CN104546151A - Numerical control laser automatic tooth preparation method and equipment thereof - Google Patents

Abstract

The invention relates to a numerical control laser automatic tooth preparation method and equipment thereof. The equipment contains an intraoral spatial digitizer, a dental laser, a numerical control laser tooth preparation control system intraoral working end, an oromaxillo-facial region cone-beam CT machine, a computer, a tooth fixator and a negative pressure aspirator. The computer is respectively connected with the intraoral spatial digitizer, the dental laser, the oromaxillo-facial region cone-beam CT machine and the negative pressure aspirator. The dental laser is connected with the numerical control laser tooth preparation control system intraoral working end which is connected with the tooth fixator. The equipment can replace part of manual operation of a doctor. A traditional mechanical grinding apparatus is replaced with laser. Clinical therapy operative level of grass-roots doctors can be effectively raised within a short period of time, and diagnosis and treatment efficiency and quality can be improved.

Description

CNC laser automatic tooth preparation method and equipment

technical field

The invention relates to a tooth body preparation method, in particular to a numerically controlled laser automated tooth body preparation method, and also relates to a numerically controlled laser automated tooth body preparation equipment. the

Background technique

1. Manual tooth preparation techniques currently used clinically at home and abroad. the

The traditional manual tooth preparation mode is difficult to meet the standard requirements in textbooks and clinical practice specifications. The level of manual tooth preparation in my country is generally low (according to the statistics of relevant experts, the pass rate accounts for about 40%). In my country, the number of high-level dentists is relatively scarce. At the same time, it often takes many years to train a high-level clinician. The above factors together lead to "to see the teeth is expensive, but difficult to see the teeth". In addition, therefore, there is an urgent need to develop a new, automated and intelligent clinical tooth preparation technology to replace the traditional manual mode. the

The traditional manual tooth preparation mode is difficult to meet the standard requirements in textbooks and clinical practice specifications. The level of manual tooth preparation in my country is generally low (according to experts' estimates, the pass rate accounts for about 40%). In my country, the number of high-level dentists is relatively scarce. At the same time, it often takes many years to train a high-level clinician. The above factors together lead to "to see the teeth is expensive, but difficult to see the teeth". In addition, therefore, there is an urgent need to develop a new, automated and intelligent clinical tooth preparation technology to replace the traditional manual mode. the

2. Dental hard tissue laser cutting technology

Some commercial Er:YAG and Er:YSGG lasers have been used for simple hand-held laser tooth preparation such as tooth decay and cavity preparation, but the tooth surface after Er laser cutting is rough and may be accompanied by microcracks. It is difficult to meet the high-precision requirements of dental restoration tooth preparation. Laser ablation of dental hard tissues has the characteristics of high precision, concentrated action, and small thermal damage. It can be ablated at a very low energy density, and its ablation threshold for enamel and dentin is 0.6-2.2, 0.3-1.4J/cm ^2. It is expected to become a tool for high-precision CNC tooth preparation. However, it has been reported in the literature that the cutting rate of laser on enamel and dentin is (0.05～3.6)×10 ^-3 , (0.12～1.90)×10-3mm ³ /s, which is lower than that of high-speed turbine handpiece (about 1mm ³ /s ). At the same time, unlike the traditional mechanical grinding head, the laser has no force feedback during operation, which is not conducive to the operator to feel the position, direction and attitude of the cutting tool, and is not conducive to controlling the accuracy of the cutting path.

3. About automatic control technology of laser light path in narrow space

At present, some numerical control laser scanning optical path control technologies have appeared abroad. For example, the three-dimensional laser processing software TbPs400 of the three-dimensional laser processing equipment of TRUMPF, a large German laser equipment manufacturer, is only effective for cutting round pipes and rectangular pipes. Strictly speaking, this is a 2.5-dimensional laser processing. However, the 3D laser processing software FORMA provided by PRIMAINUSTRIE S.P.A, a famous Italian laser equipment manufacturer, can only be used on workstations like IBMRISC Systern16000. The PEPSPentacut3D cutting system of the British CAMTEK company not only shortens the construction period, but also provides higher precision requirements and brings a huge competitive advantage. However, this type of software is expensive and does not have a popular market. At the 11th Beijing International Machine Tool Show (CIMT2009) in April 2009, more than 20 laser equipment manufacturers participating in the exhibition (including Han's Laser, Wuhan Falilai, Unity Bystronic, Unity Prima, etc.) Well-known laser equipment manufacturers), found that the exhibitors exhibited basically 2D laser cutting machine tools, only Shanghai Unity Prima exhibited the 3D laser cutting machine model SESAMO2545, but its matching 3D laser cutting machine The laser cutting automatic programming software adopts PEPS PentaCut developed by British Camtek Company, which shows that the domestic research on 3D laser cutting equipment, especially the automatic programming system is not very mature, which restricts the promotion of 3D laser cutting technology. the

