Tooth 3D modeling method for dental veneer restoration

In the repair of tooth veneer, the division method of grayscale change rate and reference point is used to adaptively select the best seed point for regional growth segmentation, which solves the problem of inaccurate seed point selection in traditional methods and improves the 3D teeth The accuracy and effectiveness of modeling.

CN118628668BActive Publication Date: 2024-11-22SHIMMER COMPUTERIZED DENTAL TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411095456.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-12
Publication Date
2024-11-22
Estimated Expiration
2044-08-12

AI Technical Summary

Technical Problem

During the repair of tooth veneer, traditional regional growth methods require artificial selection of seed points, resulting in inaccurate image segmentation effect. Especially when the grayscale value distribution of defective teeth surface is uneven, pixel points in some areas are prone to occur, which affects the impact of Accuracy of 3D modeling.

Method used

By obtaining the grayscale value of each pixel point in the grayscale image of the tooth surface, selecting the pixel point of the target grayscale value as the target point, calculate the grayscale change rate with the neighboring pixel points, obtain the reference point, and divide the grayscale change The rate range is divided into sets, and the best seed points are selected for regional growth segmentation to ensure that areas with large grayscale values ​​are not missed and the segmentation effect is improved.

Benefits of technology

It realizes more accurate segmentation of areas to be repaired in dental veneer repair, improves the quality of teeth 3D modeling, ensures that areas with large grayscale values ​​are completely detected, and reduces the impact of human intervention.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118628668B_ABST
    Figure CN118628668B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of image data processing, and in particular to a tooth 3D modeling method for tooth veneer restoration, wherein the method obtains a surface grayscale image of a defective tooth and all reference points in the surface grayscale image; divides all reference points in the surface grayscale image into at least two sets according to the grayscale change rate of all reference points; selects an optimal seed point from all reference points in the surface grayscale image according to the grayscale mean corresponding to each set, performs regional growth segmentation on the surface grayscale image based on the optimal seed point to obtain a region to be restored, performs 3D modeling of the defective tooth based on the restored region, and improves the quality of tooth 3D modeling based on the region to be restored by determining the optimal seed point that is more in line with the scene.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a tooth 3D modeling method for tooth veneer restoration. Background Art

[0002] Because some people neglect to protect their teeth, or some people are born with genetic problems of partial tooth discoloration and structural disorder, these people have defective teeth with low aesthetics or loss of use function, which is mainly manifested in uneven coloring or defects on the tooth surface. Therefore, dental veneer restoration technology is needed to improve the appearance and function of defective teeth. Dental veneer restoration technology requires specific judgment on the repair position and condition of the defective teeth. The image of the defective tooth surface is processed and segmented to obtain the area to be repaired, and then a 3D model of the defective tooth is constructed according to the area to be repaired. A specific plan for the repair of the defective tooth is made based on the 3D model, and then a repaired tooth model is constructed to obtain a dental restoration to repair the defective tooth.

[0003] However, defective teeth have missing parts or colored areas, which leads to uneven distribution of grayscale values ​​in some areas and a large gap. As a result, when obtaining the area to be repaired in the surface image of the defective tooth, it is difficult to segment the relatively complete missing parts or colored areas using traditional single threshold image segmentation, which makes the defect display incomplete and affects the accuracy of 3D modeling of defective teeth.

[0004] In the prior art, the region growing method can achieve better results in image segmentation. However, the traditional region growing method requires manual selection of seed points, and the image segmentation effect will be different depending on the selection of the seed points. At the same time, the grayscale values ​​of different areas on the surface of defective teeth are different. If the region growing method is used to segment the image according to the manually selected seed points, some colored pixels may be lost, resulting in missing image segmentation areas, resulting in inaccurate final image segmentation effect.

[0005] Therefore, how to adaptively select the best seed point position when using the region growing method to segment the image of the defective tooth surface becomes an urgent problem to be solved. Summary of the invention

[0006] In view of this, an embodiment of the present invention provides a tooth 3D modeling method for tooth veneer restoration to solve the problem of how to adaptively select the optimal seed point position when performing region segmentation on an image of a defective tooth surface using a region growing method.

[0007] An embodiment of the present invention provides a tooth 3D modeling method for tooth veneer restoration, the method comprising the following steps:

[0008] During the dental veneer restoration process, a grayscale image of the defective tooth surface is obtained;

[0009] According to the gray value of each pixel in the surface gray image, a pixel with a target gray value is selected as a target point. For any target point, the target point is used as a starting point to obtain a gray change rate between the target point and each of its neighboring pixel points. According to the gray change rate, a reference point of the target point is obtained. Any neighboring pixel of the target point is used as the target point. The method for obtaining the reference point of the target point is repeated until no reference point can be obtained, and all reference points with the target point as the starting point are obtained. According to all reference points obtained with each of the target points as the starting point, all reference points in the surface gray image are obtained.

[0010] Obtaining a maximum grayscale change rate and a minimum grayscale change rate from the grayscale change rates corresponding to all reference points in the surface grayscale image to obtain a grayscale change rate interval, dividing the grayscale change rate interval into at least two sub-intervals according to a preset interval interval, and dividing all reference points in the surface grayscale image into at least two sets according to each of the sub-intervals and the grayscale change rate corresponding to each reference point in the surface grayscale image;

[0011] The grayscale means of all the pixels included in each of the sets are obtained respectively, and according to the grayscale means corresponding to each of the sets, the best seed point is screened out from all the reference points in the surface grayscale image, and the surface grayscale image is segmented by regional growth based on the best seed point to obtain the area to be repaired, and the 3D modeling of the defective tooth is performed according to the area to be repaired.

