CN111311653B - Method for registering dental plaque fluorescent image and tooth three-dimensional model - Google Patents

Abstract

The invention discloses a method for registering dental plaque fluorescent images and a tooth three-dimensional model, which comprises the following steps: s11, acquiring three-dimensional model point cloud data P of tooth surfaces _s Fluorescence image F of occlusal surface of tooth _Img The method comprises the steps of carrying out a first treatment on the surface of the S12, fluorescence image F of occlusal surface of tooth _Img Processing to obtain tooth region and contour information Edge ₁ The method comprises the steps of carrying out a first treatment on the surface of the S13, carrying out point cloud data P on the three-dimensional model of the tooth surface _s Posture correction is carried out to obtain a corrected tooth three-dimensional model P _s The method comprises the steps of carrying out a first treatment on the surface of the S14, establishing a corrected tooth three-dimensional model P _s Fluorescence image F of occlusal surface of tooth _Img Is to be corrected into a three-dimensional tooth model P _s Mapping of points on to fluorescence image F _Img And (3) upper part. According to the invention, the mapping model is solved iteratively, so that the registration of the occlusal surface contour of the 2D fluorescent image and the 3D projection contour is realized, the corresponding relation between the dental plaque 2D fluorescent image and the dental occlusal surface 3D data is obtained, the three-dimensional spatial distribution of dental plaque on the dental surface is determined, and a foundation is provided for researching the three-dimensional distribution of dental plaque on the dental surface.

Description

A method for registering dental plaque fluorescence images with three-dimensional tooth models

Technical Field

The present invention relates to the technical field of computer-aided intelligent dental caries prevention, and in particular to a method for registering a dental plaque fluorescence image with a three-dimensional tooth model.

Background Art

Caries is a chronic and progressive disease that destroys dental hard tissues. It is prevalent worldwide. The World Health Organization has listed it as one of the three major non-communicable diseases that humans should focus on preventing and treating, along with cancer and cardiovascular disease. If treatment is delayed, it will not only cause local pain, infection, and tooth loss, but is also closely related to systemic diseases such as cardiovascular and diabetes. Caries prevention guided by caries risk assessment (CRA) is considered the basis of modern caries prevention and management. Such a system can predict the possibility of new caries in an individual within a certain period of time. The caries-related factors included in the current CRA system overlap to a large extent, such as caries experience, saliva, diet, systemic condition, and fluoride exposure, and the impact of tooth morphology on the risk of caries is insufficiently considered.

At present, visual examination, probe, X-ray and other methods are commonly used in clinical diagnosis of caries. Visual examination mainly relies on the visual observation and experience judgment of dentists, which is highly subjective and cannot diagnose early caries. Dental probing is to judge whether the tip of the probe can be hooked on the surface of the tooth enamel and whether a caries cavity is formed by scratching the tip of the probe across the tooth enamel surface, and it is also impossible to diagnose caries in the early stage. In order to transform caries from "destructive treatment" to "early prevention", non-destructive and quantifiable optical detection methods have become a research hotspot for early diagnosis of caries. Based on the material composition and structural characteristics of tooth tissue and the different mechanisms of interaction between light and tissue, the currently developed optical caries early detection methods include fluorescence technology, optical coherence tomography (OCT), Raman spectroscopy analysis, etc. Fluorescence technology uses the difference in the autofluorescence spectrum of dental plaque formed by the mixture of oral caries bacteria and food residues under ultraviolet light excitation. By collecting and analyzing the spectral distribution, the content of dental plaque can be quantified as a characterization parameter of caries risk. The fluorescence image distribution of carious and healthy teeth was analyzed by quantitative light-guided fluorescence (QLF) technology, and compared with the results of polarization microscopy, confirming the effectiveness of the fluorescence method.