For tooth preparation in such a narrow space in the oral cavity, there are no related reports to realize the automatic control of the laser light path. the

Contents of the invention

The purpose of the present invention is to provide a numerically controlled laser automatic tooth preparation method, which can replace part of the manual operation of doctors, replace traditional mechanical grinding instruments with laser, and can effectively improve the clinical treatment operation skills of grassroots doctors in a short time, and improve Efficiency and quality of care. The purpose of the present invention is to provide a numerically controlled laser automatic tooth body preparation equipment. the

In order to achieve the above object, the present invention has the following technical solutions:

A kind of numerical control laser automatic tooth body preparation method of the present invention, comprises the following steps:

1) Use an intraoral 3D scanner to obtain the 3D surface scanning data of the crown of the target tooth, use a cone beam CT machine in the mouth and maxillofacial region to obtain the 3D data of the crown of the target tooth, and store the above two sets of data separately for subsequent reading operations ;

2) Use the tooth preparation CAD software to register the above two sets of data, and unify them in the same coordinate system, extract the edge of the preparation on the computer screen, define the design parameters of the tooth preparation, and segment the enamel, dentin and pulp Cavity, complete the virtual modeling of the tooth preparation, and finally store the result as STL format data;

3) According to the tooth preparation virtual model data obtained in step 2), the laser tooth preparation CAM software automatically generates the laser focus spot diameter, spot movement path, and cutting process related parameters of speed during the tooth preparation process, and outputs them to the CNC laser In the tooth preparation control system;

4) Fix the tooth position locator on the target tooth and the mesial and distal adjacent teeth with silicone rubber, remove the silicone rubber around the crown of the target tooth, and obtain the overall three-dimensional image of the tooth position locator and the target tooth again with an intraoral 3D scanner Data, use the software in step 2) to register the virtual model data of the preparation with the overall scan data, and through the positioning column on the tooth position locator, connect the intraoral working end of the CNC laser tooth preparation control system to the tooth position locator The opening end of the occlusal surface is rigidly connected to realize the uniform fixation of the spatial position relationship of the target tooth crown, tooth position locator and the working end of the oral cavity of the CNC laser tooth preparation control system;

5) In the coordinate system formed in step 2), the laser tooth preparation CAM software is used to accurately focus the laser on the initial position of the target tooth, and the laser spot is controlled to automatically complete the tooth preparation process according to the set scanning path and scanning speed. Install a negative pressure suction device on the tooth position locator to remove tooth debris in real time;

6) Complete the tooth preparation, and sequentially remove the working end of the CNC laser tooth preparation control system and the tooth position locator from the patient's mouth. the

Wherein, the tooth preparation CAD software includes a data reading module, a preprocessing module, a data fusion module, a constraint modeling module, and a postprocessing module. the

in:

The data reading module includes:

1) CT data reconstruction: Read the 3D data of CT image reconstruction into the tooth preparation CAD software for illumination and rendering. the

2) Scanning data reconstruction: Obtain 3D surface scanning data and 3D volume data obtained by CT machine by scanning the target teeth, and read them into the software for storage to facilitate subsequent reading operations;

The preprocessing module includes:

1) Acquisition of fusion area surfaces: Firstly, the 3D surface scanning data model of the target teeth is divided into areas, which are divided into fusion areas and non-fusion areas. The point set in the non-fusion area is fixed, and the cervical margin line extraction based on a heuristic search strategy is adopted. The algorithm performs fusion surface clipping and extraction on the 3D surface scanning data model;

2) Initial registration based on manual interaction: According to the principle of three-point relocation, three points can establish a coordinate relationship, pick up no less than three corresponding feature points on the model obtained by the CT machine and the intraoral 3D scanner, and pass the feature points Alignment realizes the alignment of the model, respectively calibrates the feature points s _{i and s} ′ i on the model acquired by the CT machine and the intraoral 3D scanner, calculates the geometric transformation matrix R, T between the feature points by s′ _i = _{s i R} +T, and sets The matrix acts on the model acquired by the CT machine and the intraoral 3D scanner for initial transformation. This method can greatly reduce the translation error and rotation error between subsequent ICP model registration, and provide a good initial value for accurate registration;