[0012] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0013] In the process of tooth veneer restoration, the present invention obtains a surface grayscale image of a defective tooth; according to the grayscale value of each pixel point in the surface grayscale image, a pixel point with a target grayscale value is selected as a target point; for any target point, the target point is used as a starting point to obtain a grayscale change rate between the target point and each of its neighboring pixel points; according to the grayscale change rate, a reference point of the target point is obtained; any neighboring pixel point of the target point is used as the target point, and the method for obtaining the reference point of the target point is repeated until no reference point can be obtained, and all reference points with the target point as the starting point are obtained; according to all reference points obtained with each of the target points as the starting point, all reference points in the surface grayscale image are obtained; in all the reference points in the surface grayscale image The maximum grayscale change rate and the minimum grayscale change rate are obtained from the grayscale change rates corresponding to the reference points to obtain a grayscale change rate interval, and the grayscale change rate interval is divided into at least two sub-intervals according to a preset interval interval, and all reference points in the surface grayscale image are divided into at least two sets according to the grayscale change rates corresponding to each sub-interval and each reference point in the surface grayscale image; the grayscale mean values ​​of all pixel points contained in each of the sets are respectively obtained, and according to the grayscale mean values ​​corresponding to each of the sets, the best seed point is screened from all reference points in the surface grayscale image, and the surface grayscale image is segmented by regional growth based on the best seed point to obtain the area to be repaired, and 3D modeling of the defective tooth is performed according to the area to be repaired. Among them, the pixel point that meets the target grayscale value is obtained as the target point, firstly, the grayscale change rate of the neighboring pixel points and the target point is calculated with the target point as the center, and then the grayscale change rate of the neighboring pixel points with the surrounding pixel points is calculated with the neighboring pixel point as the new center, and the pixel point with the same grayscale change rate is found as the reference point, which is used to screen the pixel points with similar grayscale change rate that are suspected to belong to the tooth defect, and then all the reference points in the surface grayscale image are divided into multiple sets according to the maximum grayscale change rate and the minimum grayscale change rate among all the reference points, so that the reference points with the same grayscale change rate are regarded as a set. In order to ensure that the reference points with too large grayscale value difference are not omitted, the best seed point is screened from all the reference points in the surface grayscale image according to the grayscale mean value of each set, so that the obtained best seed point is more in line with the scene, so that the surface grayscale image is regionally grown and segmented based on the best seed point, and a more accurate area to be repaired is obtained, that is, the area where there is a defect on the tooth surface, further improving the quality of tooth 3D modeling according to the area to be repaired. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0015] Figure 1 is a method flow chart of a tooth 3D modeling method for tooth veneer restoration provided in the first embodiment of the present invention;

[0016] Figure 2 is a schematic diagram of a local pixel point provided by the first embodiment of the present invention;

[0017] Figure 3 is a schematic diagram of all initial reference points in a surface grayscale image provided by the first embodiment of the present invention;

[0018] Figure 4 is a schematic diagram of grayscale change rates of all reference points in a surface grayscale image provided by Embodiment 1 of the present invention;

[0019] Figure 5 This is a method provided by Embodiment 1 of the present invention. Figure 4 Schematic diagram of the locations of all reference points. DETAILED DESCRIPTION

[0020] Embodiments of the present disclosure are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, but should not be understood as limiting the present disclosure.

[0021] It should be noted that the terms "first", "second", etc. in the specification of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure.

[0022] In order to illustrate the technical solution of the present invention, specific embodiments are provided below for illustration.

[0023] See also Figure 1 , is a method flow chart of a method for 3D modeling of teeth for tooth veneer restoration provided in the first embodiment of the present invention, such as Figure 1 As shown, the method may include:

[0024] Step S101, during the tooth veneer restoration process, a grayscale image of the surface of the defective tooth is obtained.

[0025] The surface images of all teeth in the oral cavity are obtained by a 3D scanner, and the surface image of each defective tooth is screened out. Each screened surface image of the defective tooth is grayscaled separately to obtain the surface grayscale image of each defective tooth, wherein the grayscale processing belongs to the existing technology and will not be described in detail here.

[0026] Step S102, according to the grayscale value of each pixel in the surface grayscale image, select the pixel of the target grayscale value as the target point, for any target point, take the target point as the starting point, obtain the grayscale change rate between the target point and each of its neighboring pixel points, obtain the reference point of the target point according to the grayscale change rate, take any neighboring pixel point of the target point as the target point, repeat the method of obtaining the reference point of the target point until the reference point cannot be obtained, obtain all reference points with the target point as the starting point, and obtain all reference points in the surface grayscale image according to all the reference points obtained with each target point as the starting point.

[0027] Since the grayscale values ​​of different areas on the surface of defective teeth are unevenly distributed, it is necessary to use the region growing method to segment the surface grayscale image to obtain the area to be repaired corresponding to the degree of defect or coloration on the tooth surface. However, when using the traditional region growing method to segment the surface grayscale image, it is necessary to manually select seed points. If the region growing segmentation is performed according to the manually selected seed points, it will cause the loss of pixels in some areas, resulting in poor segmentation effect of the surface grayscale image. Therefore, the embodiment of the present invention takes into account that different seed points will have different segmentation results, and it is necessary to adaptively find the best seed point for region growing to obtain the best image segmentation effect.

[0028] Due to the different degrees of tooth surface defects or staining, the grayscale values ​​are also different, and the differences between some areas are large. The different positions of seed points will lead to different results. If seed points are only selected in areas with larger grayscale values, areas with smaller grayscale values ​​will not be detected, forming faults in defective areas. If seed points are selected in areas with smaller grayscale values, areas with larger grayscale values ​​will be missing. Therefore, a single grayscale value cannot be used to determine the seed points, and the possibility that seed points in different areas are different needs to be considered. Because the degree of defect or staining is gradual, there are similar or identical grayscale change rates between the pixels in the defect or staining area. These pixels with similar or identical grayscale change rates are the reference points to be found, which are used to find the best seed points among the reference points. Therefore, the specific steps to obtain all reference points in the surface grayscale image are as follows:

[0029] (1) According to the grayscale value of each pixel in the surface grayscale image, the pixel with the target grayscale value is selected as the target point.

[0030] Specifically, since the present invention is to segment the area to be repaired in the surface grayscale image, that is, the area where the tooth surface is defective or colored, the threshold segmentation method is used to obtain the obvious defect area in the surface grayscale image, calculate the grayscale mean in the obvious defect area, and round the grayscale mean as the target grayscale value, and then take the pixel point corresponding to the target grayscale value in the surface grayscale image as the target point. Assuming that the target grayscale value is 15, the pixel point with a grayscale value of 15 in the surface grayscale image is taken as the target point. At this point, all the target points in the surface grayscale image can be obtained.

[0031] (2) For any target point, take the target point as the starting point, obtain the grayscale change rate between the target point and each of its neighboring pixel points, obtain the reference point of the target point based on the grayscale change rate, take any neighboring pixel point of the target point as the target point, repeat the method of obtaining the reference point of the target point until no reference point can be obtained, and obtain all reference points with the target point as the starting point.

