the confidence level.

[0142] For example, when processing the scan data of a maxillary first premolar, a 4mm×3mm reflective area was identified in the center of the buccal surface. The curvature change rate of its boundary point was calculated to be 0.02mm. -2 , only the normal area (0.08mm -2 ), meeting the interpolation requirements. Using radial basis function interpolation, with 12 boundary points as control points, approximately 1,200 missing points were reconstructed. Furthermore, a 6 mm² occlusion was discovered in the mesial adjacent region. This required three supplementary angles: scanning at 45°, 60°, and 75° from the mesial angle, yielding approximately 1,800 additional data points. Finally, through weighted fusion, a complete point cloud consisting of 65,000 points was generated, representing a 4.8% improvement in data integrity compared to the original 62,000 points.

[0143] Through the above processing steps, even if the original scan has reflection and occlusion problems, through targeted reconstruction and supplementation strategies, complete and accurate three-dimensional tooth data can be obtained in the end; this classification processing method not only improves data quality, but also avoids the efficiency loss caused by blind supplementary scanning.

[0144] Step 500: performing registration and fusion on the corrected tooth point cloud data based on tooth anatomical landmarks to generate a dental three-dimensional model containing the complete dentition.

[0145] After the above steps, the corrected point cloud data of each tooth has been obtained. However, these data may come from multiple scans at different angles, and the coordinate system of each scan is not completely consistent. In order to construct a complete three-dimensional model of the dentition, these scattered point cloud data need to be accurately aligned to a unified coordinate system. Traditional geometric feature-based alignment methods are prone to errors when processing smooth tooth surfaces, while tooth anatomical landmarks, as stable biological features, provide a reliable benchmark for precise alignment.

[0146] In the embodiment of the present application, the registration and fusion of the corrected tooth point cloud data based on the tooth anatomical landmarks in step 500 includes steps A1 to A4:

[0147] Step A1: Identify the cusp point, ridge line, and central fossa as tooth anatomical landmarks.

[0148] Dental anatomical landmarks are characteristic locations on tooth morphology, with relatively stable topological relationships across individuals. Accurately identifying these landmarks is a prerequisite for precise registration.

[0149] The identification of tooth cusp points is based on the detection of local height extremes. For each point in the point cloud data, its relative height within a certain neighborhood (usually a radius of 2mm) is calculated. When the height value of a point is greater than that of all other points in the neighborhood, it is marked as a potential tooth cusp point. Further verification is performed through curvature analysis. The real tooth cusp point should have a large positive Gaussian curvature (usually greater than 0.5mm). -2 For anterior teeth, the highest point of the incisal edge is primarily identified; for canines, a single cusp is identified; for premolars and molars, multiple cusps need to be identified, including buccal and lingual cusps.

[0150] The ridge line is identified by using a ridge extraction algorithm. First, the principal curvature direction of each point in the point cloud is calculated, and then the ridge line is traced along the direction of the maximum principal curvature to form a continuous ridge line. The ridge line appears as a ridge connecting the cusps, such as the oblique ridge and transverse ridge of the molar. By setting the curvature threshold (minimum principal curvature greater than 0.2mm), the ridge line is extracted. -1 ) and continuity constraints to extract the main dental ridge lines.

[0151] Identification of the central fossa is based on localized concavity detection. The occlusal surface is searched for points where the Gaussian curvature is negative and has the largest absolute value. These points are typically located at the deepest part of the fossa. For molars, the central, mesial, and distal fossa need to be identified; for premolars, the lowest point of the central sulcus is primarily identified.

[0152] In an optional embodiment, a template matching method can be used to identify anatomical landmarks. Standard anatomical landmark templates for different types of teeth are pre-established, and the templates are aligned with the actual point cloud through shape registration. The locations of the landmarks on the templates are the recognition results.

[0153] In another optional embodiment, deep learning methods can be used for automatic recognition. A large number of annotated tooth point clouds are used to train a 3D convolutional neural network to directly predict the locations of anatomical landmarks, thereby improving the degree of automation of recognition.

[0154] Step A2: Establish a tooth coordinate system based on tooth anatomical landmarks to determine the spatial position of each tooth.

[0155] To achieve accurate registration of the dentition, a local coordinate system must be established for each tooth. A coordinate system based on anatomical landmarks is biologically meaningful and stable across scans.