At present, the research on caries is relatively traditional. It is this series of problems that have promoted the exploration and application of optical technology, 3D digitization technology and artificial intelligence technology in this field, and many intelligent medical information processing technologies have emerged. However, whether in the field of treatment or prevention, teeth are the host of caries, but there is little research on their morphological structure. It is of great significance to study the relationship between tooth surface morphology and caries. This research first needs to establish the connection between the plaque distribution directly related to caries and the three-dimensional morphology of teeth. Therefore, the present invention proposes a method and system for aligning dental plaque fluorescence images with tooth three-dimensional models.

Summary of the invention

The purpose of the present invention is to address the defects of the prior art and provide a method for aligning a dental plaque fluorescence image with a tooth three-dimensional model. By iteratively solving a mapping model, the occlusal surface contour of a 2D fluorescence image and a 3D projection contour are aligned, the correspondence between the 2D fluorescent image of dental plaque and the 3D data of the tooth occlusal surface is obtained, the three-dimensional spatial distribution of dental plaque on the tooth surface is determined, and a basis is provided for studying the three-dimensional distribution of dental plaque on the tooth surface.

In order to achieve the above objectives, the present invention adopts the following technical solutions:

A method for registering a dental plaque fluorescence image with a dental three-dimensional model, comprising:

S1. Obtaining the three-dimensional model point cloud data P _s of the tooth surface and the fluorescent image F _Img of the tooth occlusal surface;

S2. Processing the fluorescence image F _Img of the tooth occlusal surface to obtain tooth area and contour information Edge ₁ ;

S3. Performing posture correction on the three-dimensional model point cloud data P _s of the tooth surface to obtain a corrected three-dimensional tooth model P _s ;

S4. Establish a mapping model between the corrected three-dimensional tooth model _Ps and the fluorescent image F _Img of the tooth occlusal surface, and map the points on the corrected three-dimensional tooth model _Ps to the fluorescent image F _Img .

Furthermore, the step S1 specifically includes:

S11. Acquire 3D data of the dentition through an oral 3D scanner to obtain a 3D model P _s of the tooth surface;

S12. Obtain a fluorescent image F _Img of the plaque on the occlusal surface of the tooth using an oral fluorescent imager.

Furthermore, the step S2 specifically includes:

S21. Inputting the fluorescence data corresponding to the tooth three-dimensional model _Ps into the fluorescence image F _Img , and processing the fluorescence image F _Img of the tooth occlusal surface;

S22. Use an automatic threshold segmentation method to separate the tooth region from other regions to obtain the tooth region and contour information Edge ₁ .

Furthermore, the method for processing the fluorescence image F _Img of the tooth occlusal surface in step S21 is mean filtering.

Furthermore, the automatic threshold segmentation method adopted in step S22 is OTSU's global automatic threshold segmentation method.

Furthermore, the step S3 specifically performs posture correction on the tooth three-dimensional model point cloud data _Ps through a three-dimensional model space transformation formula.

Furthermore, in step S3, the posture correction of the three-dimensional model point cloud data _Ps is performed by using a PCA principal component analysis method.

Furthermore, the step S4 specifically includes:

S41. Calibrate the oral fluorescence imager for acquiring the fluorescence image of the tooth occlusal surface to obtain the internal parameter matrix M ₁ of the oral fluorescence imager;

S42. Input the corrected tooth three-dimensional model P _s , the contour information Edge ₁ of the tooth fluorescence image F _Img , the internal parameter matrix M ₁ , initialize the external parameter matrix M ₂ , and set the matching distance threshold _Ths ;

S43. Calculate the imaging image TPD _img of the three-dimensional tooth model in the virtual imaging system using the external parameter matrix M ₂ and the internal parameter matrix M ₁ ;

S44. Preprocessing the image TPD _img imaged by the virtual imaging system, and calculating the tooth edge information of the image TPD _img imaged by the virtual imaging system, to obtain the projection edge Edge ₂ of the tooth three-dimensional model;