[0032] Specifically, taking a target point as an example, with the target point as the starting point, first, obtain the neighboring pixels in the four neighborhoods of the target point, and obtain the grayscale change rate between the target point and each neighboring pixel according to the grayscale value difference between the target point and each neighboring pixel, which is used to characterize the grayscale change of the target point in the four neighborhood directions of up, down, left, and right. The calculation formula of the grayscale change rate is: | |*100%, where B is the grayscale value of the target point, A is the grayscale value of the neighboring pixel of the target point, and || represents the absolute value symbol.

[0033] Then, according to the grayscale change rate of the target point in the four neighborhoods, the reference point of the target point is obtained. The specific acquisition method is:

[0034] For the grayscale change rate between the target point and each of its neighboring pixel points, if there are at least two identical grayscale change rates, the neighboring pixel points corresponding to the identical grayscale change rates and the target point are used as reference points;

[0035] If there is no identical grayscale change rate, the difference between every two grayscale change rates is obtained, and the neighborhood pixel points corresponding to the difference in the preset difference interval and the target point are used as reference points.

[0036] In one embodiment, reference Figure 2 , which is a schematic diagram of a local pixel point. Assuming that the gray value of the target point is 15, then Figure 2 The grayscale values ​​of the four neighboring pixels of the target point are 16, 14, 14 and 15.5 respectively. The grayscale change rate calculation formula is as follows: The grayscale change rate between 15 and 16 is , the grayscale change rate between 15 and 14 is , the grayscale change rate between 15 and 15.5 is , the grayscale change rate between 15 and 14 is , it can be seen that there are 3 identical grayscale change rates of 7% and 1 grayscale change rate of 3%. The neighborhood pixels corresponding to the 3 identical grayscale change rates of 7% are used as reference points, that is, the neighborhood pixels with grayscale values ​​of 16, 14, and 14. At the same time, the target point is also used as a reference point.

[0037] If the grayscale change rates between the target point and each of the neighboring pixels are different, that is, the four grayscale change rates are different, then the difference between each two grayscale change rates is obtained, and the neighboring pixels and the target point corresponding to the difference in the preset difference interval are used as reference points. Assuming that the four grayscale change rates are 5%, 1%, 3% and 4% respectively, and the preset difference interval is [0%, 4%], the calculated difference between the grayscale change rates of 5% and 1% is 4%, which is in the preset difference interval, indicating that the neighboring pixels corresponding to the grayscale change rates of 5% and 1% belong to the reference points, and the target point is also used as a reference point. It should be noted that the preset difference interval is not limited here, and the implementer can adjust it according to the actual situation.

[0038] After obtaining the reference point centered on the target point, the neighborhood pixel point of the first reference point is taken as the center. Figure 2 The grayscale values ​​15.5, 16, 14, and 14 corresponding to the neighboring pixels with a grayscale value of 15 in the image are taken as the new center points to obtain the reference points. Repeat the above reference point acquisition steps. Taking 15.5 as an example, the grayscale change rate between 15.5 and each of its four neighboring pixels is obtained by using the grayscale change rate calculation formula: 0%, 6%, 3%, and the calculated 3% (that is, the grayscale change rate between 15 and 15.5 calculated when 15 is the center). At this time, two 3% can be obtained, and then the neighboring pixels corresponding to the two 3% and the center point are taken as reference points. Similarly, traverse outwards all the time. At this point, the calculated 4 grayscale change rates do not meet the reference point acquisition conditions (the same grayscale change rate or the difference between the two grayscale change rates is in the preset difference interval). At this time, the process of obtaining the reference point with the target point as the starting point ends, and the next target point can be replaced as the starting point to perform the above reference point acquisition process until all target points are traversed, and all reference points obtained with each target point as the starting point are obtained.

[0039] (3) Based on all reference points starting from each target point, all reference points in the surface grayscale image are obtained.

[0040] Specifically, for any target reference point, neighboring pixel points that are not the initial reference point are obtained from the neighboring pixel points of the target reference point as the pixel points to be analyzed, and a set of label values ​​of the grayscale change rate of each pixel point to be analyzed in each preset neighborhood direction are respectively obtained, wherein the label value refers to the value obtained according to the numerical attribute of the grayscale difference between two adjacent pixel points;

[0041] Based on the position distribution of all the initial reference points, obtaining a set of label values ​​formed by each of the initial reference points and other initial reference points in each of the preset neighborhood directions, counting the number of label value sets of each of the initial reference points, obtaining the initial reference points corresponding to the maximum number as pre-reference points, screening out reference points from all the pre-reference points, and using the label value sets of the reference points as the reference label value set;

[0042] The label value set of each pixel to be analyzed is compared with the benchmark label value set of the benchmark point to obtain supplementary reference points in the surface grayscale image. The supplementary reference points in the surface grayscale image and the initial reference points in the surface grayscale image constitute all reference points in the surface grayscale image.

[0043] In one embodiment, all reference points obtained with each target point as the starting point are used as the initial reference points in the surface grayscale image. After the initial reference points of the surface grayscale image are obtained through all the target points, the distribution of all the initial reference points in the surface grayscale image is used to obtain the edgemost initial reference point as the target reference point, that is, the initial reference point at which the grayscale change rate is finally calculated is used as the target reference point. Figure 3 , which is a schematic diagram of all initial reference points in a surface grayscale image. Since 13 in the first row of the third column, 18 in the last row of the third column, and 15 in the last column of the third row are the initial reference points for the final calculation of the grayscale change rate and are at the outermost position in the surface grayscale image, 13 in the first row of the third column, 18 in the last row of the third column, and 15 in the last column of the third row are all used as target reference points.