[0156] The coordinate system was established according to the following principles: For molars, the central fossa was used as the origin, the mesiodistal direction was the X-axis (from the mesial fossa to the distal fossa), the buccal-lingual direction was the Y-axis (from the lingual tip to the buccal tip), and the Z-axis was perpendicular to the occlusal plane. For anterior teeth, the midpoint of the labial surface was used as the origin, the mesiodistal direction was the X-axis, the labiolingual direction was the Y-axis, and the direction along the long axis of the tooth was the Z-axis.

[0157] Optimize the coordinate axis orientation using principal component analysis (PCA). Center the coordinates of the anatomical landmarks, calculate the eigenvectors of the covariance matrix, and use the eigenvector corresponding to the largest eigenvalue as the principal axis orientation. This method can reduce errors caused by manual selection.

[0158] The position and orientation of each tooth in its local coordinate system is parameterized as a 4×4 transformation matrix, which contains a 3×3 rotation matrix and a 3×1 translation vector. This transformation matrix describes the conversion relationship from the local coordinate system to the global coordinate system.

[0159] In an optional embodiment, dental arch curve constraints can be introduced. First, the dental arch curve of the entire dentition is fitted, and the position of each tooth is constrained to be near the dental arch curve to avoid unreasonable position deviation during the registration process.

[0160] Step A3: Perform overall registration of the dentition based on the proximity and occlusal relationships of adjacent teeth.

[0161] After the registration of a single tooth is completed, it is necessary to consider the relationship between the teeth and perform global optimization to ensure the coordination of the entire dentition.

[0162] Adjacency relationships are established based on contact point detection. Normally, adjacent teeth have contact points in the adjacent area. Contact is considered to exist when the closest distance between adjacent tooth surfaces is less than 0.1 mm. The normal vectors at the contact point should be opposite, indicating that the two surfaces are in contact.

[0163] The occlusal relationship is determined by the occlusal contact between the upper and lower teeth. In a normal occlusal state, the occlusal surfaces of the corresponding teeth in the upper and lower jaws should have multiple contact points. By simulating occlusal movement, the relative positions of the upper and lower jaws are adjusted to maximize the number of contact points and ensure a uniform distribution.

[0164] Global registration utilizes an energy minimization approach. The defined energy function includes: shape preservation energy for individual teeth (ensuring no tooth deformation), adjacency constraint energy (maintaining normal adjacency), occlusion constraint energy (maintaining normal occlusion), and smoothness energy (ensuring overall smoothness of the dentition). An iterative optimization algorithm minimizes the total energy to achieve the optimal registration result.

[0165] It should be noted that accurate registration of multi-source point cloud data can be achieved through the precise identification of anatomical landmarks and the establishment of a stable coordinate system based on them; through the constraints of adjacency and occlusion relationships, it is ensured that the aligned dentition is not only geometrically accurate but also conforms to the physiological characteristics of the oral cavity; even when faced with data from different scanning batches and different angles, this biological feature-based registration method can obtain coordinated and consistent results.

[0166] Step A4: When fusing multiple scan data, retain the data of the pit and fissure area and the tooth edge area.

[0167] After registration, the same tooth may have overlapping data from multiple scans. Data fusion needs to prioritize high-quality data in critical areas while ensuring integrity.

[0168] Pit and fissure areas and tooth margins are crucial reference areas for clinical diagnosis and treatment. Pit and fissure morphology is associated with caries risk, while the integrity of the margins influences the fit of restorations. Therefore, these areas are given special consideration during data fusion.

[0169] For the pit and fissure area, the identification method is to detect the linear depression with drastic curvature change on the occlusal surface. When the minimum principal curvature of a certain area is less than -0.5mm -1 If the points are distributed linearly, they are marked as pits and fissures. Within this area, the data source with the highest point cloud density and the least noise is selected as the final data.

[0170] The tooth edge is defined as the area within 1 mm of the tooth contour. The contour is determined by detecting the location where the normal vector changes suddenly. Data in this edge area directly impacts the quality of the 3D model boundary, so data sources with optimal scanning angles and no obstructions are preferred.

[0171] For other regions, a weighted averaging fusion strategy is used, with the weights determined based on the quality score of each scan. The quality score takes into account factors such as point cloud density, noise level, and scan completeness.