S45. The point with the closest Euclidean distance from each two-dimensional coordinate point in the projection edge Edge ₂ of the calculated three-dimensional tooth model to each two-dimensional coordinate point in the contour information Edge ₁ of the tooth fluorescence image is saved; the corresponding relationship and distance information are saved;

S46. Obtain the two-dimensional point coordinates of the projection edge Edge ₂ of the three-dimensional model of the tooth and the corresponding points on the corresponding three-dimensional model _Ps , and calculate the average value T _avg of the distances of all corresponding points; determine whether the average value T _avg is less than the set matching distance threshold T _hs , if so, output the current three-dimensional projection matrix M ₂ as the registration matrix between the fluorescent image and the three-dimensional model data; if not, use the three-dimensional projection matrix equation to update the external parameter matrix M ₂ , and continue to execute step S43;

S47. Project the points on the three-dimensional tooth model _Ps onto the fluorescent image using the three-dimensional projection matrix according to the final external parameter matrix _M2 and the internal parameter matrix _M1 .

Furthermore, the step S44 also includes obtaining the contour in the three-dimensional model of the tooth, and performing edge operations on the fluorescent image to obtain the contour points of the projection image and the three-dimensional space points corresponding to the contour points, and saving all the data.

Furthermore, in step S46, the extrinsic parameter matrix M ₂ is updated by using the least square method based on the obtained corresponding point information and the imaging model.

Compared with the prior art, the present invention has the following beneficial effects:

1. The processing object of the present invention is high-precision three-dimensional tooth point cloud data, which has fast processing speed, high precision, easy implementation and easy promotion.

2. The present invention proposes a method for aligning dental plaque fluorescence images with tooth three-dimensional models, which lays a foundation for studying the relationship between tooth three-dimensional morphology and caries.

3. The method for mapping dental plaque fluorescence images to tooth three-dimensional models of the present invention adopts a simple method of initializing the projection matrix based on prior knowledge, thereby reducing a large amount of calculation and making the algorithm simpler and more efficient.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of a method for aligning a dental plaque fluorescence image with a dental three-dimensional model provided in Example 1;

FIG2 is a schematic diagram of the preferred tooth point cloud data read in according to the first embodiment;

FIG3 is a schematic diagram of preferred tooth fluorescence image data read in according to the first embodiment;

FIG4 is a schematic diagram of the result after mean filtering of a tooth fluorescence image provided in Example 1;

FIG5 is a schematic diagram of the automatic threshold segmentation result of the tooth fluorescence image provided in Example 1;

FIG6 is a schematic diagram of tooth fluorescence image contour extraction results provided in Example 1;

FIG. 7 is a schematic diagram of the tooth point cloud data after correction provided in Example 1;

FIG8 is a schematic diagram of a camera calibration result diagram provided in Example 1;

FIG9 is a schematic diagram of a projection result of tooth point cloud data provided in Example 1;

FIG10 is a schematic diagram of the result after the tooth point cloud data provided in Example 1 is projected and filled with blanks;

FIG11 is a schematic diagram of a point cloud projection contour acquisition result diagram provided in Example 1;

FIG12 is a schematic diagram of the contour matching result in the iterative _M2 process provided in Example 1;

FIG13 is a schematic diagram of the result of the average distance of contours in the process of calculating _M2 provided in Example 1;

FIG. 14 is a schematic diagram of the final result of mapping the dental plaque fluorescence image onto the three-dimensional model provided in the first embodiment.

DETAILED DESCRIPTION

The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

The purpose of the present invention is to address the defects of the prior art and provide a method for registering a dental plaque fluorescence image with a three-dimensional tooth model.

Embodiment 1

This embodiment provides a method for registering a dental plaque fluorescence image with a dental three-dimensional model, as shown in FIG1 , comprising:

S11. Acquire the three-dimensional model point cloud data P _s of the tooth surface and the fluorescent image F _Img of the tooth occlusal surface;

S12. Processing the fluorescence image F _Img of the tooth occlusal surface to obtain tooth region and contour information Edge ₁ ;

S13. Performing posture correction on the three-dimensional model point cloud data P _s of the tooth surface to obtain a corrected three-dimensional tooth model P _s ;

S14. Establish a mapping model between the corrected three-dimensional tooth model _Ps and the fluorescent image F _Img of the tooth occlusal surface, and map the points on the corrected three-dimensional tooth model _Ps to the fluorescent image F _Img .