[0044] Taking a target reference point as an example, pixels that do not belong to the initial reference point in the four neighborhoods of the target reference point are obtained as pixels to be analyzed, and then for each pixel to be analyzed, a set of label values ​​of the grayscale change rate of each pixel to be analyzed in the four neighborhood directions is obtained. The four neighborhood directions are the four directions corresponding to the four neighborhoods. For any neighborhood direction, the grayscale value difference between each two adjacent pixels of the pixel to be analyzed in the neighborhood direction is calculated respectively. If the grayscale value difference is positive, the label value of the grayscale value difference is recorded as 1. If the grayscale value difference is negative, the label value of the grayscale value difference is recorded as -1. If the grayscale value difference is 0, the label value of the grayscale value difference is recorded as 0. Then, the label values ​​corresponding to each grayscale value difference of the pixel to be analyzed in the neighborhood direction are combined into a label value set. Similarly, the label value set of the pixel to be analyzed in each neighborhood direction can be obtained. Figure 3 Taking the target reference point 18 in as an example, the pixels centered at 18 that are not initial reference points in its four neighborhoods are obtained as the pixels to be analyzed, that is, the neighborhood pixels with grayscale values ​​of 15 and 9 are taken as the pixels to be analyzed. Taking the pixel 15 to be analyzed as an example, the grayscale value differences between 15 and each of its four neighborhood pixels are obtained respectively. Taking the upper neighborhood direction of 15 as an example, the grayscale value differences between every two adjacent pixels in the upper neighborhood direction are 14-15=-1, 13-14=-1, 9-13=1=-4, 10-9=1, and 2-10=-8 respectively. That is, the label value set of the pixel 15 to be analyzed in the upper neighborhood direction is {-1, -1, -1, 1, -1}. Similarly, all label value sets of the pixel to be analyzed in other neighborhood directions can be obtained.

[0045] Similarly, referring to the method for obtaining the set of label values ​​of the pixel to be analyzed in each neighborhood direction, for any initial reference point, the set of label values ​​of the initial reference point in each neighborhood direction is obtained according to the position distribution of the initial reference point and other initial reference points, thereby obtaining the set of label values ​​of each initial reference point in each neighborhood direction. It should be noted that when obtaining the set of label values ​​of the initial reference point, only the grayscale value difference between adjacent initial reference points is used. Afterwards, based on the set of label values ​​of each initial reference point, the initial reference point with the most changes in the neighborhood direction is obtained as the reference point. Specifically, first count the number of label value sets of each initial reference point, compare the number of label value sets of all initial reference points, and use the initial reference point corresponding to the largest number as the pre-reference point. For example: Assume Figure 3 The third column 13-18 and the third row 15 are the initial reference points. Figure 3 The initial reference point 15 in the third row and third column has three neighborhood direction changes, namely the upper neighborhood direction, the lower neighborhood direction, and the right neighborhood direction, corresponding to three tag value sets, and Figure 3There are only one or two neighborhood directions for other initial reference points in , corresponding to only one or two sets of marked values, so Figure 3 The initial reference point 15 in the third row and third column is Figure 3 After all pre-reference points are obtained, different screening methods are selected according to the number of pre-reference points in all pre-reference points to obtain reference points, so that the set of mark values ​​of the reference points in the four neighborhood directions is used as the reference mark value set.

[0046] Among them, the method of screening out the benchmark points from all the pre-benchmark points is:

[0047] Counting the number of the pre-reference points, and if the number of the pre-reference points is 1, taking the pre-reference point as a reference point;

[0048] If the number of the pre-reference points is not 1, for any pre-reference point, the number of elements in each label value set of the pre-reference point is counted respectively, and if the number of elements in all label value sets of the pre-reference point belongs to the maximum number of elements, the pre-reference point is taken as a candidate reference point;

[0049] The number of all candidate reference points is counted. If the number of all candidate reference points is 1, the candidate reference point is used as the reference point. If the number of all candidate reference points is not 1, any candidate reference point is selected from all candidate reference points as the reference point.

[0050] In one implementation, after obtaining the pre-reference points, the number of pre-reference points can be counted. If there is only one pre-reference point, the pre-reference point is directly used as the reference point. Otherwise, if the number of pre-reference points is greater than 1, the tag value set of each pre-reference point is analyzed. Specifically, taking any neighborhood direction as an example, the number of elements in the tag value set of each pre-reference point in the neighborhood direction is counted respectively, and the pre-reference point corresponding to the maximum number of elements in the neighborhood direction is obtained. Similarly, the pre-reference point corresponding to the maximum number of elements in each neighborhood direction is obtained, and then, based on the pre-reference point corresponding to the maximum number of elements in each neighborhood direction, the pre-reference points belonging to the maximum number of elements in all neighborhood directions are obtained as candidate reference points. For example, assuming a pre-reference point The number of elements in the marked value sets in the four neighborhood directions of the reference point are: 5 elements in the marked value set in the upper neighborhood direction, 4 elements in the marked value set in the lower neighborhood direction, 5 elements in the marked value set in the left neighborhood direction, and 6 elements in the marked value set in the right neighborhood direction. The maximum number of elements in the marked value set in the upper neighborhood direction is 5, the maximum number of elements in the marked value set in the lower neighborhood direction is 4, the maximum number of elements in the marked value set in the left neighborhood direction is 5, and the maximum number of elements in the marked value set in the right neighborhood direction is 6. It can be seen that the number of elements in the marked value set in each neighborhood direction of the pre-reference point belongs to the maximum number of elements, so the pre-reference point is a candidate reference point.

[0051] Furthermore, after the candidate benchmark points are screened out, a benchmark point is screened out from the candidate benchmark points according to the analysis method of the pre-benchmark points. Specifically, the number of candidate benchmark points is counted. If there is only one candidate benchmark point, the candidate benchmark point is directly used as the benchmark point. Otherwise, if the number of candidate benchmark points is not 1, any candidate benchmark point can be selected from all the candidate benchmark points as the benchmark point.

[0052] The step of comparing the tag value set of each pixel to be analyzed with the reference tag value set of the reference point to obtain a supplementary reference point in the surface grayscale image includes:

[0053] For any pixel to be analyzed, detect whether the label values ​​included in the label value set and the reference label value set corresponding to the pixel to be analyzed and the reference point in the same preset neighborhood direction are the same; if it is detected that the label values ​​included in the label value set and the reference label value set corresponding to the pixel to be analyzed and the reference point in the same preset neighborhood direction are the same, take the label value set of the pixel to be analyzed as the target label value set;

[0054] Counting the number of label values ​​included in each of the target label value sets, and if there is at least one target label value set in which the number of label values ​​included is greater than or equal to a preset label value number threshold, determining the pixel point to be analyzed as a supplementary reference point in the surface grayscale image;

[0055] If the number of label values ​​contained in each of the target label value sets is less than the label value number threshold, then the number of target label value sets with the same number of label values ​​is counted based on the number of label values ​​contained in each of the target label value sets, and if the number is greater than or equal to a preset number, the pixel point to be analyzed is determined to be a supplementary reference point in the surface grayscale image.

