Intelligent electronic tongue depressor and high-precision three-dimensional oral imaging method

Abstract

The present application relates to the technical field of oral imaging, in particular to an intelligent electronic tongue depressor and a high-precision three-dimensional oral imaging method. According to the structural characteristics of different frames of optical slice images, the local deformation possibility of each matching feature point in each frame of optical slice image is obtained, and the matching error feature points constituting the deformation target area are screened out. For any frame of optical slice image, according to the local deformation possibility of different matching feature points in the deformation target area, the position in the three-dimensional point cloud model, and the change trend of the corresponding moving path when the tongue depressor intersects with the deformation target area, the matching error consistency of the deformation target area on each frame of optical slice image is obtained. Combined with the position distribution of the point cloud coordinates of the matching error feature points in the deformation target area and the deformation center, the point cloud correction coordinates of the matching error feature points are obtained. The present application accurately obtains the point cloud correction coordinates of each image deformation area, and improves the accuracy of the construction of the three-dimensional oral model.

Description

Technical Field

[0001] This invention relates to the field of oral imaging technology, specifically to an intelligent electronic tongue depressor and a high-precision three-dimensional oral imaging method. Background Technology

[0002] A tongue depressor is used to open the oral cavity and obtain a visual field of the pharynx or tongue side. The tongue depressor is used to press down on the tongue for visual examination of the oral cavity and pharynx. In existing technologies, feature points are rigidly registered based on optical slice images acquired by an electronic tongue depressor to obtain a three-dimensional image of the patient's oral cavity. However, due to the contact between the tongue depressor and the soft tissues of the oral cavity, the soft tissues undergo elastic deformation, causing the same structure to exhibit significant morphological inconsistencies in different optical slice images. This leads to ghosting or tearing during feature registration, thus affecting the accuracy of three-dimensional oral cavity modeling. Summary of the Invention

[0003] To address the technical problem of significant morphological inconsistencies between different slices affecting the accuracy of oral cavity modeling, the present invention aims to provide an intelligent electronic tongue depressor and a high-precision three-dimensional oral cavity imaging method. The specific technical solution adopted is as follows:

[0004] This invention proposes an intelligent electronic tongue depressor, including a tongue depressor body, a controller, an image acquisition unit and a position sensor connected to the controller via signals. The image acquisition unit is used to acquire optical slice images of the oral cavity, and the position sensor is used to acquire the movement path of the tongue depressor. The control method of the controller includes:

[0005] Obtain optical slice images of different frames within the oral cavity along the movement path of the tongue depressor;

[0006] Based on the structural features of different frames of optical slice images, matching feature points that constitute the 3D point cloud model are obtained; based on the position and local structural features of the matching feature points between different frames of optical slice images, the local deformation probability of each matching feature point in each frame of optical slice image is obtained, and the matching error feature points that constitute the deformation target area are screened out.

[0007] For any frame of optical slice image, the deformation centroid of each deformation target region is obtained based on the local deformation probability of different matching feature points within the deformation target region and their positions in the 3D point cloud model; the movement path when the depressor intersects with the deformation target region is obtained; and the matching error consistency of the deformation target region on each frame of optical slice image is obtained based on the difference distribution of the deformation centroids within the deformation target region between adjacent frames of optical slice images and the changing trend of the corresponding movement path.

[0008] Based on the consistency distribution of matching error in the deformed target region on different frames of optical slice images, as well as the point cloud coordinates of matching error feature points and the position distribution of the deformation centroid within the deformed target region, the point cloud correction coordinates of the matching error feature points are obtained, and a new three-dimensional point cloud model is constructed.

[0009] Furthermore, the method for obtaining the matching feature points includes:

[0010] The SIFT optical matching algorithm is used to obtain matching feature points between adjacent optical slice images.

[0011] Furthermore, the method for obtaining the probability of local deformation includes:

[0012] The feature vector of each pixel is obtained, and the difference between the feature vector of each matching feature point and the feature vector of each other pixel in the neighborhood range on each frame of optical slice image is obtained as the difference vector of each other pixel.

[0013] The difference between the difference vectors of different pixels in the neighborhood of each matching feature point between adjacent optical slice images is obtained. The mean of the difference vectors is calculated after taking the modulus of the difference vectors and then normalized. This mean is used as the local deformation probability of each matching feature point on each optical slice image.

