Three-dimensional tooth model construction method, device, equipment and medium

Abstract

The invention relates to the technical field of dental digitization, and discloses a three-dimensional tooth model construction method, device, equipment and medium, and the method comprises the steps: obtaining a multi-frame oral cavity fringe pattern and an oral cavity color image, and carrying out fringe enhancement processing on the multi-frame oral cavity fringe pattern to obtain a multi-frame fringe enhancement pattern; extracting a stripe center point coordinate, and obtaining a code value of the stripe enhancement graph; performing mask area analysis on the oral cavity color image and the code value by using the segmentation model to obtain a tooth area mask; obtaining a target stripe center point coordinate and a target code value through data screening; performing data encapsulation on the target stripe center point coordinate, the target code value and the oral cavity scene color image, and sending oral cavity encapsulation data to a cloud node; performing real-time point cloud reconstruction based on the oral cavity encapsulation data to obtain three-dimensional point cloud data corresponding to the teeth of the target user; and performing point cloud optimization on the three-dimensional point cloud data to obtain a three-dimensional tooth model of the target user. According to the invention, the efficiency of constructing the three-dimensional tooth model can be improved.

Description

Technical Field

[0001] This invention relates to the field of dental digital technology, and in particular to a method, apparatus, device and medium for constructing a three-dimensional tooth model. Background Technology

[0002] Constructing a three-dimensional tooth model refers to the technical process of using optical scanning equipment to quickly and continuously acquire data on the hard and soft tissues such as teeth, gums, and jawbone in a patient's oral cavity, and then converting the acquired two-dimensional images or sequence data into a high-precision, interactive three-dimensional digital model through a series of calculations.

[0003] In the field of digital dental technology, the construction and real-time visualization of 3D tooth models are the core technologies supporting the design of dental restorations, orthodontic plans, and doctor-patient communication. However, traditional 3D tooth model reconstruction systems suffer from two key technical bottlenecks. On the one hand, if a purely local terminal processing architecture is adopted, the data collected by the oral scanning device is limited by the computing power of the terminal hardware (such as insufficient GPU performance of ordinary computers), and the processing time for a single frame often exceeds 50ms, making it difficult to break through 5fps in reconstruction frame rate, which cannot meet the needs of real-time clinical interaction. On the other hand, if a purely cloud processing architecture is adopted, the original image data must be uploaded to the cloud in its entirety. Affected by network bandwidth fluctuations, data transmission delays can easily lead to reconstruction stuttering, and the lack of edge preprocessing will cause the cloud to receive a large amount of non-target data, consuming additional computing resources and further reducing reconstruction efficiency.

[0004] Therefore, in the existing technology, traditional three-dimensional tooth model reconstruction systems lack an efficient edge-cloud collaborative architecture, resulting in insufficient real-time reconstruction, limited terminal interaction, and lack of multi-user collaboration capabilities, which in turn leads to low efficiency in the construction of three-dimensional tooth models. Summary of the Invention

[0005] This invention provides a method, apparatus, device, and medium for constructing a three-dimensional tooth model, the main purpose of which is to solve the problem of low efficiency in constructing a three-dimensional tooth model.

[0006] To achieve the above objectives, the present invention provides a method for constructing a three-dimensional tooth model, comprising: A preset oral scanner is used to acquire multiple frames of oral stripe patterns and oral color images of the target user. Stripe enhancement processing is performed on the multiple frames of oral stripe patterns to obtain multiple frames of stripe-enhanced images. Extract the coordinates of the center points of the stripes in the stripe enhancement images of multiple frames, and decode the position information of the stripe enhancement images of multiple frames to obtain the code value of the stripe enhancement images; A mask region analysis is performed on the oral cavity color image and the code value using a preset segmentation model to obtain the tooth region mask; Based on the tooth region mask, the coordinates of the stripe center point and the code value are filtered to obtain the target stripe center point coordinates and the target code value. The target stripe center point coordinates, the target code value, and the oral cavity scene color image are encapsulated according to a preset frame order to obtain oral cavity encapsulation data, and the oral cavity encapsulation data is sent to a preset cloud node; Real-time point cloud reconstruction is performed on the oral cavity encapsulation data in the cloud node to obtain the three-dimensional point cloud data corresponding to the target user's teeth; The three-dimensional point cloud data is optimized to obtain a three-dimensional tooth model of the target user.

[0007] Optionally, the step of performing stripe enhancement processing on multiple frames of the oral cavity stripe map to obtain multiple frames of enhanced stripe maps includes: The oral cavity stripe pattern is divided into multiple stripe images according to a preset segmentation rule; Calculate the minimum and maximum gray values ​​for each of the striped images; The minimum gray value among the minimum gray values ​​corresponding to multiple stripe images is selected as the global minimum gray value in the oral cavity stripe image; The maximum value among the maximum gray values ​​corresponding to multiple stripe images is selected as the global maximum gray value in the oral cavity stripe image; Based on the global maximum gray value and the global minimum gray value, gray-level enhancement processing is performed on each pixel in the multiple frames of the oral cavity stripe map to obtain a multi-frame stripe enhancement map.

[0008] Optionally, extracting the coordinates of the stripe center points in the stripe enhancement images across multiple frames includes: The grayscale profile data of each stripe in the stripe enhancement image of multiple frames is extracted and the position of the peak pixel in the grayscale profile data is located. The grayscale data at the peak pixel position is extracted using a preset window as a fitting sample; A logarithmic transformation is performed on the fitted samples to obtain the linear equation of the stripe enhancement pattern; The coordinates of the fringe center point of the fringe enhancement pattern are calculated based on the linear equation.

[0009] Optionally, the step of performing mask region analysis on the oral cavity color image and the code value using a preset segmentation model to obtain a tooth region mask includes: Based on a preset segmentation model, the oral color image and the code value are fused using multi-source features to obtain a classification probability map corresponding to the oral region of the target user. The classification probability map is thresholded to obtain a binary image; Connected component filtering is performed on the binary image to obtain the tooth region mask.