3) Precise registration based on the ICP algorithm: through the initial registration, the ICP algorithm is used for precise registration, and the position registration between the models is between the model obtained by scanning the intraoral 3D scanner and the coordinate system of the 3D volume data model obtained by the CT machine Rotation and translation to minimize the distance between homologous points. By calculating the optimal rotation matrix R ^k and translation vector T ^k of the k-th iteration, the CT model N is approximated by the least squares method of the scanning model M after spatial changes The objective function reaches f(R,T) to a minimum:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; Rv</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&Right\; Arrow;</span>\mathrm{<span\; class="notranslate">\; min</span>}$

The data fusion module includes:

1) Establishment of differential coordinates: introduce the Laplacian coefficient matrix L to ensure that the conversion between Cartesian coordinates and differential coordinates can be performed in the form of matrix multiplication: Δ=LV, L=ID ^-1 A

Where Δ={δ _i } is the grid differential coordinate, D is a diagonal matrix, where D _ii =d _i , and A is the adjacency matrix of the grid;

2) Constraint condition establishment: after the deformed surface N of the 3D volume data model acquired by the CT machine is registered with the fixed surface M of the model scanned by the intraoral 3D scanner, the scanning model M is realized by combining the grid saliency with Morse theory On the extraction of prominent feature points, the feature points extracted by searching M correspond to the nearest point in N as a constraint;

3) Iterative processing of fusion deformation:

In the process of deforming the deformed surface N to the fixed surface M, a single deformation is likely to cause mesh distortion. Therefore, a strategy of multiple iterative deformations is adopted, which decomposes a single deformation into multiple operations. The iterative strategy not only makes the deformed The grid is naturally gentle, and the self-intersection phenomenon of the grid is avoided. The amount of deformation b each time is:

$\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1</span>}<span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}$

4) Deformation weight coefficient design: Quantify and evaluate the proximity between the deformation point and the target point by designing the weight matrix, through the following definition:

w _i =w _d (k,d)×w _a (α)

Among them, w _d (k, d) is the distance weight function, k is the current iteration number, d is the distance between the current deformation point g _i and the target deformation point v _i , w _a (α) is the angle weight function, and α is The angle between the normal vector direction of the deformation point g _i and the ray direction l _i from the center of gravity of the scanning model to the deformation point;

5) Model grid reconstruction: introduce the above weight matrix W into the linear equation and energy function, and get:

$\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; L</span>}\\ {\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; m</span>}}\end{array}\right){\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; L</span>}}^{<span\; class="notranslate">\; \&prime;</span>}{\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; \&delta;</span>}\\ \mathrm{<span\; class="notranslate">\; wxya</span>}\end{array}\right)$

$\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; L</span>}}^{<span\; class="notranslate">\; \&prime;</span>}{\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&delta;</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; C</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; w</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}$

When the energy function is minimized, in order to achieve the expected deformation position of the deformed grid, the pre-computation acceleration solution based on Cholesky decomposition is adopted in the solution process, and x′, y′, z′ are obtained respectively;

The constraint modeling modules include:

1) Axis extraction of spatial point cloud data: use the optimization method of spatial straight line fitting to extract

Take the long axis of the tooth, set the direction vector in space as S=(m,n,p) and the equation of the line passing through the point (x ₀ ,y ₀ ,z ₀ ) is

$\frac{\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; p</span>}}$

Wherein x ₀ , y ₀ , z ₀ are the three-dimensional coordinate values of passing points, m, n, p are the coordinate representation of the direction vector,

According to the principle of best square approximation: the error equation is $f(S_{t1},S_{t2},S_{t3})={\mathrm{\&Sigma;}}_{i=0}^R({S_{t1}}^2+{S_{t2}}^2{+S_{t3}}^2)$ , where, S _t1 , S _t2 , S _t3 are the error components of the points in the X-axis, Y-axis, and Z-axis respectively, and N is the number of points,

Using the optimal gradient method to solve this ternary quadratic nonlinear equation, the optimal direction vector (m, n, p) can be obtained. Using this direction vector and the known midpoint, a straight line can be obtained, which is the tooth axis;