[0014] Furthermore, the method for obtaining the matching error feature points includes:

[0015] If the local deformation probability of a matching feature point is greater than or equal to a preset probability threshold, the corresponding matching feature point is taken as a matching error feature point, and the region formed by consecutive matching error feature points is taken as the deformation target region.

[0016] Furthermore, the method for obtaining the deformation center of gravity includes:

[0017] The first accumulated value of the product between the local deformation probability of different matching feature points in each deformation target region and the coordinate value of each direction in the corresponding 3D point cloud model is obtained, and the second accumulated value of the local deformation probability of different matching feature points is obtained.

[0018] The ratio of the first accumulated value and the second accumulated value is obtained as the deformation barycenter coordinate in each direction; the deformation barycenter coordinates in all spatial directions are obtained to form the deformation barycenter.

[0019] Furthermore, the method for obtaining the consistency of the matching error includes:

[0020] Based on the difference distribution of the centroid of deformation within the deformation target region between adjacent optical slice images, the maximum deformation displacement of the deformation target region is obtained.

[0021] The modulus of the maximum displacement of the deformation is obtained and the ratio is taken as the moving path corresponding to the deformation target area on each frame of optical slice image. This ratio is used as the consistency of the matching error of the deformation target area on each frame of optical slice image.

[0022] Furthermore, the method for obtaining the maximum displacement of the deformation includes:

[0023] The coordinate difference between the centroid of deformation between each frame and the next frame of the optical slice image is obtained, which is used as the deformation displacement of the centroid of deformation on each frame of the optical slice image.

[0024] The cumulative value of the deformation displacement ratio between each frame and the optical slice images of different frames before and after is obtained. The deformation displacement of the frame with the largest cumulative value is selected as the maximum deformation displacement.

[0025] Furthermore, the method for obtaining the point cloud corrected coordinates includes:

[0026] Based on the consistency distribution of matching errors in the deformed target region on different frames of optical slice images, the influence of the tongue depressor in the deformed target region is obtained.

[0027] The relative distance between the point cloud coordinates and the deformation centroid is obtained and negatively correlated, which is used as the first correction coefficient. The product of the first correction coefficient and the influence of the tongue depressor is used as the correction weight. The point cloud coordinates are weighted based on the correction weight to obtain the corrected point cloud coordinates.

[0028] Furthermore, the method for obtaining the influence of the tongue depressor includes:

[0029] The difference in matching error consistency between the next frame and each frame of optical slice image is obtained as the consistency difference between the corresponding frame images.

[0030] The mean ratio of consistency differences between different frame images is obtained, the difference between the mean ratio result and the positive integer 1 is obtained, and a negative correlation mapping is performed as the influence of the tongue depressor.

[0031] This invention also proposes a high-precision three-dimensional oral cavity imaging method, the method comprising:

[0032] Obtain optical slice images of different frames within the oral cavity along the movement path of the tongue depressor;

[0033] Based on the structural features of different frames of optical slice images, the matching feature points that constitute the 3D point cloud model and the feature vectors of the matching feature points are obtained; based on the position and feature vector distribution of the matching feature points between different frames of optical slice images, the local deformation probability of each matching feature point in each frame of optical slice image is obtained, and the matching error feature points that constitute the deformation target area are selected.

[0034] For any frame of optical slice image, the deformation centroid of each deformation target region is obtained based on the local deformation probability of different matching feature points within the deformation target region and their positions in the 3D point cloud model; the movement path when the depressor intersects with the deformation target region is obtained; and the matching error consistency of the deformation target region on each frame of optical slice image is obtained based on the difference distribution of the deformation centroids within the deformation target region between adjacent frames of optical slice images and the changing trend of the corresponding movement path.

[0035] Based on the consistency distribution of matching error in the deformed target region on different frames of optical slice images, as well as the point cloud coordinates of matching error feature points and the position distribution of the deformation centroid within the deformed target region, the point cloud correction coordinates of the matching error feature points are obtained, and a new three-dimensional point cloud model is constructed.