[0010] Optionally, the step of fusing multi-source features of the oral cavity color image and the code value based on a preset segmentation model to obtain a classification probability map corresponding to the oral cavity region of the target user includes: Multi-scale features in the oral cavity color image are extracted using the encoder in the preset segmentation model; The code value is vector-converted to obtain the code value feature vector; The multi-scale features are fused with the code value vector to obtain a fused feature vector. The fused feature vector is upsampled and normalized to obtain the classification probability map corresponding to the oral cavity region of the target user.

[0011] Optionally, the step of performing real-time point cloud reconstruction based on the oral cavity encapsulation data in the cloud node to obtain three-dimensional point cloud data corresponding to the target user's teeth includes: Obtain the preset calibration parameters from the oral scanner, and extract the coordinates of the target stripe center point of each frame of stripe enhancement image in the oral encapsulation data; Calculate the three-dimensional spatial coordinates corresponding to the center point of the target stripe based on the coordinates of the center point of the target stripe and the calibration parameters; Local three-dimensional point cloud data of the stripe enhancement map for each frame is generated based on the three-dimensional spatial coordinates; The local 3D point cloud data is stitched together according to a preset frame order to obtain dense 3D point cloud data.

[0012] Optionally, the step of optimizing the three-dimensional point cloud data to obtain the three-dimensional tooth model of the target user includes: The three-dimensional point cloud data is subjected to point cloud denoising processing to obtain denoised three-dimensional point cloud data; The denoised 3D point cloud data is subjected to point cloud segmentation processing to obtain point cloud sub-blocks; The adjacent point cloud sub-blocks are subjected to fine point cloud registration processing, and all point cloud sub-blocks are fused into registered 3D point cloud data according to the pose in the fine point cloud registration. Global consistency correction is performed on the registered 3D point cloud data to obtain corrected 3D point cloud data. The three-dimensional point cloud data for orthodontic treatment is reconstructed to obtain a mesh model of the tooth surface; The texture information in the oral cavity color image is mapped onto the surface of the tooth surface mesh model to obtain the three-dimensional tooth model of the target user.

[0013] To address the above problems, the present invention also provides a three-dimensional tooth model construction device, the device comprising: The oral striate pattern enhancement module is used to acquire multiple frames of oral striate patterns and oral color images of the target user using a preset oral scanner, and to perform striate enhancement processing on the multiple frames of oral striate patterns to obtain multiple frames of enhanced striate patterns; The stripe center point coordinate extraction module is used to extract the stripe center point coordinates from multiple frames of the stripe enhancement image and decode the position information of the multiple frames of the stripe enhancement image to obtain the code value of the stripe enhancement image; The tooth region masking analysis module is used to perform masking region analysis on the oral cavity color image and the code value using a preset segmentation model to obtain the tooth region mask. The data filtering module is used to filter the center point coordinates of the stripes and the code value according to the tooth region mask to obtain the center point coordinates of the target stripes and the target code value. The oral cavity encapsulation data encapsulation module is used to encapsulate the target stripe center point coordinates, the target code value, and the oral cavity scene color image according to a preset frame order to obtain oral cavity encapsulation data, and send the oral cavity encapsulation data to a preset cloud node; The three-dimensional point cloud data reconstruction module is used to perform real-time point cloud reconstruction based on the oral cavity encapsulation data in the cloud node to obtain the three-dimensional point cloud data corresponding to the target user's teeth. The 3D tooth model optimization module is used to optimize the 3D point cloud data to obtain the 3D tooth model of the target user.

[0014] To address the above problems, the present invention also provides an electronic device, the electronic device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the three-dimensional tooth model construction method described above.

[0015] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one computer program, which is executed by a processor in an electronic device to implement the three-dimensional tooth model construction method described above.

[0016] This invention, through acquiring multiple frames of oral ridge patterns and color images of a target user and enhancing the ridge patterns, significantly improves image contrast. By extracting the coordinates of the ridge center points and decoding their positional information, the geometric features and spatial codes of the ridges can be accurately obtained. Furthermore, a preset segmentation model is used to perform mask analysis on the color images and code values, automatically and accurately distinguishing tooth regions from non-target tissues and effectively filtering out invalid data. Screening of ridge center points and code values ​​based on the mask further ensures the effectiveness and relevance of subsequent data processing. The filtered target data is encapsulated frame by frame and uploaded to the cloud, achieving efficient collaboration between the edge and cloud, ensuring data integrity and temporal consistency. Real-time point cloud reconstruction is performed on cloud nodes based on the encapsulated data, fully utilizing the parallel computing capabilities of cloud GPUs to quickly generate dense and accurate 3D point clouds. Finally, by denoising, registering, globally optimizing, reconstructing surfaces, and mapping textures from the point cloud, a high-precision 3D tooth model is obtained, thereby improving the efficiency of 3D tooth model construction. Therefore, the 3D tooth model construction method, apparatus, device, and medium proposed in this invention can solve the problem of low efficiency in constructing 3D tooth models. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a method for constructing a three-dimensional tooth model according to an embodiment of the present invention. Figure 2 This is a flowchart illustrating a method for extracting the center point coordinates of stripes according to an embodiment of the present invention. Figure 3 This is a flowchart illustrating a tooth region masking analysis method provided in an embodiment of the present invention. Figure 4 A functional block diagram of a three-dimensional tooth model construction device provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of an electronic device for implementing the three-dimensional tooth model construction method according to an embodiment of the present invention.

[0018] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0020] This application provides a method for constructing a three-dimensional tooth model. The execution subject of the three-dimensional tooth model construction method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the three-dimensional tooth model construction method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0021] Reference Figure 1 The diagram shown is a flowchart illustrating a method for constructing a three-dimensional tooth model according to an embodiment of the present invention. In this embodiment, the method for constructing a three-dimensional tooth model includes: S1. Use a preset oral scanner to acquire multiple frames of oral stripe patterns and oral color images of the target user, and perform stripe enhancement processing on the multiple frames of oral stripe patterns to obtain multiple frames of stripe enhancement images.

[0022] In this embodiment of the invention, the preset oral scanner can be a high-precision oral optical scanning device that integrates a structured light projection module, a color image acquisition module and a data transmission module. It can simultaneously acquire structured light stripe images and real color images of areas such as teeth and gums in the oral cavity, and transmit the acquired raw data to a subsequent processing unit (such as a microcontroller development board).