[0056] In one embodiment, after obtaining the label value set of each pixel to be analyzed and the reference label value set of the reference point, taking a pixel to be analyzed as an example, the label value set of the pixel to be analyzed and the reference point in the same neighborhood direction are compared. For the upper neighborhood direction, the label value set of the pixel to be analyzed in the upper neighborhood direction is compared with the reference label value set of the reference point in the upper neighborhood direction to detect whether the label values ​​in the two label value sets are the same. If they are the same, the label value set of the pixel to be analyzed in the upper neighborhood direction is used as the target label value set. Similarly, the label value set of the lower neighborhood direction of the pixel to be analyzed is respectively compared with the reference label value set of the lower neighborhood direction of the pixel to be analyzed. The upper domain direction, the left neighbor direction and the right neighbor direction are compared to obtain the set of all target label values ​​of the pixel to be analyzed. For example: assuming that the benchmark label value set of the benchmark point in the upper neighborhood direction is {-1, -1}, and the benchmark label value set of the pixel to be analyzed in the upper neighborhood direction is {-1, -1, -1, -1}, it can be seen that the elements in these two label value sets are all -1. At this time, it is considered that the label values ​​in the two label value sets are the same. If the benchmark label value set of the pixel to be analyzed in the upper neighborhood direction is {-1, 0, -1, -1}, the label values ​​in the two label value sets are different.

[0057] After obtaining the target tag value set of the pixel to be analyzed, the number of tag values ​​contained in each target tag value set is counted to detect whether there is a representative target tag value set for subsequent judgment of whether the pixel to be analyzed is a supplementary reference point. Since the number of tag values ​​contained in the target tag value set is too small, it is not representative and cannot reflect the grayscale change of the pixel to be analyzed. Therefore, if there is at least one target tag value set with a number of tag values ​​greater than or equal to 2, it is determined that the pixel to be analyzed is a supplementary reference point in the surface grayscale image, indicating that the grayscale change rate of the pixel to be analyzed in at least one neighborhood direction belongs to the grayscale change rate in the defective or colored defect area.

[0058] On the contrary, if the number of tag values ​​contained in each target tag value set of the pixel to be analyzed is less than 2, the number of target tag value sets with the same number of tag values ​​is detected. If there are 3 target tag value sets with the same number of tag values, the pixel to be analyzed is also determined to be a supplementary reference point in the surface grayscale image, otherwise the pixel to be analyzed does not belong to the reference point being sought. Similarly, by judging each pixel to be analyzed, all supplementary reference points in the surface grayscale image can be obtained.

[0059] At this point, all initial reference points in the surface grayscale image are obtained with all target points as starting points, and all supplementary reference points are obtained based on all initial reference points in the surface grayscale image, and then the supplementary reference points in the surface grayscale image and the initial reference points in the surface grayscale image constitute all the reference points in the surface grayscale image.

[0060] Step S103, obtaining the maximum grayscale change rate and the minimum grayscale change rate among the grayscale change rates corresponding to all reference points in the surface grayscale image, obtaining the grayscale change rate interval, dividing the grayscale change rate interval into at least two sub-intervals according to a preset interval interval, and dividing all reference points in the surface grayscale image into at least two sets according to the grayscale change rate corresponding to each sub-interval and each reference point in the surface grayscale image.

[0061] Since some areas with smaller grayscale values ​​are located in the middle of areas with larger grayscale values, data loss may occur during region growing segmentation. Therefore, in order to avoid the loss of areas with smaller grayscale values, the seed point should be as close to the area as possible. However, since the grayscale values ​​of the reference points in the area are lower than the grayscale values ​​of most reference points, it is necessary to determine the grayscale value interval of the seed point by finding the middle value, so that the grayscale value interval of the seed point should have as much correlation as possible with other reference points, so that the seed point can cover all areas with high and low grayscale values. That is, the position of the seed point is preferably where two places with larger grayscale values ​​can be connected in the area, and at the same time, it is associated with as many other reference points as possible. Therefore, according to the grayscale change rate of all reference points in the surface grayscale image, all reference points in the surface grayscale image are divided into at least two sets, which are used for subsequent reference points based on the division to find one or more reference points with the highest correlation, and filter out qualified reference points as possible. Seek as possible seed points.

[0062] The specific process of dividing all reference points in the surface grayscale image into at least two sets according to the grayscale change rates of all reference points in the surface grayscale image is as follows:

[0063] (1) Obtaining a maximum grayscale change rate and a minimum grayscale change rate from the grayscale change rates corresponding to all reference points in the surface grayscale image to obtain a grayscale change rate interval, and dividing the grayscale change rate interval into at least two sub-intervals according to a preset interval interval.

[0064] Specifically, the maximum grayscale change rate and the minimum grayscale change rate are obtained according to the grayscale change rate of all reference points in the surface grayscale image to form a grayscale change rate interval. After determining the grayscale change rate interval of all reference points in the surface grayscale image, the grayscale change rate interval is divided into sub-intervals according to a preset interval interval. Assuming that the preset interval interval is 4%, the grayscale change rate interval can be divided into the following sub-intervals according to 4%: [0%, 4%] [5%, 9%] [10%, 14%] [15%, 20%] ..., so that the grayscale change rate interval is divided into multiple sub-intervals.

[0065] (2) Dividing all reference points in the surface grayscale image into at least two sets according to the grayscale change rate corresponding to each sub-interval and each reference point in the surface grayscale image.

[0066] Specifically, for any sub-interval, any reference point in the sub-interval where there is at least one grayscale change rate in the surface grayscale image is taken as the first traversal point, and according to the grayscale change rate between the first traversal point and each of its neighborhood reference points, the neighborhood reference point corresponding to the grayscale change rate in the sub-interval is obtained as the first reference point, and the neighborhood reference point is taken as the first traversal point, and the step of obtaining the first reference point is repeated to obtain all first reference points including the first traversal point, and according to the positions of all first reference points, at least two first reference points with consecutive positions are grouped into a set to obtain at least one set under the sub-interval; according to the sets under each of the sub-intervals, at least two sets after dividing all reference points in the surface grayscale image are obtained.