[0036] The present invention has the following beneficial effects:

[0037] This invention obtains matching feature points constituting a three-dimensional point cloud model based on the structural features of different frames of optical slice images, which helps in describing structural features; based on the position and local structural features of the matching feature points between different frames of optical slice images, it obtains the local deformation probability of each matching feature point in each frame of optical slice images, and filters out the matching error feature points constituting the deformation target region; for any frame of optical slice image, based on the local deformation probability of different matching feature points in the deformation target region and their position in the three-dimensional point cloud model, it obtains the deformation centroid of each deformation target region, which is a reference point reflecting the overall deformation state of the region; considering the changes in the physical motion of the tongue depressor during its slow movement in the oral cavity, the contact... Changes in the soft tissue structure cause localized elastic deformation in the target area. The movement path of the tongue depressor when it intersects with the deformed target area is obtained. Based on the difference in the distribution of the deformation centroid within the target area between adjacent optical slice images, and the changing trend of the corresponding movement path, the consistency of the matching error of the deformed target area on each frame of the optical slice image is obtained. Based on the distribution of the consistency of the matching error of the deformed target area on different frames of optical slice images, and the point cloud coordinates of the matching error feature points and the positional distribution of the deformation centroid within the deformed target area, the point cloud correction coordinates of the matching error feature points are obtained. This eliminates most of the systematic errors caused by soft tissue deformation, more realistically reflects the structure of the oral cavity, and constructs a new three-dimensional point cloud model. This invention accurately obtains the point cloud correction coordinates of the deformed area in each image, improving the accuracy of the three-dimensional model construction of the oral cavity. Attached Figure Description

[0038] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This is a flowchart of an intelligent electronic tongue depressor and a high-precision three-dimensional oral imaging method provided in one embodiment of the present invention. Detailed Implementation

[0040] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an intelligent electronic tongue depressor and a high-precision three-dimensional oral imaging method proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0042] The following description, in conjunction with the accompanying drawings, details the specific scheme of the intelligent electronic tongue depressor and the high-precision three-dimensional oral imaging method provided by the present invention.

[0043] This invention proposes an intelligent electronic tongue depressor, comprising a tongue depressor body, specifically a groove at one end of the handle connected to the tongue depressor, with a light source for illumination and a miniature depth camera embedded above the groove to form an image acquisition device; it also includes a controller, and the image acquisition device and a position sensor connected to the controller via signals. The image acquisition device, consisting of the light source for illumination and the miniature depth camera embedded above the groove in the handle end connected to the tongue depressor, is used to acquire optical slice images of the oral cavity. The position sensor is used to acquire the movement path of the tongue depressor. The control method of the controller is described in [reference needed]. Figure 1 The diagram illustrates a flowchart of a control method provided in one embodiment of the present invention, the specific method including:

[0044] Step S1: Obtain optical slice images of different frames inside the oral cavity along the movement path of the tongue depressor.

[0045] In an embodiment of the present invention, firstly, the miniature depth camera is a visual device with a 55° field of view, capable of acquiring clear oral optical slices within a working distance of 2-10mm. The electronic tongue depressor moves smoothly along different paths within the oral cavity, ensuring that the field of view overlap of adjacent frames is greater than or equal to 30%, thereby acquiring optical slice images of different frames within the oral cavity along the movement path of the tongue depressor.

[0046] Step S2: Based on the structural features of different frames of optical slice images, obtain the matching feature points that constitute the 3D point cloud model; based on the position and local structural features of the matching feature points between different frames of optical slice images, obtain the local deformation probability of each matching feature point in each frame of optical slice image, and filter out the matching error feature points that constitute the deformation target area.

[0047] The internal background of the oral cavity is complex, and the acquired oral cavity structural features are relatively one-sided. Therefore, it is necessary to perform feature matching on consecutive frames of optical slice images to obtain matching feature points that constitute a three-dimensional model based on the structural features of different frames of optical slice images.

[0048] Preferably, in one embodiment of the present invention, the method for obtaining matching feature points includes:

[0049] The SIFT optical matching algorithm is used to obtain matching feature points between adjacent optical slice images.

[0050] It should be noted that the SIFT algorithm finds points with similar features from different perspectives using techniques well-known to those skilled in the art, which will not be elaborated here.

[0051] It should be noted that, considering the complex background inside the oral cavity and the large amount of information contained in the optical slice image, in order to reduce the computational load, threshold segmentation can be used to segment the optical slice before image matching to avoid interference from irrelevant features.

[0052] Because the tissue structure within the patient's oral cavity is consistent, the vectors pointing to each point in the neighborhood from the same feature point on adjacent optical slices are more similar during the matching process. However, considering the compression of the oral soft tissue by the tongue depressor, which causes elastic deformation in the structure of optical slices at different locations, local analysis is performed on the feature vectors of the matching feature points in different frames of images to quantify the possibility of local deformation. Based on the position and local structural features of the matching feature points between different frames of optical slice images, the possibility of local deformation for each matching feature point in each frame of optical slice image is obtained.