In step S11, the three-dimensional model point cloud data _Ps of the tooth surface and the fluorescence image F _Img of the tooth occlusal surface are acquired.

In this embodiment, firstly, an intraoral three-dimensional scanner is used to obtain the point cloud data _Ps of the three-dimensional model of the tooth surface, and then a plaque fluorescence imager is used to obtain the fluorescence image F _Img of the tooth occlusal surface. Specifically:

S111. Acquire 3D data of the dentition through an oral 3D scanner, then cut out a single molar to obtain a 3D model P _s of the tooth surface, as shown in FIG2 ;

S112. Obtain a fluorescent image F _Img of the plaque on the occlusal surface of the tooth using an oral fluorescent imager, as shown in FIG3 .

In step S12, the fluorescence image F _Img of the tooth occlusal surface is processed to obtain tooth region and contour information Edge ₁ .

In this embodiment, the tooth fluorescence image F _Img is processed: the tooth fluorescence image F _Img is processed to obtain tooth area and contour information. Specifically:

S121. Input the fluorescence data corresponding to the tooth three-dimensional model _Ps into the fluorescence image F _Img , and process the fluorescence image F _Img of the tooth occlusal surface;

The preprocessing method used for the plaque fluorescence image is mean filtering. Since the dental plaque fluorescence image is collected in a dark environment in the oral cavity, it contains a lot of noise. The mean filtering can effectively filter out the noise and improve the accuracy of extracting the tooth area from the entire image. The mean filtering result is shown in Figure 4.

S122. Use an automatic threshold segmentation method to separate the tooth region from other regions to obtain the tooth region and contour information Edge ₁ .

The automatic threshold segmentation method used is a commonly used image segmentation method. It has a simple implementation principle, small amount of calculation, and stable performance. Therefore, it is widely used in the field of image segmentation. Assume that the image f(x, y) consists of a dark background and a brighter target object. There is a significant difference in the grayscale value between the two. Therefore, a threshold T can be selected. The following formula is used to separate the target object from the background in the image. The newly generated image is g(x, y)

After the thresholding operation, the background part of the image is set to black, and the color of the target part remains its own color. In the process of image thresholding operation, the selection of gray threshold T is very important. The appropriate T value will effectively remove irrelevant information. In tooth fluorescence images, due to the complexity of the acquisition environment, the brightness of multiple fluorescence images is very different. Therefore, it is necessary to adopt an algorithm that can determine the threshold T according to the information of the image itself to achieve accurate extraction of the tooth area. This embodiment adopts a global automatic threshold segmentation method based on OTSU.

The principle is to divide the image into two parts, background and foreground, according to the grayscale characteristics of the image. The greater the difference between the background and foreground, the greater the gap between the two parts that constitute the image. The specific description of the algorithm is as follows;

Assume that the total number of pixels in an image is N, the pixel grayscale range is [0～L-1], and the number of pixels with grayscale value i is n _i , then the probability of occurrence of a pixel with grayscale value i is:

Set the segmentation threshold to T ₀ and divide the grayscale value of the entire image into two categories: C ₀ and C _1. The grayscale value of C ₀ is distributed between [C ₀ ～T ₀ -1], and the grayscale value corresponding to C ₁ is between [T ₀ ～L-1]. The probabilities w ₀ and w ₁ of C ₀ and C ₁ are respectively:

The mean u ₀ ,u ₁ of C ₁ and C ₁ is:

The grayscale mean of the entire image is:

μ＝ω ₀ μ ₀ +ω ₁ μ ₁

Define the inter-class variance of two grayscale partitions as:

σ ² =ω ₀ (μ ₀ -μ) ² +ω ₁ (μ ₁ -μ) ²

When determining the segmentation threshold T ₀ , let T ₀ start from 0 and increase to L-1 in steps of 1, record the inter-class variance σ ² corresponding to each T ₀ , and the T ₀ with the largest σ ² is the optimal segmentation threshold. This method is used to automatically segment the tooth fluorescence image, and the segmentation result is shown in Figure 5.