2) Spatial curve projection: Use the line segment encryption projection method to discretize the cervical margin line into a data point set, project point by point along the curvature direction onto the three-dimensional volume data model obtained by the CT machine, and connect each projection point to form a projection curve;

3) Discrete model bias: Use the point-based bias algorithm to bias the model. The multi-vectors on the grid vertices are classified by type and given different weights. The calculation equation for the bias direction is as follows:

${\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; V</span>}}}_{\mathrm{<span\; class="notranslate">\; Offset</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>{\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; N</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}$

Among them, V _offset is the offset direction of the point, 1...n is the number of triangles in a circle of vertices, W _j is different weight values of different types of triangles, and N _ij is the normal vector corresponding to a circle of triangles in a vertex. The model after the offset point is obtained at each point normal vector direction offset point;

4) Analysis and solution of the model constraint equations: by analyzing the size chain relationship of the constraint model, list the model constraint equations:

$\left\{\begin{array}{c}\frac{\mathrm{<span\; class="notranslate">\; L</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}\\ \mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}\\ \mathrm{<span\; class="notranslate">\; the\; tan</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}}{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}\end{array}\right.$

Among them, L ₁ , L ₂ , and α are three parameterized variable values. For a specified tooth, H, L, and L4 are fixed values. When the parameter values are given, the equation system is an equation system containing three unknowns. The three known equations can uniquely determine a set of values of L _x , Ly _and L ₃ , so that the preparation model can be uniquely determined. By solving this constraint equation, the spatial position of each part of the model can be determined;

5) Parametric modeling of multi-constraint model: use the method of parametric operation to drive the modeling of the preparation, and use the method based on history to realize the parametric dynamic modification of the model;

The post-processing module includes:

Check the satisfaction of the constraints: By calculating the angle and size constraints of each section of the generated preparation, detect the size of the error of the generated preparation model, and use the color cloud image to visually display the distribution of the error. the

Among them, the history-based method is to record the parameterized operations and parameter values together in the order of model construction, form a model construction tree, and attach a mark to each operation. When the parameter size is modified, find the operation corresponding to the mark And start to rebuild the model with new parameter values according to the construction history, and complete the update of the new model. the

A numerically controlled laser automatic tooth preparation equipment of the present invention includes an intraoral three-dimensional scanner, a dental laser, a working end of the numerically controlled laser tooth preparation control system in the oral cavity, an oral and maxillofacial cone beam CT machine, a computer, a tooth fixator, Negative pressure suction device, the computer is respectively connected with intraoral three-dimensional scanner, dental laser, oral and maxillofacial cone beam CT machine, negative pressure suction device, and dental laser is connected with the working end of the oral cavity of the numerical control laser tooth preparation control system, The working end in the oral cavity of the numerical control laser tooth preparation control system is connected with the tooth fixer. the

Wherein, the working end of the oral cavity of the numerical control laser tooth preparation control system includes a light guide arm, a reflector cover, a positioner interface, a base, a motor base, a swing motor 1, a swing motor 2, a double vibrating mirror system, a linear motor, a focusing Lens seat, linear guide rail, grating sensor, the light guide arm is fixed on the left side of the base, the reflector cover is located at the end of the light guide arm, the locator interface is located under the reflector cover, and the linear guide rail is located on the base, The focusing lens seat is located on the linear guide rail, the grating sensor is located under the focusing lens seat, the linear motor is fixed on the base, the swing motor 1 and the swing motor 2 are fixed on the motor base, the double vibrating mirror system is connected with the swing motor 1 and the swing motor Two connections. the

Wherein, the double vibrating mirror system includes a vibrating mirror one, a vibrating mirror two, a focusing lens, and a reflecting mirror, the vibrating mirror one is located under the vibrating mirror two, the focusing lens is located between the vibrating mirror two and the reflecting mirror, and the vibrating mirror one is positioned under the vibrating mirror two. The vibrating mirror 2 is driven to rotate by the swing motor 1 and the swing motor 2 respectively, and the focusing lens is driven by a linear motor. the

Owing to taking above technical scheme, the advantage of the present invention is:

The invention can replace part of the manual operation of doctors, replace traditional mechanical grinding instruments with laser, effectively improve the clinical treatment operation technical level of grass-roots doctors in a short time, and improve the efficiency and quality of diagnosis and treatment. the

Description of drawings

Fig. 1 is the flowchart of numerical control laser automatic tooth preparation method of the present invention;

Fig. 2 is a schematic diagram of the restraint of the preparation body model of the present invention;