[0053] Preferably, in one embodiment of the present invention, the method for obtaining the probability of local deformation includes:

[0054] The feature vector of each pixel is obtained, and the difference between the feature vector of each matching feature point and the feature vector of each other pixel in the neighborhood range on each frame of optical slice image is obtained as the difference vector of each other pixel.

[0055] It should be noted that, in one embodiment of the present invention, the feature vector can represent the local structural features of the image, and the feature vector of each pixel is generated based on the SIFT algorithm. In other embodiments of the present invention, the SURF algorithm can also be used to obtain the feature vector. The specific means are well known to those skilled in the art and will not be described in detail here.

[0056] The difference between the difference vectors of different pixels in the neighborhood of each matching feature point between adjacent optical slice images is obtained. The mean of the difference vectors is calculated after taking the modulus of the difference vectors and then normalized. This mean is used as the local deformation probability of each matching feature point on each optical slice image.

[0057] It should be noted that, in the embodiments of the present invention, normalization is performed by linear normalization or a normalization function. The specific means are well known to those skilled in the art and will not be described in detail here.

[0058] It should be noted that, in one embodiment of the present invention, the method for obtaining the neighborhood range is to obtain the range of pixels within a spherical sliding window with a radius of 5, centered on each matching feature point. In other embodiments of the present invention, the size of the neighborhood range can be set according to specific circumstances, and will not be limited or elaborated here.

[0059] The greater the likelihood of local deformation, the more pronounced the squeezing effect of the tongue depressor on the oral soft tissue, and the more likely it is to affect the matching results, making it a more likely match error feature point; thus, the match error feature points constituting the deformation target area are selected.

[0060] Preferably, in one embodiment of the present invention, the method for obtaining matching error feature points includes:

[0061] If the local deformation probability of a matching feature point is greater than or equal to a preset probability threshold, the corresponding matching feature point is taken as a matching error feature point, and the region formed by consecutive matching error feature points is taken as the deformation target region.

[0062] It should be noted that, in one embodiment of the present invention, the preset probability threshold is 0.5. In other embodiments of the present invention, the preset probability threshold can be set according to specific circumstances, and will not be limited or elaborated here.

[0063] Step S3: For any frame of optical slice image, obtain the deformation centroid of each deformation target region based on the local deformation probability of different matching feature points in each deformation target region and the different directional coordinates in the 3D point cloud model; obtain the movement path when the depressor plate intersects with the deformation target region; and obtain the matching error consistency of the deformation target region on each frame of optical slice image based on the difference distribution of deformation centroids in the deformation target region between adjacent frames of optical slice images and the changing trend of the corresponding movement path.

[0064] The deformable target area exhibits inconsistent matching or prominent discontinuities among multiple optical slices within a local area. The greater the possibility of local deformation, the more it is affected by deformation interference from the tongue depressor operation, and the more necessary it is to analyze the deformation centroid. Based on the local deformation probability of different matching feature points within each deformable target area, as well as the coordinates of different directions, the deformation centroid of each deformable target area is obtained.

[0065] Preferably, in one embodiment of the present invention, the method for obtaining the deformation center of gravity includes:

[0066] The first accumulated value of the product between the local deformation probability of different matching feature points in each deformation target region and the coordinate value of each direction in the corresponding 3D point cloud model is obtained, and the second accumulated value of the local deformation probability of different matching feature points is obtained.

[0067] The ratio of the first accumulated value and the second accumulated value is obtained as the deformation barycenter coordinate in each direction; the deformation barycenter coordinates in all spatial directions are obtained to form the deformation barycenter.

[0068] During the slow movement of the tongue depressor within the oral cavity, the changes in its physical motion cause changes in the soft tissue structure at the contact point, resulting in local elastic deformation. As the tongue depressor moves away, the deformed target area changes from being compressed to relaxed, and the deformation recovers. There is an influence between the movement of the tongue depressor and the change in the deformation center of gravity. Therefore, the movement path of the tongue depressor when it intersects with the deformed target area is obtained. Based on the difference distribution of the deformation center of gravity in the deformed target area between adjacent frames of optical slice images, and the changing trend of the corresponding movement path, the consistency of the matching error of the deformed target area on each frame of optical slice image is obtained.