[0023] In detail, the structured light projection module of the dental scanner projects pre-coded striped structured light into the target user's oral cavity. Simultaneously, it triggers the color image acquisition module and the stripe image acquisition module to work synchronously at a frame rate of 15fps. Through the scanner's USB 3.0 or Bluetooth connection channel, it receives the raw image data transmitted by the scanner in real time. Each set of images contains multiple frames of stripe patterns (e.g., 6 frames) and 1 frame of color image. Each received frame is labeled with a frame number according to the acquisition sequence. The acquired images then undergo image preprocessing, which includes image denoising (e.g., Gaussian filtering) and stripe enhancement, among other preprocessing methods.

[0024] Specifically, the oral ridge pattern is a grayscale image reflecting the surface morphology of the oral cavity, characterized primarily by alternating light and dark ridges. The oral color image is an RGB image reflecting the true color information of the oral cavity region, clearly showing the enamel color of teeth, the pink hue of gums, and the texture details of soft tissues. The ridge enhancement image is a grayscale image obtained by enhancing the original oral ridge pattern.

[0025] In this embodiment of the invention, the step of performing stripe enhancement processing on multiple frames of the oral cavity stripe pattern to obtain multiple frames of enhanced stripe patterns includes: The oral cavity stripe pattern is divided into multiple stripe images according to a preset segmentation rule; Calculate the minimum and maximum gray values ​​for each of the striped images; The minimum gray value among the minimum gray values ​​corresponding to multiple stripe images is selected as the global minimum gray value in the oral cavity stripe image; The maximum value among the maximum gray values ​​corresponding to multiple stripe images is selected as the global maximum gray value in the oral cavity stripe image; Based on the global maximum gray value and the global minimum gray value, gray-level enhancement processing is performed on each pixel in the multiple frames of the oral cavity stripe map to obtain a multi-frame stripe enhancement map.

[0026] In detail, the pixel range of each image block is calculated according to a preset block division rule (such as dividing it into 4 or 8 blocks). The pixel coordinates of the single frame stripe pattern are traversed in a loop, and the pixel data belonging to the same pixel range are extracted to form independent sub-image blocks. This process is repeated to process the oral stripe patterns of all frames to obtain multiple stripe images corresponding to each frame stripe pattern.

[0027] Next, for each sub-image, the minimum gray value variable is initialized to 255 (the maximum gray value of the grayscale image), and the maximum gray value variable is initialized to 0. By using a double loop to traverse each pixel in the sub-image, the gray value of each pixel is read. If the gray value of the current pixel is less than the initialized minimum gray value, the minimum gray value is updated to the current pixel gray value. Similarly, the maximum gray value is updated, thereby obtaining the minimum and maximum gray values ​​of each striped image.

[0028] Specifically, by iterating through and comparing the minimum grayscale value of each sub-image block, the smallest grayscale value is selected and determined as the global minimum grayscale value of the stripe pattern in that frame. Similarly, the global maximum grayscale value is obtained, and the global maximum grayscale value and global minimum grayscale value of each frame are determined.

[0029] Furthermore, the grayscale enhancement processing of each pixel in the multiple frames of the oral cavity stripe map can be performed using the following formula to obtain a multi-frame stripe enhancement map: in, For the output stripe enhancement pattern, This is the input oral cavity stripe pattern. This is the global minimum grayscale value. This is the global maximum grayscale value. Represents the x-coordinate of a pixel in the image. This represents the vertical coordinate of a pixel in the image.

[0030] Specifically, if equal If the original grayscale value of all pixels is replaced with the stretched grayscale value, the stripe enhancement map of that frame is obtained. The oral stripe maps of all frames are processed in the same way, ultimately resulting in a multi-frame stripe enhancement map. The stripe enhancement processing methods here include, but are not limited to, grayscale histogram stretching and Gamma correction.

[0031] In this embodiment of the invention, the synchronously acquired stripe pattern and color image can provide dual data support of structural information and color information for subsequent three-dimensional reconstruction, avoiding the problem of insufficient data dimension of a single image. Stripe enhancement processing can effectively improve the contrast of the stripe pattern and eliminate noise and grayscale unevenness interference in the original image.

[0032] S2. Extract the coordinates of the center points of the stripes in the stripe enhancement images of multiple frames, and decode the position information of the stripe enhancement images of multiple frames to obtain the code value of the stripe enhancement images.

[0033] In this embodiment of the invention, the coordinates of the stripe center point refer to the sub-pixel level coordinates corresponding to the peak value of the stripe grayscale distribution, calculated by algorithms such as Gaussian fitting in the stripe enhancement image. The code value is a decimal value obtained by decoding multiple frames of stripe enhancement images according to a preset encoding rule (such as binary stripe time encoding), which can uniquely identify the position of the stripe in the projection direction.

[0034] In the embodiments of the present invention, see Figure 2 As shown, the extraction of the center point coordinates of the stripes in the stripe enhancement image across multiple frames includes: S21. Traverse and extract the grayscale profile data of each stripe in the stripe enhancement image of multiple frames, and locate the peak pixel position in the grayscale profile data. S22. Use a preset window to extract the grayscale data of the peak pixel position as a fitting sample; S23. Perform a logarithmic transformation on the fitted samples to obtain the linear equation of the stripe enhancement pattern; S24. Calculate the coordinates of the center point of the fringe in the fringe enhancement pattern according to the linear equation.

[0035] In detail, for each frame of stripe enhancement image, first determine the stripe extension direction (e.g., horizontal extension), then traverse the image at 1-pixel intervals along the direction perpendicular to the stripe extension (vertical direction), extract the gray value sequence (i.e., the gray-scale profile data of the stripe) on each scan line, and for each gray-scale profile data, compare each gray value by iteratively traversing, record the pixel coordinates corresponding to the gray value with the largest value, and locate the peak pixel position of the gray-scale profile.