[0067] In one embodiment, reference Figure 4 , which is a schematic diagram of the grayscale change rate of all reference points in a surface grayscale image. Taking a subinterval [0%, 4%] as an example, the reference points with grayscale change rates in the same subinterval and adjacent to each other are classified into the same category set. Figure 4 The reference point 13 in the last row of the second column is the first traversal point. The two grayscale change rates of the reference point 13 are both in the sub-interval [0%, 4%]. Figure 4 The reference point 13 in the second to last row of the second column and the reference point 13 in the last row of the third column are the first reference points to be found, and then the first reference points of the grayscale change rate in the sub-interval [0%, 4%] are searched with these two first reference points as the center, until the first reference point of the grayscale change rate in the sub-interval [0%, 4%] is not found, and these first reference points and the first traversal points are combined into a set. Figure 4 There are multiple identical reference points in the image. In order to distinguish the positional relationship of the identical reference points, the positions of the identical reference points are distinguished by using the identifiers a, b, c, and d. Figure 5 , which is a Figure 4 Schematic diagram of the position identification of all reference points in the subinterval [0%, 4%] based on the Figure 5 Find the set A{13g, 13f, 13h, 13i, 13.5, 14e, 14.5, 14a, 14d} of grayscale change rate that matches the subinterval. Similarly, obtain the set of reference points corresponding to the grayscale change rate of each subinterval. Then combine Figure 4 and Figure 5 , can Figure 5 The set obtained by dividing all reference points in is:

[0068] Set A{13g, 13f, 13h, 13i, 13.5, 14e, 14.5, 14a, 14d}; Set B{13b, 13c, 13d, 13e}; Set C{15.5d, 15.5c, 15.5b, 15.5a, 15, 16b}; Set D{10.8, 11a, 10.5}; Set E{12d, 12e, 12f, 12g} ; Set F{19, 18, 17, 16a, 15, 14a, 13a, 12c, 14b, 13b, 12d, 11b, 14e, 13i, 14d}; Set G{12b, 13c, 12e}; Set H{14f, 13j, 14c}; Set I{14c, 15.5b, 16b, 14d}; Set J{14b, 12b, 14a, 12a}.

[0069] At this point, all reference points in the grayscale surface image are divided into multiple sets according to sub-intervals.

[0070] Step S104, respectively obtain the grayscale mean of all pixel points contained in each set, and select the best seed point from all reference points in the surface grayscale image according to the grayscale mean corresponding to each set, perform regional growth segmentation on the surface grayscale image based on the best seed point to obtain the area to be repaired, and perform 3D modeling of the defective tooth according to the area to be repaired.

[0071] Since the location of the best seed point to be found needs to be where two grayscale values ​​with large differences can be connected, and at the same time it needs to be associated with all reference points to ensure the best effect when performing regional growing, therefore, in an embodiment of the present invention, the grayscale mean of all pixel points contained in each set is obtained respectively, and according to the grayscale mean corresponding to each set, the best seed point is screened from all reference points in the surface grayscale image.

[0072] Among them, according to the grayscale mean value corresponding to each set, the best seed point is screened from all reference points in the surface grayscale image, including:

[0073] (1) According to the grayscale values ​​of the reference points in all the sets, the average grayscale value is obtained as the total grayscale mean, and according to the grayscale mean corresponding to each of the sets and the total grayscale mean, an intermediate grayscale value interval is constructed, and according to the grayscale mean corresponding to each of the sets, a set whose grayscale mean is in the intermediate grayscale value interval is obtained as the target set.

[0074] In one embodiment, based on the grayscale value of each reference point in the surface grayscale image, the grayscale value mean is calculated as the total grayscale mean. At the same time, the grayscale means of all reference points in each set are calculated. One set corresponds to one grayscale mean. Based on the grayscale means corresponding to each of the sets, the minimum grayscale mean is obtained, and the minimum grayscale mean and the total grayscale mean are combined to form an intermediate grayscale value interval. According to the grayscale mean corresponding to each set, the set with the grayscale mean in the middle grayscale value interval is obtained as the target set. For example, the existing data can be divided into 4 sets N, M, K, and L. The elements in the sets are N{8, 9, 10, 11}; M{19, 17, 15, 13}; K{20, 21, 22, 23, 24}; L{19, 20, 21, 22}, where the total grayscale mean is 17.3, the grayscale mean of set N is 9.5, the grayscale mean of set M is 16, the grayscale mean of set K is 22, and the grayscale mean of set L is 20.5. The corresponding middle grayscale value interval is (17.3, 9.5). The set that satisfies the middle grayscale value interval is set M, and set M is used as the target set for subsequent acquisition of the best seed point according to the target set.

[0075] (2) Obtaining reference points whose grayscale values ​​correspond to the intermediate grayscale value interval in the target set as candidate reference points, counting the number of occurrences of each candidate reference point in all sets, obtaining the candidate reference point with the largest number of occurrences as the candidate reference point, and screening out the best seed point from all candidate reference points.

[0076] Specifically, according to the gray value of each reference point in the target set, reference points with gray values ​​in the middle gray value interval are obtained as candidate reference points, the number of occurrences of the candidate reference points in all sets is counted, and the candidate reference points with the largest number of occurrences are obtained as candidate reference points, and then the best seed points are screened from all candidate reference points. For example: the target set is M{19, 17, 15, 13}, and the middle gray value interval is {15, 9}. Reference points that meet the middle gray value interval are selected in the target set M as candidate reference points, that is, 15 and 13. If there is only one candidate reference point, it is used as a candidate reference point. If the number of candidate reference points is not 1, further screening is required, and the number of occurrences of each candidate reference point in all sets is counted. One or more candidate reference points with the largest number of occurrences are used as candidate reference points to screen out reference points that may be the best seed points.

[0077] Among them, the best seed point is screened out from all candidate reference points, including:

[0078] If the number of the candidate reference points is 1, the candidate reference point is used as the best seed point;

[0079] If the number of the candidate reference points is not 1, for any candidate reference point, in the surface grayscale image, calculate the abscissa difference between the abscissa value of the candidate reference point and the abscissa value of each of the reference points, obtain the mean absolute value of the abscissa difference as the abscissa offset of the candidate reference point, calculate the ordinate difference between the ordinate value of the candidate reference point and the ordinate value of each of the reference points, obtain the mean absolute value of the ordinate difference as the ordinate offset of the candidate reference point, and calculate the sum of the abscissa offset and the ordinate offset of the candidate reference point as the total offset;

[0080] According to the total offset of each candidate reference point, the candidate reference point corresponding to the minimum total offset is used as the best seed point.