[0069] Preferably, in one embodiment of the present invention, the method for obtaining the consistency of matching errors includes:

[0070] Based on the difference distribution of the centroid of deformation within the deformation target region between adjacent optical slice images, the maximum deformation displacement of the deformation target region is obtained.

[0071] Preferably, in one embodiment of the present invention, the method for obtaining the maximum deformation displacement includes:

[0072] The coordinate difference between the centroid of deformation between each frame and the next frame of the optical slice image is obtained, which is used as the deformation displacement of the centroid of deformation on each frame of the optical slice image.

[0073] The cumulative value of the deformation displacement ratio between each frame and the optical slice images of different frames before and after is obtained. The deformation displacement of the frame with the largest cumulative value is selected as the maximum deformation displacement.

[0074] The modulus of the maximum displacement of the deformation is obtained and the ratio is taken as the moving path corresponding to the deformation target area on each frame of optical slice image. This ratio is used as the consistency of the matching error of the deformation target area on each frame of optical slice image.

[0075] Step S4: Based on the consistency distribution of matching error in the deformed target region on different frames of optical slice images, and the point cloud coordinates of matching error feature points and the position distribution of the deformation centroid within the deformed target region, obtain the point cloud correction coordinates of the matching error feature points and construct a new three-dimensional point cloud model.

[0076] The higher the consistency of the matching error, the clearer the deformation of the target area and the movement of the tongue depressor. The point cloud coordinates reflect the geometric structure of the feature points, and the deformation centroid represents the position of the deformation area in the overall space. By analyzing the point cloud coordinates and the position distribution of the deformation centroid of the matching error feature points, it is helpful to show the distance of the feature points from the deformation area. The closer the position is, the closer it is to the tongue depressor, and the more correction is needed. Therefore, based on the distribution of the consistency of the matching error of the target area in different frames of optical slice images, and the distribution of the point cloud coordinates and the position distribution of the deformation centroid of the matching error feature points in the target area, the point cloud correction coordinates of the matching error feature points can be obtained.

[0077] Preferably, in one embodiment of the present invention, the method for obtaining point cloud correction coordinates includes:

[0078] Based on the consistency distribution of matching errors in the deformed target region on different frames of optical slice images, the influence of the tongue depressor in the deformed target region is obtained.

[0079] The consistency of the matching error reflects the correlation between the deformation changes in the target area and the movement path of the tongue depressor; preferably, in one embodiment of the present invention, the method for obtaining the influence of the tongue depressor includes:

[0080] The difference in matching error consistency between the next frame and each frame of optical slice image is obtained as the consistency difference between the corresponding frame images.

[0081] The mean ratio of consistency differences between different frame images is obtained, the difference between the mean ratio result and the positive integer 1 is obtained, and a negative correlation mapping is performed as the influence of the tongue depressor.

[0082] Based on this, the closer the mean ratio is to 1, the smaller the difference, the greater the possibility that the matching error will undergo elastic deformation due to the action of the tongue depressor, and the greater the influence of the tongue depressor; the closer the mean ratio is to 1, the smaller the difference, the smaller the possibility that the matching error will undergo elastic deformation due to the action of the tongue depressor, and the smaller the influence of the tongue depressor.

[0083] It should be noted that, in the embodiments of the present invention, the method involves finding the reciprocal or an exponential function with the natural constant as its base. When performing negative correlation mapping, in order to avoid the formula being meaningless with a denominator of 0 when calculating the reciprocal, a threshold such as 0.01 needs to be artificially added to the denominator. The specific method is a well-known technique to those skilled in the art and will not be elaborated here.

[0084] Obtain the relative distance between the point cloud coordinates and the deformation centroid, and perform negative correlation mapping as the first correction coefficient; obtain the product of the first correction coefficient and the influence of the tongue depressor as the correction weight;

[0085] The point cloud coordinates are weighted based on the corrected weights to obtain the corrected point cloud coordinates.

[0086] It should be noted that, in the embodiments of the present invention, the relative distance is obtained by Euclidean distance and Manhattan distance, etc.; the specific means are well known to those skilled in the art and will not be described in detail here.

[0087] Based on this, 3D reconstruction technology is used to reconstruct each frame of optical slice image after the point cloud coordinates of the matching error feature points in the deformed target area are corrected, generating a smooth and detailed three-dimensional mesh surface. Multiple local optical slice images from different parts of the oral cavity are then fused into a unified three-dimensional model of the oral cavity, displaying a clear structure of the patient's oral cavity.