[0036] Specifically, for the peak pixel position of each stripe, a preset size of the cropping window (such as a window of ±3 pixels) is set. Centered on the peak pixel coordinates, the gray values ​​of all pixels within the window range are cropped. The cropped gray values ​​are stored in an array according to the pixel position order to form the fitting sample of the stripe.

[0037] Next, for each grayscale value in the fitted sample, first check if it is 0 (to avoid errors in logarithmic calculations). If it is 0, adjust it to 1. Perform a natural logarithmic operation on the adjusted grayscale values ​​to obtain logarithmically transformed grayscale data. Based on the characteristic that the stripe grayscale profile approximates a Gaussian distribution, the Gaussian distribution formula is as follows: in, Peak amplitude, Represents the pixel coordinates perpendicular to the direction of the stripe extension. The coordinates of the fringe center are Stripe width information Background brightness.

[0038] Furthermore, (Peak amplitude) (Stripe width) is obtained by Gaussian fitting of the stripe grayscale profile data. (Background brightness) is obtained by statistically analyzing the mean grayscale value of the area outside the stripes. The logarithm of the Gaussian formula is then taken and rearranged to obtain a linear equation, as follows: in, Represents grayscale profile data. Peak amplitude, Represents the pixel coordinates perpendicular to the direction of the stripe extension. The coordinates of the fringe center are Stripe width information.

[0039] Next, the linear equation is expanded as follows: ,set up , c The equation is transformed into Using least squares to fit a quadratic polynomial Fitting , , The center point can then be obtained by inverse solving using the coefficients. , and Combined with the scan line where the stripes are located coordinates, to obtain the complete coordinates of the fringe center point. .

[0040] Furthermore, the acquisition sequence reads multiple frames of stripe enhancement images (e.g., 3 frames of binary stripe images), and performs thresholding processing on each frame of stripe enhancement image (set to 1 if grayscale value > 128, otherwise set to 0) to obtain a binary image and extract the binary sequence of each stripe (e.g., the sequence of frame 1 is 100010, frame 2 is 001010, and frame 3 is 111111). The binary sequences of multiple frames are aligned according to the stripe position, and the binary states of the same stripe position in the multiple frames are combined into a binary code (e.g., (1,0,1)). The binary code is converted into a decimal value, which is the code value of the corresponding stripe. After traversing all stripe positions, the code value set of each frame is obtained.

[0041] In this embodiment of the invention, sub-pixel level stripe center point coordinates can significantly improve the spatial accuracy of subsequent 3D reconstruction, avoid reconstruction errors caused by pixel-level coordinates, and the generation of code values ​​can provide a unique identifier for stripe positions, ensuring the accuracy of subsequent projection coordinate mapping.

[0042] S3. Using a preset segmentation model, perform mask region analysis on the oral cavity color image and the code value to obtain the tooth region mask.

[0043] In this embodiment of the invention, the preset segmentation model is a deep learning model capable of pixel-level classification of oral cavity images. Commonly used models include the lightweight MobileNet network and the SegFormer segmentation model. The tooth region mask is a binary image, where regions with a pixel value of 1 correspond to tooth regions in the oral cavity, and regions with a pixel value of 0 correspond to non-tooth regions such as gums and soft tissues.

[0044] In the embodiments of the present invention, see Figure 3 As shown, the step of using a preset segmentation model to perform mask region analysis on the oral cavity color image and the code value to obtain the tooth region mask includes: S31. Based on a preset segmentation model, multi-source feature fusion is performed on the oral cavity color image and the code value to obtain a classification probability map corresponding to the oral cavity region of the target user; S32. Threshold the classification probability map to obtain a binary image; S33. Perform connected component filtering on the binary image to obtain the tooth region mask.

[0045] In detail, a large number of oral color images and corresponding manually labeled masks (labeling teeth, gums, and soft tissue categories) are obtained from the database and divided into training, validation, and test sets in an 8:1:1 ratio. The training set images are preprocessed by normalization (mapping pixel values ​​to 0-1) and size scaling (adapting to model input), and the code values ​​are converted into feature vectors as auxiliary inputs to the model. The preset segmentation model is iteratively trained using the training set data. After each round of training, the model performance (such as Dice coefficient) is evaluated using the validation set. The learning rate, weights, and other parameters are adjusted according to the evaluation results. Training stops when the Dice coefficient of the model on the test set is >0.9, resulting in the optimized segmentation model.

[0046] In this embodiment of the invention, the classification probability map is an image with the same resolution as the oral cavity color image, and each pixel location stores the probability value of the pixel belonging to a preset category such as teeth, gums, or soft tissue.

[0047] In this embodiment of the invention, the step of fusing multi-source features of the oral cavity color image and the code value based on a preset segmentation model to obtain a classification probability map corresponding to the oral cavity region of the target user includes: Multi-scale features in the oral cavity color image are extracted using the encoder in the preset segmentation model; The code value is vector-converted to obtain the code value feature vector; The multi-scale features are fused with the code value vector to obtain a fused feature vector. The fused feature vector is upsampled and normalized to obtain the classification probability map corresponding to the oral cavity region of the target user.

[0048] In detail, the optimized segmentation model encoder adopts a hierarchical structure. The first layer performs convolution operations on the oral color image to extract low-order features (such as edges and textures), and at the same time performs a fully connected operation on the code value feature vector to transform it into a vector that matches the image feature dimension. The middle layer continues to downsample the image features and extract high-order semantic features. Each layer concatenates and fuses the code value features with the corresponding layer's image features to obtain a fused feature vector.

[0049] Next, the fused features are input into the decoder. The decoder first upsamples the lowest-level semantic features to a higher resolution, and then concatenates and fuses them layer by layer with the corresponding level of shallow features from the encoder that have undergone convolution processing, gradually restoring spatial details. The decoder outputs a raw score containing all categories (such as teeth, gums, and background) for each pixel through a convolutional classification layer. The class score vector of each pixel output by the decoder is normalized by applying the Softmax function, compressing all elements of the vector to the (0,1) interval and making their sum equal to 1, thereby converting the value of each element into the probability that the pixel belongs to the corresponding category, and finally generating a pixel-level classification probability map corresponding to the spatial size of the input image.