[0172] Furthermore, during the overall registration of the dentition, the present application may also simultaneously perform the following tests and analyses, including steps B1 to B3:

[0173] Step B1: Measure the distance between adjacent tooth anatomical landmarks. If the distance between adjacent tooth anatomical landmarks exceeds the width of a single normal tooth, it is determined to be a tooth missing area. Generate a three-dimensional boundary of the missing space based on the adjacent tooth morphology and the position of the opposing teeth.

[0174] During the registration process, tooth loss can be detected by analyzing the distribution of anatomical landmarks. In a normal dentition, the distance between corresponding anatomical landmarks of adjacent teeth (such as the mesial to distal contact points) should match the standard width of the teeth at that location.

[0175] Specifically, the standard width for anterior teeth is 7-9 mm, for premolars 7-8 mm, and for molars 10-12 mm. When the distance between adjacent anatomical landmarks exceeds 1.5 times the standard width, it is determined that a tooth is missing at that location.

[0176] Determining the three-dimensional boundaries of the missing space requires comprehensive consideration of multiple factors. First, the mesiodistal boundaries of the missing space are determined based on the morphology of the adjacent teeth, particularly the proximal contours facing the missing area. Next, the occlusal boundaries of the missing space are determined based on the occlusal morphology of the opposing teeth and the normal occlusal space (2-3 mm). Finally, the apical boundaries of the missing space are estimated based on the morphological trends of the alveolar bone.

[0177] The generated 3D boundary is represented as a closed surface and can be used for subsequent implant planning or restoration design. The smoothness of the boundary is ensured by B-spline surface fitting.

[0178] Step B2: Detect the area where the normal vector of the tooth surface is locally reversed, calculate the depth value of the locally reversed area, and if the ratio of the depth value to the average curvature radius of the tooth surface represents a concave morphology, mark the locally reversed area as a lesion area and calculate the three-dimensional volume of the lesion area.

[0179] Tooth surface lesions, such as caries and defects, appear geometrically as localized depressions. These abnormal areas can be automatically detected by analyzing changes in surface normal vectors.

[0180] Local normal vector reversal is detected by calculating the angle between the normal vectors of adjacent points. An angle greater than 90° indicates the presence of a depression. To exclude normal anatomical structures (such as pits and fissures), further analysis of the morphological characteristics of the depression is necessary.

[0181] The depth value is calculated by fitting a reference plane to the normal surface surrounding the concave area and calculating the vertical distance from the plane to the deepest point of the concave area. The mean radius of curvature is determined by analyzing the curvature of the entire tooth surface and reflects the overall morphological scale of the tooth.

[0182] When the ratio of the depth to the mean radius of curvature was greater than 0.15 and the shape of the concave area was irregular (different from regular pits and fissures), it was marked as a lesion. The lesion volume was calculated by triangulating the concave area and calculating the volume below the reference plane.

[0183] Step B3: Generate an evaluation index of the tooth health status based on the ratio of the three-dimensional volume of the lesion area to the total volume of the tooth.

[0184] Quantitative assessment of dental health status is important for clinical diagnosis and treatment planning. Volumetric analysis based on three-dimensional models provides an objective assessment method.

[0185] The total volume of a tooth is calculated by triangulating the point cloud of the entire crown to form a closed triangular mesh, and then calculating the volume enclosed by this mesh. For adults, statistically speaking, the normal volume range for different tooth positions is: 200-400 mm³ for anterior teeth, 400-600 mm³ for premolars, and 600-1000 mm³ for molars.

[0186] The health status assessment index is defined as: (1-lesion volume / total tooth volume) × 100%. When this index is greater than 95%, the tooth is considered healthy; 90%-95% is mild lesion; 85%-90% is moderate lesion; and less than 85% is severe lesion.

[0187] Furthermore, the location of the lesion is combined to generate a more detailed assessment. Occlusal lesions can affect mastication, interproximal lesions can affect adjacent teeth, and labial / buccal lesions can affect aesthetics. Lesion location is weighted to generate a comprehensive health index.

[0188] Preferably, through the simultaneous detection and analysis during the registration process, not only is the construction of the dentition 3D model completed, but also rich clinical diagnostic information is obtained; when tooth loss is detected, the automatically generated 3D boundary provides an accurate reference for restoration design; when surface lesions are found, quantitative volume analysis and health assessment provide an objective basis for treatment decisions; furthermore, this integrated processing method makes 3D scanning not only a tool for morphological recording, but also an important means of digital diagnosis.