The method for extracting the edge points of the tooth contour is to traverse a pixel and determine whether the eight adjacent pixels contain both tooth region points and non-tooth region points. If so, it is a contour boundary point. If not, it is not a contour boundary point. The tooth edge Edge ₁ corresponding to the fluorescence image is obtained, where the data structure contained in Edge ₁ is {{I _x1 ,I _y1 },{I _x2 ,I _y2 },...,{I _xi ,I _yi }}, where {I _xi ,I _yi } is the image coordinate corresponding to the i-th edge point of the fluorescence image. The extracted contour structure is shown in Figure 6.

In step S13, posture correction is performed on the three-dimensional model point cloud data _Ps of the tooth surface to obtain a corrected three-dimensional tooth model _Ps .

Input the tooth three-dimensional point cloud data _Ps , calculate the center of gravity and main direction of the tooth point cloud data, and use the three-dimensional model space transformation formula to correct the posture of the tooth three-dimensional point cloud data so that the origin of the spatial coordinate system is at the center of gravity of the tooth point cloud data model, the occlusal surface is parallel to the XOY plane of the spatial coordinate system, and is perpendicular to the Z axis of the spatial coordinate system. The corrected tooth three-dimensional model point set is obtained as the new _Ps ; the corrected point cloud model is shown in Figure 7.

The posture correction of the tooth model Ps uses the PCA principal component analysis method to calculate the main direction of the point cloud, which is the Z-axis direction. The three-dimensional space coordinate conversion matrix is used to convert the main direction to the Z-axis direction. The three-dimensional space coordinate conversion formula is:

The PCA method is used to determine the R and T matrices.

In step S14, a mapping model between the corrected three-dimensional tooth model _Ps and the fluorescence image F _Img of the tooth occlusal surface is established, and points on the corrected three-dimensional tooth model _Ps are mapped to the fluorescence image F _Img .

In this embodiment, a mapping model between the three-dimensional tooth model _Ps and the plaque fluorescence image F _Img is established, and the points on the three-dimensional tooth model _Ps are mapped to the plaque fluorescence image F _Img . Specifically, the imaging image of the three-dimensional tooth model is calculated according to the camera imaging model, and the contour of the imaging image is compared with the contour of the fluorescence image, and the projection matrix is updated to achieve the same contour of the tooth imaging and the contour of the fluorescence image, so as to realize the mapping of the points on the three-dimensional tooth model _Ps to the fluorescence image. Specifically,

S141. Calibrate the oral fluorescence imager for acquiring the fluorescence image of the tooth occlusal surface to obtain the internal parameter matrix M ₁ of the oral fluorescence imager;

The intraoral fluorescence imaging device is calibrated using Zhang Zhengyou's calibration method to obtain the camera intrinsic parameter matrix M ₁ . The calibration plate image collected by the camera calibration is shown in Figure 8 . The camera intrinsic parameter matrix M ₁ is:

The Zhang Zhengyou calibration method was used to take multiple chessboard calibration plate pictures with an oral fluorescence imager to calibrate the imaging device and obtain its camera imaging internal parameter matrix.

S142. Input the corrected tooth three-dimensional model P _s , the contour information Edge ₁ of the tooth fluorescence image F _Img , the internal parameter matrix M ₁ , initialize the external parameter matrix M ₂ , and set the matching distance threshold _Ths ;

Input the corrected tooth three-dimensional model P _s , the corresponding tooth fluorescence image F _Img contour data Edge ₁ , and the camera internal parameter matrix M ₁ ; initialize the external parameter matrix M ₂ , and set the matching distance threshold _Ths .