[0050] Specifically, for the classification probability map, focus on the tooth category probability value of each pixel and set a preset threshold (such as 0.8); traverse each pixel in the probability map through a double loop. If the tooth category probability value of the pixel is greater than the threshold, set the pixel value to 1, otherwise set it to 0; after traversing all pixels, obtain a binary image containing only 0 and 1, where the 1 region is the initially determined tooth region.

[0051] Next, an eight-neighbor connected component analysis algorithm can be used to traverse all pixels in the binary image, mark all connected regions consisting of 1s, count the number of pixels in each connected region, set a minimum connected region pixel threshold (e.g., 50 pixels), remove small regions (noise points) with fewer pixels than the threshold, set the pixel values ​​of the retained connected regions to 1, and set the values ​​of the remaining pixels to 0, thus obtaining the final tooth region mask.

[0052] In this embodiment of the invention, the model can accurately segment and automatically distinguish between tooth and non-tooth regions, avoiding subjective errors and tedious operations of manual segmentation. The code value, as an auxiliary feature, can improve the segmentation accuracy, especially in areas where the boundary between teeth and gums is blurred, ensuring that the mask can accurately select the tooth range.

[0053] S4. Based on the tooth region mask, perform data filtering on the coordinates of the stripe center point and the code value to obtain the coordinates of the target stripe center point and the target code value.

[0054] In this embodiment of the invention, the target stripe center point coordinates and the target code value refer to the stripe center point coordinates and corresponding code values ​​retained only within the tooth region after filtering by the tooth region mask. The target stripe center point coordinates will have additional "region type information" (marked as "teeth") added to clarify the region to which they belong.

[0055] In detail, first, a pixel mapping relationship is established between the stripe center point coordinates and the tooth region mask (the coordinates (x, y) of the stripe center point in the stripe enhancement image correspond one-to-one with the pixels at coordinates (x, y) in the mask); by iterating through the set of stripe center point coordinates and corresponding code values ​​for each frame, for each stripe center point coordinate (x_c, y_c), the pixel value of coordinates (x_c, y_c) in the mask is read; if the pixel value is 1 (tooth region), the center point coordinates (with the "region type = tooth" identifier) ​​and the corresponding code value are stored in a temporary list; if the pixel value is 0 (non-tooth region), the center point coordinates and code value are directly removed; after the traversal is completed, the target stripe center point coordinates and target code values ​​are obtained.

[0056] In this embodiment of the invention, invalid data from non-dental areas such as gums and soft tissues is filtered out, reducing the volume of subsequent data encapsulation and transmission bandwidth usage. At the same time, invalid data is prevented from entering the cloud reconstruction process, which would lead to a decrease in accuracy and computational redundancy.

[0057] S5. Encapsulate the target stripe center point coordinates, the target code value, and the oral cavity scene color image according to the preset frame order to obtain oral cavity encapsulation data, and send the oral cavity encapsulation data to the preset cloud node.

[0058] In this embodiment of the invention, the oral cavity encapsulation data refers to structured data formed by integrating the coordinates of the target stripe center point (including region type information), the target code value, the color image of the oral cavity scene, and the frame number according to a preset structure. The preset cloud node refers to a cloud computing unit composed of a multi-node GPU server cluster, which has high-speed parallel computing capabilities and data storage functions, pre-stores the calibration parameters of the oral scanner and the reconstruction algorithm, and its core function is to receive the oral cavity encapsulation data transmitted from the edge and perform three-dimensional point cloud reconstruction and optimization.

[0059] In detail, multiple frames of data are collected and sorted by frame number (from smallest to largest). A structured data format is defined (including frame number field, target stripe center point coordinate array (including region type), target code value array, and oral scene color image pixel data field). Each field is filled according to the format to form a single frame data block. All single frame data blocks are integrated and compressed. An identifier and data length field are added to the header of the compressed data stream to form oral encapsulated data. A connection request is sent to a preset cloud node. After the connection is successful, the encapsulated data is sent and the cloud confirmation is waited for until the cloud receives the data and the data transmission is completed.

[0060] In this embodiment of the invention, structured encapsulation ensures data orderliness and integrity, avoids data corruption during transmission, compression reduces transmission bandwidth usage and improves transmission speed, and the high-performance computing capabilities of cloud nodes can handle subsequent intensive reconstruction tasks, solving the problem of insufficient computing power at the edge.

[0061] S6. Real-time point cloud reconstruction is performed on the cloud node based on the oral cavity encapsulation data to obtain the three-dimensional point cloud data corresponding to the target user's teeth.

[0062] In this embodiment of the invention, the three-dimensional point cloud data refers to a dataset composed of a large number of three-dimensional spatial coordinates (x, y, z) of the tooth surface. Each coordinate point corresponds to a sampling point on the tooth surface. The density of the points is determined by the number of fringe center points (usually a dense point cloud), which can accurately reflect the three-dimensional morphology and spatial structure of the tooth.

[0063] In this embodiment of the invention, the step of performing real-time point cloud reconstruction based on the oral cavity encapsulation data in the cloud node to obtain the three-dimensional point cloud data corresponding to the target user's teeth includes: Obtain the preset calibration parameters from the oral scanner, and extract the coordinates of the target stripe center point of each frame of stripe enhancement image in the oral encapsulation data; Calculate the three-dimensional spatial coordinates corresponding to the center point of the target stripe based on the coordinates of the center point of the target stripe and the calibration parameters; Local three-dimensional point cloud data of the stripe enhancement map for each frame is generated based on the three-dimensional spatial coordinates; The local 3D point cloud data is stitched together according to a preset frame order to obtain dense 3D point cloud data.

[0064] Specifically, the cloud node reads the preset calibration parameters of the dental scanner from the local database, including the camera intrinsic parameter matrix (focal length). , Principal point coordinates , Camera extrinsic parameters (rotation matrix R, translation vector T) and the equation of the light plane Analyze the oral cavity packaging data and extract the coordinates of the center point of the target stripe in each frame. Based on the camera's intrinsic parameters, using the formula , ( v , (Using three-dimensional coordinates) Derive the direction vector of the imaging ray, and calculate the coordinates of the intersection point between the ray and the light plane using the equation of the light plane. The intersection point is the three-dimensional spatial coordinate corresponding to the center point of the target stripe. The coordinates of all center points are calculated frame by frame.