[0189] In an optional implementation, artificial intelligence-assisted diagnosis can be introduced. Detected lesion features are input into a trained classification model to automatically determine the lesion type (caries, wear, defect, etc.), thereby improving diagnostic accuracy.

[0190] For example, when processing the full-mouth scan data of a patient, 68 landmarks (including 28 cusp points, 24 ridge lines, and 16 central pits) were identified in the maxilla through anatomical landmark point recognition. 14 local coordinate systems (one for each tooth) were established based on these landmark points. During the registration process, it was detected that the distance between the landmarks at positions 16 and 17 reached 18 mm, which exceeded 1.5 times the width of a normal molar. 16 positions were determined to be missing, and a missing space boundary with a volume of approximately 800 mm³ was generated. At the same time, an irregular depression with a depth of 2.1 mm was detected on the occlusal surface at position 36. Its volume was calculated to be 12 mm³, accounting for 1.6% of the total volume of 750 mm³ of the tooth. The health index was 98.4%, and it was diagnosed as mild occlusal caries. The final generated three-dimensional model of the dentition is not only morphologically complete, but also contains these important clinical information.

[0191] Through the processing of step 500, what is finally obtained is not only a geometrically accurate dental three-dimensional model, but also a digital oral health record containing rich diagnostic information, which provides comprehensive data support for subsequent treatment planning and long-term monitoring.

[0192] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0193] Based on the same inventive concept, embodiments of the present application also provide a dental 3D model data processing system. The solution provided by this system is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more dental 3D model data processing system embodiments provided below can be found in the aforementioned limitations of the dental 3D model data processing method and will not be further elaborated here.

[0194] In an exemplary embodiment, Figure 2 As shown, a data processing system for a dental three-dimensional model is provided, comprising a tooth recognition module, a scanning strategy module, a data acquisition module, a data correction module and a model construction module, wherein:

[0195] The tooth recognition module is used to obtain the initial data of the oral scan, identify the tooth type in the initial data, and determine the corresponding tooth surface reflection parameters according to the tooth type;

[0196] The scanning strategy module is used to generate scanning parameter combinations for different tooth positions based on tooth surface reflection parameters and tooth anatomical structure;

[0197] The data acquisition module is used to collect tooth point cloud data according to the scanning parameter combination and identify the enamel reflective area and shadow occlusion area in the tooth point cloud data;

[0198] The data correction module is used to perform de-reflection processing on the reflective area of ​​enamel, complete the data of the shadow-occluded area, and generate the corrected tooth point cloud data;

[0199] The model building module is used to align and fuse the corrected tooth point cloud data based on tooth anatomical landmarks to generate a dental three-dimensional model containing a complete dentition.

[0200] Each module in the aforementioned dental 3D model data processing system may be implemented in whole or in part through software, hardware, or a combination thereof. Each module may be embedded in or independent of a processor within a computer device in the form of hardware, or may be stored in a computer device memory in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0201] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 3 As shown. The computer device includes a processor, memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication, and the wireless communication can be achieved via Wi-Fi, a mobile cellular network, near-field communication (NFC), or other technologies. When executed by the processor, the computer program implements a method for processing data of a dental three-dimensional model. The display unit of the computer device is used to form a visually visible image, and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse.

[0202] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0203] In one embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0204] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0205] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0206] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0207] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), data processing logic based on quantum computing, artificial intelligence (AI) processors, and the like.

[0208] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0209] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A data processing method for a dental three-dimensional model, characterized in that: include: Acquiring initial data of an oral scan, identifying tooth types in the initial data, and determining corresponding tooth surface reflection parameters according to the tooth types; generating scanning parameter combinations for different tooth positions according to the tooth surface reflection parameters and tooth anatomical structure; Collecting tooth point cloud data according to the scanning parameter combination, and identifying enamel reflective areas and shadow-blocked areas in the tooth point cloud data; Performing de-reflection processing on the enamel reflective area and completing data on the shadow-blocked area to generate corrected tooth point cloud data; The corrected tooth point cloud data is registered and fused based on the tooth anatomical landmarks to generate a dental 3D model containing the complete dentition. The identifying the tooth type in the initial data and determining the corresponding tooth surface reflection parameter according to the tooth type includes: Identify the tooth position distribution of anterior teeth, premolars, and molars through morphological analysis; Determine the corresponding curvature distribution value according to the crown morphology of different teeth; Determining the surface reflection parameters of each tooth position by combining the curvature distribution value and the optical parameters of the tooth enamel; The de-reflection processing is performed on the enamel reflective area, and the data of the shadow-blocked area is completed to generate the corrected tooth point cloud data, including: For the enamel reflective area, calculating the curvature change rate of the boundary point; if the curvature change rate is less than a predetermined ratio of the curvature change rate of the adjacent normal area, reconstructing the missing surface data using an interpolation algorithm; For the shadow-blocked area, determining the number of angles and angle intervals for supplementary scanning according to the ratio of the area of ​​the shadow-blocked area to the amount of data reconstructed by the interpolation algorithm, and performing supplementary scanning to obtain supplementary tooth point cloud data; The surface data reconstructed by the interpolation algorithm, the supplemented tooth point cloud data and the valid data in the tooth point cloud data are fused to generate corrected tooth point cloud data.