Fig. 3 is the schematic block diagram of numerical control laser automatic tooth body preparation equipment of the present invention;

Fig. 4 is the schematic diagram of the working end in the oral cavity of the numerical control laser tooth preparation control system of the present invention;

Fig. 5 is a schematic diagram of the dual vibrating mirror system of the present invention. the

In the figure: 1. Light guide arm; 2. Reflector cover; 3. Positioner interface; 4. Lens mount; 5. Grating sensor; 6. Linear guide rail; 7. Linear motor; 8. Base; 9. Double lens System; 10. Swing motor one; 11. Swing motor two; 12. Motor base; 13. Vibrating mirror one; 14. Vibrating mirror two; 15. Focusing lens; 16. Mirror; 17. Target teeth; 18. Laser beam . the

Detailed ways

The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention. the

A kind of numerical control laser automatic tooth body preparation method of the present invention of the present invention is made up of the following steps:

1) Use an intraoral 3D scanner to obtain the 3D surface scanning data of the crown of the target tooth, use a cone beam CT machine in the mouth and maxillofacial region to obtain the 3D data of the crown of the target tooth, and store the above two sets of data separately for subsequent reading operations ;

2) Use the tooth preparation CAD software to register the above two sets of data, and unify them in the same coordinate system, extract the edge of the preparation on the computer screen, define the design parameters of the tooth preparation, and segment the enamel, dentin and pulp Cavity, complete the virtual modeling of the tooth preparation, and finally store the result as STL format data;

3) According to the virtual model data of the tooth preparation obtained in step 2), the laser tooth preparation CAM software automatically generates cutting process-related parameters such as the focused spot diameter, spot movement path, and speed of the laser during the tooth preparation process, and outputs them to the CNC laser In the tooth preparation control system;

The above-mentioned laser tooth preparation adopts the method of layer-by-layer removal to obtain the preparation body. The whole path planning includes two links, one is slice layering, and the other is generating an efficient laser cutting path for the cross-sectional outline obtained by slice layering; 

Fast slicing and layering link: firstly, group the triangular faces in the 3D model of tooth STL format data, and divide the triangular faces into several groups according to the different intersecting tangent planes, and the triangular faces in each group are equal to Intersect with the tangent plane; then, intersect the triangles in each group with the tangent plane, each triangle intersects with the tangent plane to get an intersecting line segment, according to the triangles in the 3D model of STL format data Continuity, which can form an undirected and closed cross-sectional contour line, without the need to reconstruct the topology of the triangular patch or sort the obtained intersection segments;

High-efficiency laser cutting path generation of cross-sectional contours: firstly, use the straight line filling method to solve the intersection line segment between the laser cutting area and the filling line, calculate the laser cutting line, and then sort and divide these cutting lines to divide the complex cutting area into Small monotonous cutting areas, resulting in efficient laser cutting paths;

4) Fix the tooth position locator on the target tooth and the mesial and distal adjacent teeth with silicone rubber, and remove the silicone rubber around the crown of the target tooth. Use the intraoral 3D scanner to obtain the overall 3D data of the tooth position locator and the target tooth again, and use the software in step 2) to register the virtual model data of the preparation with the overall scan data. Through the positioning column on the tooth position locator, the intraoral working end of the CNC laser tooth preparation control system is rigidly connected to the occlusal surface opening end of the tooth position locator, so as to realize the crown of the target tooth, the tooth position locator and the numerical control Unified fixation of the spatial position relationship of the working end of the laser tooth preparation control system in the oral cavity;

5) In the coordinate system formed in step 2), the laser tooth preparation CAM software is used to accurately focus the laser on the initial position of the target tooth, and the laser spot is controlled to automatically complete the tooth preparation process according to the set scanning path and scanning speed. Install a negative pressure suction device on the tooth position locator to remove tooth debris in real time;

Laser tooth preparation CAM software works as follows:

Autofocus at the initial position: First, the lens position is initialized to make the lens return to the initial zero position; then, the highest point coordinate value of the target tooth (the maximum Z value of the target tooth) can be obtained through the target Yahoo STL model; finally, according to the obtained The highest coordinates of the tooth and the parameters of the laser light path can calculate the final moving distance of the lens, and finally the focal position of the laser can be accurately located at the highest point of the target tooth;