[0065] Specifically, the cloud nodes utilize the parallel computing capabilities of a multi-node GPU server cluster to allocate independent GPU computing units to each frame of data. Each unit traverses the coordinates of the target stripe center point in the frame in pixel order. For each center point coordinate, the GPU unit performs parallel 3D spatial coordinate solving (ray back projection + light plane intersection calculation). After all center points in each frame are calculated, a local 3D point cloud for that frame is formed. The cloud then stitches together all local point clouds in frame order, removes duplicate coordinate points, and finally forms dense 3D point cloud data covering the tooth area.

[0066] In this embodiment of the invention, the parallel computing capability of the cloud GPU cluster can ensure that the reconstruction frame rate is maintained above 15fps, and the calculation of three-dimensional coordinates in combination with calibration parameters can ensure the accuracy of point cloud, avoid spatial position deviation caused by lack of calibration, and the dense point cloud data can completely present the details of the tooth surface.

[0067] S7. Perform point cloud optimization on the three-dimensional point cloud data to obtain the three-dimensional tooth model of the target user.

[0068] In this embodiment of the invention, the three-dimensional tooth model refers to a digital model that can realistically reflect the three-dimensional morphology and color characteristics of teeth, obtained by performing surface reconstruction and texture mapping on optimized three-dimensional point cloud data.

[0069] In this embodiment of the invention, the step of optimizing the three-dimensional point cloud data to obtain the three-dimensional tooth model of the target user includes: The three-dimensional point cloud data is subjected to point cloud denoising processing to obtain denoised three-dimensional point cloud data; The denoised 3D point cloud data is subjected to point cloud segmentation processing to obtain point cloud sub-blocks; The adjacent point cloud sub-blocks are subjected to fine point cloud registration processing, and all point cloud sub-blocks are fused into registered 3D point cloud data according to the pose in the fine point cloud registration. Global consistency correction is performed on the registered 3D point cloud data to obtain corrected 3D point cloud data. The three-dimensional point cloud data for orthodontic treatment is reconstructed to obtain a mesh model of the tooth surface; The texture information in the oral cavity color image is mapped onto the surface of the tooth surface mesh model to obtain the three-dimensional tooth model of the target user.

[0070] In detail, point cloud denoising processing of 3D point cloud data in cloud nodes can be performed using statistical filtering algorithms. For each point in the 3D point cloud data, the average distance and standard deviation of its 50 neighboring points are calculated. A distance threshold is set as the average distance + 2 × standard deviation. If the average distance between a point and its neighboring points is greater than the threshold, it is determined to be a noise point and removed. Denoising is completed by traversing all points to obtain denoised 3D point cloud data with isolated noise points removed.

[0071] Specifically, using a density clustering algorithm, a cluster radius (e.g., 2mm) and a minimum point count threshold (e.g., 10) are set. Each unlabeled point in the denoised point cloud is traversed. If the number of points in the neighborhood of a point is greater than the threshold, it is used as a core point to expand and form a cluster. This process is repeated to divide the point cloud into multiple clusters, each corresponding to a tooth region. The point cloud data of each cluster is extracted as an independent point cloud sub-block.

[0072] Next, the cloud node estimates the relative pose (rotation matrix, translation vector) of adjacent point cloud sub-blocks through a deep neural network, which is used as the initial value for the iterative nearest point algorithm. The initial pose is substituted into the iterative nearest point algorithm to calculate the distance error between corresponding points of adjacent sub-blocks. The pose is iteratively adjusted to minimize the error (convergence threshold 0.1mm). After all sub-blocks are finely registered, they are stitched together to form complete point cloud data, resulting in registered 3D point cloud data.

[0073] Furthermore, fast point feature histogram feature descriptors for each frame in the point cloud can be extracted and updated through cloud nodes. The feature similarity between adjacent and non-adjacent frames can be calculated. If the similarity is >0.85, it is determined to be a loop. A pose graph is constructed (the frame pose is a node, and the loop is an edge). The pose graph is optimized using a nonlinear least squares algorithm to eliminate accumulated errors. The point cloud coordinates are adjusted according to the optimized pose to obtain globally consistent corrected 3D point cloud data.

[0074] Next, the cloud node uses the Poisson reconstruction algorithm to estimate the normal vector and unify the direction of the correction point cloud, constructs the three-dimensional Poisson equation, solves the equation with the point cloud data as constraints to obtain the three-dimensional implicit function, uses the moving cube algorithm to extract the isosurface of the implicit function, generates the tooth surface triangular mesh, simplifies the mesh (removes redundant patches) and smooths it, repairs mesh defects, and obtains the tooth surface mesh model.

[0075] Furthermore, the cloud node establishes a mapping relationship between the vertices of the mesh model and the color image of the oral cavity. Based on the 3D coordinates of the vertices, it infers their pixel coordinates in the color image, reads the RGB color value of the corresponding pixel, assigns it to the mesh vertex, calculates the color value of the points inside the mesh patch through bilinear interpolation, making the texture transition natural, and saves the textured mesh model as an STL format to obtain the 3D tooth model of the target user.

[0076] Next, after obtaining the textured 3D tooth mesh model, the cloud node converts it into a lightweight format suitable for web browsers and streams it over the network to all connected collaborative terminals. The cloud establishes a real-time synchronization service that continuously listens for and receives interactive operation commands from any user terminal (such as the browser of a doctor or patient), updates these operations uniformly on the cloud model state, and immediately distributes the state changes to the terminals of all other online users.

[0077] In this embodiment of the invention, point cloud optimization can eliminate noise and accumulated errors, ensuring the geometric accuracy of the model. Surface reconstruction and texture mapping enable the model to have both accurate structure and realistic visual effects, making it easier for doctors to intuitively observe lesions (such as tooth decay and defects) and also helping patients understand their condition.