[0081] In one embodiment, after obtaining the candidate reference point, if the number of candidate reference points is only one, the candidate reference point is directly used as the best seed point; otherwise, if the number of candidate reference points is greater than 1, considering that the closer the seed point is to the center position, the better, for any candidate reference point, according to the position difference between the candidate reference point and each reference point in the surface grayscale image, the position offset (total offset) of the candidate reference point is obtained, and then the candidate reference point corresponding to the smallest position offset is selected as the best seed point, wherein the calculation formula of the position offset of the candidate reference point is:

[0082]

[0083]

[0084]

[0085] Where c(x) represents the horizontal coordinate offset of the candidate reference point, c(y) represents the vertical coordinate offset of the candidate reference point, C represents the total offset of the candidate reference points, and n represents the number of reference points in the surface grayscale image. represents the abscissa value of the i-th reference point in the surface grayscale image, x represents the abscissa value of the candidate reference point, represents the ordinate value of the i-th reference point in the surface grayscale image, y represents the ordinate value of the candidate reference point, and || represents the absolute value sign.

[0086] It should be noted that because the closer the best seed point is to the center position, the better the data accuracy, the total offset between each candidate reference point and all reference points is calculated respectively, and the one with the smaller total offset is the best seed point.

[0087] After determining the best seed point in the surface grayscale image, the surface grayscale image can be segmented by region growing based on the best seed point to obtain the area to be repaired, and then the 3D modeling of the defective tooth is performed according to the area to be repaired. The specific process is as follows:

[0088] After obtaining the best seed point, the best seed point is used as the seed point of the region growing algorithm, and the region is grown in the surface grayscale image with the best seed point. The area to be repaired is segmented in the surface grayscale image, that is, the defective area with defects in the defective teeth or the defective area with colored teeth. According to the defect or coloring location, depth, and the needs of the doctor or patient, the area to be repaired is 3D modeled, and the corresponding restoration is obtained by 3D printing technology. The suitability and comfort of the restoration are tested in the patient's mouth, and then adjustments are made to obtain the final restoration for tooth veneer restoration. It should be noted that the region growing algorithm belongs to the existing technology and will not be repeated here.

[0089] At this point, it is possible to find the optimal seed point when using the region growing method to segment the image of the defective tooth surface, and obtain a more accurate area to be repaired, that is, the area where there is a defect on the tooth surface, further improving the quality of tooth 3D modeling based on the area to be repaired.

[0090] In summary, in the process of tooth veneer restoration, the embodiment of the present invention obtains a surface grayscale image of a defective tooth; according to the grayscale value of each pixel point in the surface grayscale image, a pixel point with a target grayscale value is selected as a target point; for any target point, the target point is used as a starting point to obtain a grayscale change rate between the target point and each of its neighboring pixel points; according to the grayscale change rate, a reference point of the target point is obtained; any neighboring pixel point of the target point is used as the target point, and the method for obtaining the reference point of the target point is repeated until no reference point can be obtained, and all reference points with the target point as the starting point are obtained; according to all the reference points obtained with each of the target points as the starting point, all reference points in the surface grayscale image are obtained; in the surface grayscale image The maximum grayscale change rate and the minimum grayscale change rate are obtained from the grayscale change rates corresponding to all reference points in the surface grayscale image to obtain a grayscale change rate interval, the grayscale change rate interval is divided into at least two sub-intervals according to a preset interval interval, and all reference points in the surface grayscale image are divided into at least two sets according to the grayscale change rate corresponding to each sub-interval and each reference point in the surface grayscale image; the grayscale mean values ​​of all pixel points contained in each of the sets are obtained respectively, and the best seed point is screened from all reference points in the surface grayscale image according to the grayscale mean values ​​corresponding to each of the sets, and the surface grayscale image is segmented by regional growth based on the best seed point to obtain the area to be repaired, and 3D modeling of the defective tooth is performed according to the area to be repaired. Among them, the pixel point that meets the target grayscale value is obtained as the target point, firstly, the grayscale change rate of the neighboring pixel points and the target point is calculated with the target point as the center, and then the grayscale change rate of the neighboring pixel points with the surrounding pixel points is calculated with the neighboring pixel point as the new center, and the pixel point with the same grayscale change rate is found as the reference point, which is used to screen the pixel points with similar grayscale change rate that are suspected to belong to the tooth defect, and then all the reference points in the surface grayscale image are divided into multiple sets according to the maximum grayscale change rate and the minimum grayscale change rate among all the reference points, so that the reference points with the same grayscale change rate are regarded as a set. In order to ensure that the reference points with too large grayscale value difference are not omitted, the best seed point is screened from all the reference points in the surface grayscale image according to the grayscale mean value of each set, so that the obtained best seed point is more in line with the scene, so that the surface grayscale image is regionally grown and segmented based on the best seed point, and a more accurate area to be repaired is obtained, that is, the area where there is a defect on the tooth surface, further improving the quality of tooth 3D modeling according to the area to be repaired.

[0091] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A tooth 3D modeling method for tooth veneer restoration, characterized in that: The method comprises: During the dental veneer restoration process, a grayscale image of the defective tooth surface is obtained; According to the gray value of each pixel in the surface gray image, a pixel with a target gray value is selected as a target point, and for any target point, the gray change rate between the target point and each of its neighboring pixels is obtained with the target point as the starting point, and a reference point of the target point is obtained according to the gray change rate, and any neighboring pixel of the target point is used as a new starting point, and the reference point obtaining method is repeated until no reference point can be obtained, and all reference points with the target point as the starting point are obtained, and all reference points in the surface gray image are obtained according to all reference points obtained with each of the target points as the starting point; Obtaining a maximum grayscale change rate and a minimum grayscale change rate from the grayscale change rates corresponding to all reference points in the surface grayscale image to obtain a grayscale change rate interval, dividing the grayscale change rate interval into at least two sub-intervals according to a preset interval interval, and dividing all reference points in the surface grayscale image into at least two sets according to each of the sub-intervals and the grayscale change rate corresponding to each reference point in the surface grayscale image; Respectively obtaining the grayscale mean of all the pixel points included in each of the sets, screening out the best seed point from all the reference points in the surface grayscale image according to the grayscale mean corresponding to each of the sets, performing region growing segmentation on the surface grayscale image based on the best seed point to obtain the area to be repaired, and performing 3D modeling of the defective tooth according to the area to be repaired; The step of acquiring a reference point of the target point according to the grayscale change rate includes: For the grayscale change rate between the target point and each of its neighboring pixel points, if there are at least two identical grayscale change rates, the neighboring pixel points corresponding to the identical grayscale change rates and the target point are used as reference points; If there is no identical grayscale change rate, the difference between every two grayscale change rates is obtained, and the neighborhood pixel points corresponding to the difference in the preset difference interval and the target point are used as reference points.