The projection matrix is initialized using empirical values. When collecting dental fluorescence images, the acquisition angle is perpendicular to the tooth occlusal surface and is about 1 to 2 cm away from the tooth surface. Therefore, the external parameter matrix _M2 can be initialized as

S143. Calculate the imaging image TPD _img of the three-dimensional tooth model in the virtual imaging system using the external parameter matrix M ₂ and the internal parameter matrix M ₁ ;

The imaging relationship model of the imaging system is used to obtain the three-dimensional tooth model in the fluorescence imaging system. The imaging model is:

Among them, _ax = c/ _dx , is the scale factor on the u-axis, also called the equivalent (normalized) focal length on the u-axis; _ay = c/ _dy , is the scale factor on the v-axis, also called the equivalent (normalized) focal length on the v-axis; M is a 3×4 matrix, called the projection matrix, _M1 is determined by the camera parameters _ax , _ay , _u0 , _v0 , called the intrinsic parameter matrix, _{and M2} is determined by the external parameters of the camera relative to the world coordinate system, called the extrinsic parameter matrix.

According to the camera perspective imaging matrix, the projection matrix of the tooth point cloud data in the _M1 and _M2 cases can be obtained. Here, because only the contour is used as the eigenvalue, it is not necessary to calculate the distance value between each point and the viewpoint camera coordinate to generate a grayscale image, thereby reducing a lot of calculations and speeding up the algorithm. Therefore, another simple method is used here, that is, to obtain a black and white image of the point cloud projection. When there is a corresponding projected three-dimensional point on the two-dimensional image point, the point is set to white, and other places are set to black. After the operation, a silhouette image of a tooth three-dimensional model is obtained, and its projection image is shown in Figure 9.

S144. Preprocessing the imaging image TPD _img under the virtual imaging system, and calculating the tooth edge information of the imaging image TPD _img under the virtual imaging system to obtain the projection edge Edge ₂ of the tooth three-dimensional model;

The imaging image TPD _img is preprocessed and the tooth edge information of the imaging image TPD _img is calculated to obtain the projection edge Edge ₂ of the tooth three-dimensional model. The data structure contained in Edge ₂ is {{{I _x1 ,I _y1 },{X _p1 ,Y _p1 ,Z _p1 }},{{I _x2 ,I _y2 },{X _p2 ,Y _p2 ,Z _p2 }},...,{{I _xj ,I _yj },{X _pj ,Y _pj ,Z _pj }}}, where {I _xj ,I _yj } is the image coordinate corresponding to the j-th edge point, and {X _pj ,Y _pj ,Z _pj } is the three-dimensional space point corresponding to the j-th projection edge point.

Since the tooth point cloud is relatively sparse and the resolution of the fluorescent image is relatively large, the obtained tooth model silhouette image forms points one by one. In order to accurately extract the contour, the image needs to be operated. The closing operation in image morphological processing is used here. The image is first expanded to fill small gaps, and then the original edge is restored to its original shape using corrosion of the same size. In this way, a complete and gap-free white silhouette of the tooth model can be obtained, and the image is shown in Figure 10.

When obtaining the projection contour, the same edge operation as the fluorescence image is used to obtain the projection image contour points and the three-dimensional space points corresponding to the contour points, and all data are saved. The contour image is shown in Figure 11.

S145. Save the corresponding relationship and distance information from each two-dimensional coordinate point in Edge ₂ of the projection edge of the calculated three-dimensional tooth model to the point with the closest Euclidean distance to each two-dimensional coordinate point in Edge ₁ of the contour information of the tooth fluorescence image;

In this embodiment, the point with the shortest Euclidean distance from each two-dimensional coordinate point {I _xj ,I _yj } in Edge ₂ to each two-dimensional coordinate point {I _xi ,I _yi } in Edge ₁ is calculated, and the corresponding relationship and distance information are saved.