[0078] In this embodiment of the invention, multiple frames of oral striate patterns and color images are acquired and enhanced using a preset oral scanner. High-precision striate center point coordinates and position code values ​​are extracted, and the data is filtered by combining the tooth region mask generated by the preset segmentation model to ensure that only valid tooth region information is retained. Subsequently, the target data is encapsulated and transmitted to cloud nodes for real-time point cloud reconstruction and optimization, and finally a three-dimensional tooth model is generated, which improves the efficiency of constructing a three-dimensional tooth model.

[0079] like Figure 4 The diagram shown is a functional block diagram of a three-dimensional tooth model construction device provided in an embodiment of the present invention.

[0080] The three-dimensional tooth model construction device 400 of this invention can be installed in an electronic device. Depending on the functions implemented, the three-dimensional tooth model construction device 400 may include an oral striate enhancement module 401, a striate center point coordinate extraction module 402, a tooth region mask analysis module 403, a data filtering module 404, an oral striate encapsulation data encapsulation module 405, a three-dimensional point cloud data reconstruction module 406, and a three-dimensional tooth model optimization module 407. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, stored in the memory of the electronic device.

[0081] In this embodiment, the functions of each module / unit are as follows: The oral striate enhancement module 401 is used to acquire multiple frames of oral striate images and oral color images of a target user using a preset oral scanner, and to perform striate enhancement processing on the multiple frames of oral striate images to obtain multiple frames of enhanced striate images. The stripe center point coordinate extraction module 402 is used to extract the stripe center point coordinates in multiple frames of the stripe enhancement image and decode the position information of the multiple frames of the stripe enhancement image to obtain the code value of the stripe enhancement image. The tooth region masking analysis module 403 is used to perform masking region analysis on the oral cavity color image and the code value using a preset segmentation model to obtain the tooth region mask. The data filtering module 404 is used to filter the center point coordinates of the stripe and the code value according to the tooth region mask to obtain the center point coordinates of the target stripe and the target code value. The oral cavity encapsulation data encapsulation module 405 is used to encapsulate the target stripe center point coordinates, the target code value, and the oral cavity scene color image according to a preset frame order to obtain oral cavity encapsulation data, and send the oral cavity encapsulation data to a preset cloud node; The three-dimensional point cloud data reconstruction module 406 is used to perform real-time point cloud reconstruction based on the oral cavity encapsulation data in the cloud node to obtain the three-dimensional point cloud data corresponding to the target user's teeth. The 3D tooth model optimization module 407 is used to optimize the 3D point cloud data to obtain the 3D tooth model of the target user.

[0082] In detail, each module in the three-dimensional tooth model construction device 400 described in this embodiment of the invention uses the same technical means as the three-dimensional tooth model construction method described in the above figures, and can produce the same technical effect, which will not be repeated here.

[0083] like Figure 5 The diagram shown is a schematic diagram of an electronic device for implementing a three-dimensional tooth model construction method according to an embodiment of the present invention.

[0084] The electronic device 500 may include a processor 501, a memory 502, a communication bus 503, and a communication interface 504. It may also include a computer program, such as a three-dimensional tooth model construction program, stored in the memory 502 and capable of running on the processor 501.

[0085] In some embodiments, the processor 501 may be composed of an integrated circuit. The processor 501 is the control unit of the electronic device. It connects various components of the electronic device through various interfaces and lines. It performs various functions of the electronic device and processes data by running or executing programs or modules stored in the memory 502 (such as executing a three-dimensional tooth model construction program) and calling data stored in the memory 502.

[0086] The memory 502 includes at least one type of readable storage medium, including flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, disk, optical disk, etc. The memory 502 can be used not only to store application software and various types of data installed in electronic devices, such as the code of a 3D tooth model construction program, but also to temporarily store data that has been output or will be output.

[0087] The communication bus 503 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. This bus can be divided into an address bus, a data bus, a control bus, etc. The bus is configured to enable communication between the memory 502 and at least one processor 501, etc.

[0088] The communication interface 504 is used for communication between the above-mentioned electronic device and other devices, including a network interface and a user interface.

[0089] The figure only shows an electronic device with components. Those skilled in the art will understand that the structure shown in the figure does not constitute a limitation on the electronic device and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0090] For example, although not shown, the electronic device may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 501 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0091] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following: A preset oral scanner is used to acquire multiple frames of oral stripe patterns and oral color images of the target user. Stripe enhancement processing is performed on the multiple frames of oral stripe patterns to obtain multiple frames of stripe-enhanced images. Extract the coordinates of the center points of the stripes in the stripe enhancement images of multiple frames, and decode the position information of the stripe enhancement images of multiple frames to obtain the code value of the stripe enhancement images; A mask region analysis is performed on the oral cavity color image and the code value using a preset segmentation model to obtain the tooth region mask; Based on the tooth region mask, the coordinates of the stripe center point and the code value are filtered to obtain the target stripe center point coordinates and the target code value. The target stripe center point coordinates, the target code value, and the oral cavity scene color image are encapsulated according to a preset frame order to obtain oral cavity encapsulation data, and the oral cavity encapsulation data is sent to a preset cloud node; Real-time point cloud reconstruction is performed on the oral cavity encapsulation data in the cloud node to obtain the three-dimensional point cloud data corresponding to the target user's teeth; The three-dimensional point cloud data is optimized to obtain a three-dimensional tooth model of the target user.

[0092] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0093] Specifically, the specific implementation method of the processor 501 of the above instructions can be referred to the description of the relevant steps in the corresponding embodiment of the accompanying drawings, and will not be repeated here.

[0094] Furthermore, if the modules / units integrated in the electronic device 500 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0095] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for constructing a three-dimensional tooth model, characterized in that, The method includes: A preset oral scanner is used to acquire multiple frames of oral stripe patterns and oral color images of the target user. Stripe enhancement processing is performed on the multiple frames of oral stripe patterns to obtain multiple frames of stripe enhancement images. Extract the coordinates of the center points of the stripes in the stripe enhancement images of multiple frames, and decode the position information of the stripe enhancement images of multiple frames to obtain the code value of the stripe enhancement images; A mask region analysis is performed on the oral cavity color image and the code value using a preset segmentation model to obtain the tooth region mask; Based on the tooth region mask, the coordinates of the stripe center point and the code value are filtered to obtain the target stripe center point coordinates and the target code value. The target stripe center point coordinates, the target code value, and the oral cavity scene color image are encapsulated according to a preset frame order to obtain oral cavity encapsulation data, and the oral cavity encapsulation data is sent to a preset cloud node; Real-time point cloud reconstruction is performed on the oral cavity encapsulation data in the cloud node to obtain the three-dimensional point cloud data corresponding to the target user's teeth; The three-dimensional point cloud data is optimized to obtain a three-dimensional tooth model of the target user.