2. The tooth 3D modeling method for tooth veneer restoration according to claim 1, characterized in that: The step of acquiring all reference points in the surface grayscale image based on all reference points acquired with each target point as a starting point comprises: After obtaining the initial reference point in the surface grayscale image with each target point as a starting point, according to the position distribution of all the initial reference points in the surface grayscale image, the initial reference point at the edge of the surface grayscale image and the last grayscale change rate analyzed is obtained as the target reference point; For any target reference point, obtain neighboring pixel points that are not the initial reference point in the neighboring pixel points of the target reference point as the pixel points to be analyzed, and respectively obtain a set of label values ​​of the grayscale change rate of each pixel point to be analyzed in each preset neighborhood direction, wherein the label value is obtained according to the numerical attribute of the grayscale difference between two adjacent pixel points; Based on the position distribution of all the initial reference points, obtaining a set of label values ​​formed by each of the initial reference points and other initial reference points in each of the preset neighborhood directions, counting the number of label value sets of each of the initial reference points, obtaining the initial reference points corresponding to the maximum number as pre-reference points, screening out reference points from all the pre-reference points, and using the label value sets of the reference points as the reference label value set; The label value set of each pixel to be analyzed is compared with the benchmark label value set of the benchmark point to obtain supplementary reference points in the surface grayscale image. The supplementary reference points in the surface grayscale image and the initial reference points in the surface grayscale image constitute all reference points in the surface grayscale image.

3. The tooth 3D modeling method for tooth veneer restoration according to claim 2, characterized in that: The step of screening out the reference points from all the pre-reference points includes: Counting the number of the pre-reference points, and if the number of the pre-reference points is 1, taking the pre-reference point as a reference point; If the number of the pre-reference points is not 1, for any pre-reference point, the number of elements in each label value set of the pre-reference point is counted respectively, and if the number of elements in all label value sets of the pre-reference point belongs to the maximum number of elements, the pre-reference point is taken as a candidate reference point; The number of all candidate reference points is counted. If the number of all candidate reference points is 1, the candidate reference point is used as the reference point. If the number of all candidate reference points is not 1, any candidate reference point is selected from all candidate reference points as the reference point.

4. The tooth 3D modeling method for tooth veneer restoration according to claim 2, characterized in that: The step of comparing the tag value set of each pixel to be analyzed with the reference tag value set of the reference point to obtain a supplementary reference point in the surface grayscale image includes: For any pixel to be analyzed, detect whether the label values ​​included in the label value set and the reference label value set corresponding to the pixel to be analyzed and the reference point in the same preset neighborhood direction are the same; if it is detected that the label values ​​included in the label value set and the reference label value set corresponding to the pixel to be analyzed and the reference point in the same preset neighborhood direction are the same, take the label value set of the pixel to be analyzed as the target label value set; Counting the number of label values ​​included in each of the target label value sets, and if there is at least one target label value set in which the number of label values ​​included is greater than or equal to a preset label value number threshold, determining the pixel point to be analyzed as a supplementary reference point in the surface grayscale image; If the number of label values ​​contained in each of the target label value sets is less than the label value number threshold, then the number of target label value sets with the same number of label values ​​is counted based on the number of label values ​​contained in each of the target label value sets, and if the number is greater than or equal to a preset number, the pixel point to be analyzed is determined to be a supplementary reference point in the surface grayscale image.

5. The tooth 3D modeling method for tooth veneer restoration according to claim 1, characterized in that: The step of dividing all reference points in the surface grayscale image into at least two sets according to the grayscale change rate corresponding to each sub-interval and each reference point in the surface grayscale image comprises: For any sub-interval, any reference point in the sub-interval where at least one grayscale change rate exists in the surface grayscale image is taken as a first traversal point, and according to the grayscale change rate between the first traversal point and each of its neighborhood reference points, a neighborhood reference point corresponding to the grayscale change rate in the sub-interval is obtained as a first reference point, and the neighborhood reference point is taken as the first traversal point, and the step of obtaining the first reference point is repeated to obtain all first reference points including the first traversal point, and according to the positions of all the first reference points, at least two first reference points with consecutive positions are grouped into a set to obtain at least one set under the sub-interval; According to the set under each of the subintervals, at least two sets are obtained after all reference points in the surface grayscale image are divided.

6. The tooth 3D modeling method for tooth veneer restoration according to claim 1, characterized in that: The step of selecting the best seed point from all reference points in the surface grayscale image according to the grayscale mean value corresponding to each set includes: According to the grayscale values ​​of the reference points in all the sets, the average grayscale value is obtained as the total grayscale mean, and according to the grayscale mean corresponding to each of the sets and the total grayscale mean, an intermediate grayscale value interval is constructed, and according to the grayscale mean corresponding to each of the sets, a set with the grayscale mean in the intermediate grayscale value interval is obtained as the target set; In the target set, reference points whose grayscale values ​​correspond to the intermediate grayscale value interval are obtained as candidate reference points, the number of occurrences of each candidate reference point in all sets is counted, the candidate reference point with the largest number of occurrences is obtained as the candidate reference point, and the best seed point is screened from all the candidate reference points.

7. The tooth 3D modeling method for tooth veneer restoration according to claim 6, characterized in that: The step of constructing an intermediate grayscale value interval according to the grayscale mean corresponding to each of the sets and the total grayscale mean comprises: According to the grayscale mean corresponding to each of the sets, a minimum grayscale mean is obtained, and the minimum grayscale mean and the total grayscale mean are combined to form an intermediate grayscale value interval.

8. The tooth 3D modeling method for tooth veneer restoration according to claim 6, characterized in that: The best seed point is obtained by screening all candidate reference points, including: If the number of the candidate reference points is 1, the candidate reference point is used as the best seed point; If the number of the candidate reference points is not 1, for any candidate reference point, in the surface grayscale image, calculate the abscissa difference between the abscissa value of the candidate reference point and the abscissa value of each of the reference points, obtain the mean absolute value of the abscissa difference as the abscissa offset of the candidate reference point, calculate the ordinate difference between the ordinate value of the candidate reference point and the ordinate value of each of the reference points, obtain the mean absolute value of the ordinate difference as the ordinate offset of the candidate reference point, and calculate the sum of the abscissa offset and the ordinate offset of the candidate reference point as the total offset; According to the total offset of each candidate reference point, the candidate reference point corresponding to the minimum total offset is used as the best seed point.