[0088] In summary, this invention obtains the local deformation probability of each matching feature point in each frame of optical slice images based on the structural characteristics of different frames, and filters out the matching error feature points that constitute the deformation target region. For any frame of optical slice images, based on the local deformation probability of different matching feature points in the deformation target region, their position in the 3D point cloud model, and the changing trend of the corresponding movement path when the tongue depressor intersects with the deformation target region, the consistency of the matching error in the deformation target region on each frame of optical slice images is obtained. Combining the point cloud coordinates of the matching error feature points in the deformation target region and the position distribution of the deformation centroid, the point cloud correction coordinates of the matching error feature points are obtained. This invention accurately obtains the point cloud correction coordinates of the deformation region of each image, improving the accuracy of the oral cavity 3D model construction.

[0089] This invention also proposes a high-precision three-dimensional oral cavity imaging method, the method comprising:

[0090] Obtain optical slice images of different frames within the oral cavity along the movement path of the tongue depressor;

[0091] Based on the structural features of different frames of optical slice images, matching feature points that constitute the 3D point cloud model are obtained; based on the position and local structural features of the matching feature points between different frames of optical slice images, the local deformation probability of each matching feature point in each frame of optical slice image is obtained, and the matching error feature points that constitute the deformation target area are screened out.

[0092] For any frame of optical slice image, the deformation centroid of each deformation target region is obtained based on the local deformation probability of different matching feature points within the deformation target region and their positions in the 3D point cloud model; the movement path when the depressor intersects with the deformation target region is obtained; and the matching error consistency of the deformation target region on each frame of optical slice image is obtained based on the difference distribution of the deformation centroids within the deformation target region between adjacent frames of optical slice images and the changing trend of the corresponding movement path.

[0093] Based on the consistency distribution of matching error in the deformed target region on different frames of optical slice images, as well as the point cloud coordinates of matching error feature points and the position distribution of the deformation centroid within the deformed target region, the point cloud correction coordinates of the matching error feature points are obtained, and a new three-dimensional point cloud model is constructed.

[0094] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0095] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. An intelligent electronic tongue depressor comprising a tongue depressor body, characterized in that, It also includes a controller, an image acquisition unit and a position sensor connected to the controller via signals. The image acquisition unit is used to acquire optical slice images of the oral cavity, and the position sensor is used to acquire the movement path of the tongue depressor. The control method of the controller includes: Obtain optical slice images of different frames within the oral cavity along the movement path of the tongue depressor; Based on the structural features of different frames of optical slice images, matching feature points that constitute the 3D point cloud model are obtained; based on the position and local structural features of the matching feature points between different frames of optical slice images, the local deformation probability of each matching feature point in each frame of optical slice image is obtained, and the matching error feature points that constitute the deformation target area are screened out. For any frame of optical slice image, the deformation centroid of each deformation target region is obtained based on the local deformation probability of different matching feature points within the deformation target region and their positions in the 3D point cloud model; the movement path when the depressor intersects with the deformation target region is obtained; and the matching error consistency of the deformation target region on each frame of optical slice image is obtained based on the difference distribution of the deformation centroids within the deformation target region between adjacent frames of optical slice images and the changing trend of the corresponding movement path. Based on the consistency distribution of matching error in the deformed target region on different frames of optical slice images, and the point cloud coordinates of matching error feature points and the position distribution of the deformation centroid in the deformed target region, the point cloud correction coordinates of the matching error feature points are obtained, and a new three-dimensional point cloud model is constructed. The method for obtaining the point cloud corrected coordinates includes: obtaining the tongue depressor influence of the deformed target region based on the consistency distribution of matching error of the deformed target region on different frames of optical slice images; obtaining the relative distance between the point cloud coordinates and the deformation centroid and performing negative correlation mapping as the first correction coefficient; obtaining the product of the first correction coefficient and the tongue depressor influence as the correction weight; and weighting the point cloud coordinates based on the correction weight to obtain the point cloud corrected coordinates.

2. The intelligent electronic tongue depressor according to claim 1, wherein, The method for obtaining the matching feature points includes: The SIFT optical matching algorithm is used to obtain matching feature points between adjacent optical slice images.