2. The method for constructing a three-dimensional tooth model as described in claim 1, characterized in that, The step of performing stripe enhancement processing on multiple frames of the oral cavity stripe map to obtain multiple frames of enhanced stripe maps includes: The oral cavity stripe pattern is divided into multiple stripe images according to a preset segmentation rule; Calculate the minimum and maximum gray values ​​for each of the striped images; The minimum gray value among the minimum gray values ​​corresponding to multiple stripe images is selected as the global minimum gray value in the oral cavity stripe image; The maximum value among the maximum gray values ​​corresponding to multiple stripe images is selected as the global maximum gray value in the oral cavity stripe image; Based on the global maximum gray value and the global minimum gray value, gray-level enhancement processing is performed on each pixel in the multiple frames of the oral cavity stripe map to obtain a multi-frame stripe enhancement map.

3. The method for constructing a three-dimensional tooth model as described in claim 1, characterized in that, The extraction of the center point coordinates of the stripes in the stripe enhancement image from multiple frames includes: The grayscale profile data of each stripe in the stripe enhancement image of multiple frames is extracted and the position of the peak pixel in the grayscale profile data is located. The grayscale data at the peak pixel position is extracted using a preset window as a fitting sample; A logarithmic transformation is performed on the fitted samples to obtain the linear equation of the stripe enhancement pattern; The coordinates of the fringe center point of the fringe enhancement pattern are calculated based on the linear equation.

4. The method for constructing a three-dimensional tooth model as described in claim 1, characterized in that, The step of performing mask region analysis on the oral cavity color image and the code value using a preset segmentation model to obtain a tooth region mask includes: Based on a preset segmentation model, the oral color image and the code value are fused using multi-source features to obtain a classification probability map corresponding to the oral region of the target user. The classification probability map is thresholded to obtain a binary image; Connected component filtering is performed on the binary image to obtain the tooth region mask.

5. The method for constructing a three-dimensional tooth model as described in claim 4, characterized in that, The step of fusing multi-source features of the oral cavity color image and the code value based on a preset segmentation model to obtain a classification probability map corresponding to the oral cavity region of the target user includes: Multi-scale features in the oral cavity color image are extracted using the encoder in the preset segmentation model; The code value is vector-converted to obtain the code value feature vector; The multi-scale features are fused with the code value vector to obtain a fused feature vector. The fused feature vector is upsampled and normalized to obtain the classification probability map corresponding to the oral cavity region of the target user.

6. The method for constructing a three-dimensional tooth model as described in claim 1, characterized in that, The step of performing real-time point cloud reconstruction based on the oral cavity encapsulation data in the cloud node to obtain the three-dimensional point cloud data corresponding to the target user's teeth includes: Obtain the preset calibration parameters from the oral scanner, and extract the coordinates of the target stripe center point of each frame of stripe enhancement image in the oral encapsulation data; Calculate the three-dimensional spatial coordinates corresponding to the center point of the target stripe based on the coordinates of the center point of the target stripe and the calibration parameters; Local three-dimensional point cloud data of the stripe enhancement map for each frame is generated based on the three-dimensional spatial coordinates; The local 3D point cloud data is stitched together according to a preset frame order to obtain dense 3D point cloud data.

7. The method for constructing a three-dimensional tooth model as described in claim 1, characterized in that, The step of optimizing the three-dimensional point cloud data to obtain the three-dimensional tooth model of the target user includes: The three-dimensional point cloud data is subjected to point cloud denoising processing to obtain denoised three-dimensional point cloud data; The denoised 3D point cloud data is subjected to point cloud segmentation processing to obtain point cloud sub-blocks; The adjacent point cloud sub-blocks are subjected to fine point cloud registration processing, and all point cloud sub-blocks are fused into registered 3D point cloud data according to the pose in the fine point cloud registration. Global consistency correction is performed on the registered 3D point cloud data to obtain corrected 3D point cloud data. The three-dimensional point cloud data for orthodontic treatment is reconstructed to obtain a mesh model of the tooth surface; The texture information in the oral cavity color image is mapped onto the surface of the tooth surface mesh model to obtain the three-dimensional tooth model of the target user.

8. A three-dimensional tooth model construction device, characterized in that, The device includes: The oral striate pattern enhancement module is used to acquire multiple frames of oral striate patterns and oral color images of the target user using a preset oral scanner, and to perform striate enhancement processing on the multiple frames of oral striate patterns to obtain multiple frames of enhanced striate patterns; The stripe center point coordinate extraction module is used to extract the stripe center point coordinates from multiple frames of the stripe enhancement image and decode the position information of the multiple frames of the stripe enhancement image to obtain the code value of the stripe enhancement image; The tooth region masking analysis module is used to perform masking region analysis on the oral cavity color image and the code value using a preset segmentation model to obtain the tooth region mask. The data filtering module is used to filter the center point coordinates of the stripe and the code value according to the tooth region mask to obtain the center point coordinates of the target stripe and the target code value. The oral cavity encapsulation data encapsulation module is used to encapsulate the target stripe center point coordinates, the target code value, and the oral cavity scene color image according to a preset frame order to obtain oral cavity encapsulation data, and send the oral cavity encapsulation data to a preset cloud node; The three-dimensional point cloud data reconstruction module is used to perform real-time point cloud reconstruction based on the oral cavity encapsulation data in the cloud node to obtain the three-dimensional point cloud data corresponding to the target user's teeth. The 3D tooth model optimization module is used to optimize the 3D point cloud data to obtain the 3D tooth model of the target user.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the three-dimensional tooth model construction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the three-dimensional tooth model construction method as described in any one of claims 1 to 7.

Images