S146. Obtain the two-dimensional point coordinates of the projection edge Edge ₂ of the three-dimensional model of the tooth and the corresponding point on the corresponding three-dimensional model _Ps , and calculate the average value T _avg of the distances of all corresponding points; determine whether the average value T _avg is less than the set matching distance threshold T _hs , if so, output the current three-dimensional projection matrix M ₂ as the registration matrix between the fluorescence image and the three-dimensional model data; if not, use the three-dimensional projection matrix equation to update the external parameter matrix M ₂ , and continue to execute step S143;

According to the 3σ principle, the wild points in Edge ₂ are deleted from the distance information. If there are too many remaining paired points, a part of the points are randomly selected from Edge ₂ to form a paired set U of about 200 points. The set U contains the coordinates of the two-dimensional points of the fluorescence image {I _xi ,I _yi } and the corresponding points {X _pj ,Y _pj ,Z _pj } on the corresponding three-dimensional P _s . The average value T _avg of the distances of all corresponding points is calculated. If T _avg <T _hs , the current projection matrix M ₂ is output as the registration matrix between the fluorescence image 2D and the three-dimensional model 3D data. Otherwise, the projection matrix M ₂ is updated using the three-dimensional projection matrix equation, and then the process returns to step S143 to continue iteration.

Using the corresponding point information and imaging model, the camera extrinsic matrix M ₂ is updated using the least squares method. The contour update process is shown in Figure 12. The final M ₂ matrix is

During the iteration process, the average value of all point distances decreases as shown in Figure 13.

According to the final obtained external parameter projection matrix _M2 and the fluorescence camera internal parameter matrix _M1 , the three-dimensional projection matrix can be used to project a certain point on the three-dimensional tooth model _Ps onto the fluorescence image. The mapping result is shown in Figure 14.

S147. Project the points on the three-dimensional tooth model _Ps onto the fluorescent image using the three-dimensional projection matrix according to the final external parameter matrix _M2 and the internal parameter matrix _M1 .

In this embodiment, according to the finally obtained projection matrix _M2 and the fluorescence camera internal parameter matrix _M1 , a three-dimensional projection matrix can be used to project a certain point on the three-dimensional tooth model _Ps onto the fluorescence image.

Compared with the prior art, the beneficial effects of this embodiment are:

1. The processing object of this embodiment is high-precision three-dimensional tooth point cloud data, which has fast processing speed, high precision, easy implementation and easy promotion.

2. This embodiment proposes a method for registering a dental plaque fluorescence image with a three-dimensional tooth model, which lays a foundation for studying the relationship between the three-dimensional morphology of teeth and caries.

3. This embodiment of the method for mapping dental plaque fluorescence images to tooth three-dimensional models adopts a simple method of initializing the projection matrix based on prior knowledge, thereby reducing a large amount of calculation and making the algorithm simpler and more efficient.

The specific embodiments described herein are merely examples of the spirit of the present invention. A person skilled in the art of the present invention may make various modifications or additions to the specific embodiments described or replace them in a similar manner, but this will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

Claims (8)