2. The method for processing dental three-dimensional model data according to claim 1, wherein: Generating scanning parameter combinations for different tooth positions according to the tooth surface reflection parameters and tooth anatomical structure includes: The ratio of occlusal pit and fissure depth to crown height was calculated as an index of morphological complexity; Determining a scanning angle range and a scanning density based on the morphological complexity index; wherein the scanning density is a function of the morphological complexity index and a basic density value; The laser power is determined according to the scanning density so that the product of the laser power and the scanning density remains constant, and the corresponding exposure time is determined according to the laser power.

3. The method for processing dental three-dimensional model data according to claim 2, wherein: The identifying of the enamel reflective area and the shadow-blocked area in the tooth point cloud data includes: Analyze the point cloud density distribution and identify areas where the point cloud density is lower than a first density value as potential problem areas; Based on the relationship between the rate of change of the tooth surface normal vector and the distance between adjacent points, it is determined whether the potential problem area is a data loss caused by enamel reflection or a shadow area caused by occlusion of adjacent teeth; Extract the gum line contour and identify the data collection blind area at the tooth neck.

4. The method for processing dental three-dimensional model data according to claim 3, wherein: The registering and fusing of the corrected tooth point cloud data based on tooth anatomical landmarks includes: Identify the cusp point, ridge line, and central fossa as dental anatomical landmarks; Establishing a tooth coordinate system based on the tooth anatomical landmarks to determine the spatial position of each tooth; Perform overall registration of the dentition based on the adjacent teeth's proximity and occlusal relationships; When fusing multiple scans, data from the pit and fissure areas and tooth margins are retained.

5. The method for processing dental three-dimensional model data according to claim 4, wherein: Also includes: Measuring the distance between adjacent tooth anatomical landmarks. If the distance between adjacent tooth anatomical landmarks exceeds the width of a single normal tooth, it is determined to be a tooth-missing area. The three-dimensional boundary of the missing space is generated based on the adjacent tooth morphology and the position of the opposing teeth. Detecting an area where a normal vector of a tooth surface is locally reversed, calculating a depth value of the locally reversed area, and if a ratio of the depth value to an average radius of curvature of the tooth surface represents a concave morphology, marking the locally reversed area as a lesion area and calculating a three-dimensional volume of the lesion area; An evaluation index of the tooth health status is generated according to the ratio of the three-dimensional volume of the lesion area to the total volume of the tooth.

6. A dental three-dimensional model data processing system, using the dental three-dimensional model data processing method according to any one of claims 1 to 5, characterized in that: include: Teeth recognition module, scanning strategy module, data acquisition module, data correction module and model building module, wherein: The tooth recognition module is used to obtain initial data of the oral scan, identify the tooth type in the initial data, and determine the corresponding tooth surface reflection parameters according to the tooth type; The scanning strategy module is used to generate scanning parameter combinations for different tooth positions according to the tooth surface reflection parameters and tooth anatomical structure; The data acquisition module is used to collect tooth point cloud data according to the scanning parameter combination, and identify enamel reflective areas and shadow-blocked areas in the tooth point cloud data; The data correction module is used to perform a de-reflection process on the enamel reflective area, perform data completion on the shadow-blocked area, and generate corrected tooth point cloud data; The model building module is used to perform registration and fusion on the corrected tooth point cloud data based on tooth anatomical landmarks to generate a dental three-dimensional model containing a complete dentition.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the dental three-dimensional model data processing method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the dental three-dimensional model data processing method according to any one of claims 1 to 5 are implemented.

Images