3. The intelligent electronic tongue depressor according to claim 1, wherein, The method for obtaining the probability of local deformation includes: The feature vector of each pixel is obtained, and the difference between the feature vector of each matching feature point and the feature vector of each other pixel in the neighborhood range on each frame of optical slice image is obtained as the difference vector of each other pixel. The difference between the difference vectors of different pixels in the neighborhood of each matching feature point between adjacent optical slice images is obtained. The mean of the difference vectors is calculated after taking the modulus of the difference vectors and then normalized. This mean is used as the local deformation probability of each matching feature point on each optical slice image.

4. The intelligent electronic tongue depressor according to claim 3, characterized in that, The method for obtaining the matching error feature points includes: If the local deformation probability of a matching feature point is greater than or equal to a preset probability threshold, the corresponding matching feature point is taken as a matching error feature point, and the region formed by consecutive matching error feature points is taken as the deformation target region.

5. The intelligent electronic tongue depressor according to claim 1, wherein, The method for obtaining the deformation centroid includes: The first accumulated value of the product between the local deformation probability of different matching feature points in each deformation target region and the coordinate value of each direction in the corresponding 3D point cloud model is obtained, and the second accumulated value of the local deformation probability of different matching feature points is obtained. The ratio of the first accumulated value and the second accumulated value is obtained as the deformation barycenter coordinate in each direction; the deformation barycenter coordinates in all spatial directions are obtained to form the deformation barycenter.

6. The intelligent electronic tongue depressor according to claim 1, wherein, The method for obtaining the consistency of the matching error includes: Based on the difference distribution of the centroid of deformation within the deformation target region between adjacent optical slice images, the maximum deformation displacement of the deformation target region is obtained. The modulus of the maximum displacement of the deformation is obtained and the ratio is taken as the movement path corresponding to the deformation target area on each frame of optical slice image. This ratio is used as the consistency of the matching error of the deformation target area on each frame of optical slice image.

7. The intelligent electronic tongue depressor according to claim 6, wherein, The method for obtaining the maximum displacement of the deformation includes: The coordinate difference between the centroid of deformation between each frame and the next frame of the optical slice image is obtained, which is used as the deformation displacement of the centroid of deformation on each frame of the optical slice image. The cumulative value of the deformation displacement ratio between each frame and the optical slice images of different frames before and after is obtained. The deformation displacement of the frame with the largest cumulative value is selected as the maximum deformation displacement.

8. The intelligent electronic tongue depressor according to claim 1, wherein, The method for obtaining the influence of the tongue depressor includes: The difference in matching error consistency between the next frame and each frame of optical slice image is obtained as the consistency difference between the corresponding frame images. The mean ratio of consistency differences between different frame images is obtained, the difference between the mean ratio result and the positive integer 1 is obtained, and a negative correlation mapping is performed as the influence of the tongue depressor.

9. A high-precision three-dimensional oral imaging method, characterized by, The method includes: Obtain optical slice images of different frames within the oral cavity along the movement path of the tongue depressor; Based on the structural features of different frames of optical slice images, the matching feature points that constitute the 3D point cloud model and the feature vectors of the matching feature points are obtained; based on the position and feature vector distribution of the matching feature points between different frames of optical slice images, the local deformation probability of each matching feature point in each frame of optical slice image is obtained, and the matching error feature points that constitute the deformation target area are selected. For any frame of optical slice image, the deformation centroid of each deformation target region is obtained based on the local deformation probability of different matching feature points within the deformation target region and their positions in the 3D point cloud model; the movement path when the depressor intersects with the deformation target region is obtained; and the matching error consistency of the deformation target region on each frame of optical slice image is obtained based on the difference distribution of the deformation centroids within the deformation target region between adjacent frames of optical slice images and the changing trend of the corresponding movement path. Based on the consistency distribution of matching error in the deformed target region on different frames of optical slice images, and the point cloud coordinates of matching error feature points and the position distribution of the centroid of deformation in the deformed target region, the point cloud correction coordinates of the matching error feature points are obtained, and a new three-dimensional point cloud model is constructed. The method for obtaining the point cloud corrected coordinates includes: obtaining the tongue depressor influence of the deformed target region based on the consistency distribution of matching error of the deformed target region on different frames of optical slice images; obtaining the relative distance between the point cloud coordinates and the deformation centroid and performing negative correlation mapping as the first correction coefficient; obtaining the product of the first correction coefficient and the tongue depressor influence as the correction weight; and weighting the point cloud coordinates based on the correction weight to obtain the point cloud corrected coordinates.

Images