1. A method for registering a dental plaque fluorescence image with a dental three-dimensional model, comprising: S1. Obtaining the three-dimensional model point cloud data P _s of the tooth surface and the fluorescent image F _Img of the tooth occlusal surface; S2. Processing the fluorescence image F _Img of the tooth occlusal surface to obtain tooth area and contour information Edge ₁ ; S3. Performing posture correction on the three-dimensional model point cloud data P _s of the tooth surface to obtain a corrected three-dimensional tooth model P _s ; S4. Establishing a mapping model between the corrected three-dimensional tooth _{model Ps} and the fluorescent image F _Img of the tooth occlusal surface, mapping points on the corrected three-dimensional tooth _model Ps to the fluorescent image F _Img ; Step S2 specifically includes: S21. Inputting the fluorescence data corresponding to the tooth three-dimensional _model Ps into the fluorescence image F _Img , and processing the fluorescence image F _Img of the tooth occlusal surface; S22. Using an automatic threshold segmentation method to separate the tooth region from other regions, obtaining the tooth region and contour information Edge ₁ ; Step S4 specifically includes: S41. Calibrate the oral fluorescence imager for acquiring the fluorescence image of the tooth occlusal surface to obtain the internal parameter matrix M ₁ of the oral fluorescence imager; S42. Input the corrected tooth three-dimensional model P _s , the contour information Edge ₁ of the tooth fluorescence image F _Img , the internal parameter matrix M ₁ , initialize the external parameter matrix M ₂ , and set the matching distance _{threshold Ths} ; S43. Calculate the imaging image TPD _img of the three-dimensional tooth model in the virtual imaging system using the external parameter matrix M ₂ and the internal parameter matrix M ₁ ; S44. Preprocessing the image TPD _img imaged by the virtual imaging system, and calculating the tooth edge information of the image TPD _img imaged by the virtual imaging system, to obtain the projection edge Edge ₂ of the tooth three-dimensional model; S45. The point with the closest Euclidean distance from each two-dimensional coordinate point in the projection edge Edge ₂ of the calculated three-dimensional tooth model to each two-dimensional coordinate point in the contour information Edge ₁ of the tooth fluorescence image is saved; the corresponding relationship and distance information are saved; S46. Obtain the two-dimensional point coordinates of the projection edge Edge ₂ of the three-dimensional model of the tooth and the corresponding points on the corresponding three-dimensional _model Ps , and calculate the average value T _avg of the distances of all corresponding points; determine whether the average value T _avg is less than the set matching distance threshold T _hs , if so, output the current three-dimensional projection matrix M ₂ as the registration matrix between the fluorescent image and the three-dimensional model data; if not, use the three-dimensional projection matrix equation to update the external parameter matrix M ₂ , and continue to execute step S43; S47. Project the points on the three _- dimensional tooth model Ps onto the fluorescent image using the three-dimensional _projection matrix according to the final external parameter matrix M2 and _the internal parameter matrix M1 . 2. The method for registering a dental plaque fluorescence image with a dental three-dimensional model according to claim 1, wherein step S1 specifically comprises: S11. Acquire 3D data of the dentition through an oral 3D scanner to obtain a 3D model P _s of the tooth surface; S12. Obtain a fluorescent image F _Img of the plaque on the occlusal surface of the tooth using an oral fluorescent imager. 3. The method for aligning a dental plaque fluorescence image with a tooth three-dimensional model according to claim 1, characterized in that the processing method for processing the fluorescence image F _Img of the tooth occlusal surface in step S21 is mean filtering. 4. The method for aligning a dental plaque fluorescence image with a dental three-dimensional model according to claim 1, characterized in that the automatic threshold segmentation method used in step S22 is an OTSU global automatic threshold segmentation method. 5. The method for aligning a dental plaque fluorescence image with a dental three-dimensional model according to claim 1, characterized in that step S3 specifically performs posture correction _{on the dental three-dimensional model point cloud data Ps} through a three-dimensional model space transformation formula. 6. A method for aligning a dental plaque fluorescence image with a dental three-dimensional model according to claim 5, characterized in that the posture correction of the three-dimensional model point cloud data Ps _in step S3 is performed by a PCA principal component analysis method. 7. A method for aligning a dental plaque fluorescence image with a dental three-dimensional model according to claim 1, characterized in that step S44 also includes obtaining the contour in the dental three-dimensional model, and operating the edges of the fluorescent image to obtain the projection image contour points and the three-dimensional space points corresponding to the contour points, and saving all data. 8 . The method for registering a dental plaque fluorescence image with a dental three-dimensional model according to claim 1 , wherein in step S46 , the extrinsic parameter matrix M ₂ is updated by using the least square method through the obtained corresponding point information and the imaging model.

Images