Automatic process control of tooth preparation: In laser tooth preparation path planning, tooth preparation is performed at a uniform speed by straight line cutting. The cutting path of the laser is composed of laser cutting line segments of different lengths. To achieve uniform laser cutting, it is necessary to interpolate laser cutting line segments, that is, to divide laser scanning line segments of different lengths into small-pitch laser cutting points, and convert the three-dimensional coordinates of small-pitch laser cutting points into laser cutting points through kinematic inverse operations. The angular displacement of galvanometer 1 and galvanometer 2 in the working head and the linear displacement of the lens. Since both the vibrating mirror and the lens are directly driven, the simultaneous movement of the vibrating mirror 1 motor, the vibrating mirror 2 motor and the linear motor of the lens can control the laser spot to track the interpolated laser cutting point, and finally realize the laser along the planned cutting path at a uniform speed Realize cutting and complete the automatic tooth preparation process;

6) Complete the tooth preparation, and sequentially remove the working end of the CNC laser tooth preparation control system and the tooth position locator from the patient's mouth. the

The tooth preparation CAD software includes a data reading module, a preprocessing module, a data fusion module, a constraint modeling module and a postprocessing module. see picture 1. the

in:

The data reading module includes:

1) CT data reconstruction: Read the 3D data of CT image reconstruction into the tooth preparation CAD software for illumination and rendering. the

2) Scanning data reconstruction: Obtain 3D surface scanning data and 3D volume data obtained by CT machine by scanning the target teeth, and read them into the software for storage to facilitate subsequent reading operations;

The preprocessing module includes:

1) Acquisition of fusion area surface: Firstly, the three-dimensional surface scanning data model of the scanning target tooth is divided into areas, which are divided into fusion area and non-fusion area, and the point set of the non-fusion area is fixed. The neck edge line extraction algorithm based on the heuristic search strategy is used to extract the 3D surface scanning data model by fused surface clipping. The heuristic search strategy is to obtain the optimal search position by evaluating the cost of each target to be searched, and then from this search to the final target, thereby omitting a large number of unnecessary paths and improving search efficiency;

2) Initial registration based on manual interaction: According to the principle of three-point relocation, three points can establish a coordinate relationship, pick up no less than three corresponding feature points on the model obtained by the CT machine and the intraoral 3D scanner, and pass the feature points Alignment implements the alignment of the model. The feature points s _i and s′ i are respectively calibrated on the models acquired by the CT machine and the intraoral three-dimensional scanner, and the geometric _{transformation matrix R and T between the feature points are calculated by s′ i =} s _i R+T, and the matrix is applied to the CT machine and The model acquired by the intraoral 3D scanner is initially transformed. This method can greatly reduce the translation error and rotation error between subsequent ICP model registrations, and provide a good initial value for accurate registration;

3) Precise registration based on the ICP algorithm: through the initial registration, the ICP algorithm is used for precise registration, and the position registration between the models is between the model obtained by scanning the intraoral 3D scanner and the coordinate system of the 3D volume data model obtained by the CT machine Rotation and translation to minimize the distance between homologous points. By calculating the optimal rotation matrix R ^k and translation vector T ^k of the k-th iteration, the CT model N is approximated by the least squares method of the scanning model M after spatial changes The objective function reaches f(R,T) to a minimum:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; Rv</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&Right\; Arrow;</span>\mathrm{<span\; class="notranslate">\; min</span>}$

The data fusion module includes:

1) Establishment of differential coordinates: Introduce the Laplacian coefficient matrix L to ensure that the conversion between Cartesian coordinates and differential coordinates can be performed in the form of matrix multiplication:

Δ=LV, L=ID ^-1 A

Where Δ={δ _i } is the grid differential coordinate, D is a diagonal matrix, where D _ii =d _i , and A is the adjacency matrix of the grid;

2) Constraint condition establishment: after the deformation surface N of the 3D volume data model obtained by the CT machine is registered with the fixed surface M of the model scanned by the intraoral 3D scanner, the scanning model M is realized by combining the grid saliency with Morse theory On the extraction of prominent feature points, the feature points extracted by searching M correspond to the nearest point in N as a constraint;

3) Iterative processing of fusion deformation:

In the process of deforming the deformed grid N to the fixed grid M, a single deformation is likely to cause grid distortion. Therefore, a strategy of multiple iterative deformations is adopted, that is, a single deformation is decomposed into multiple operations. The iterative strategy not only makes After the deformation, the grid is naturally smooth, and the self-intersection phenomenon of the grid is avoided. The amount of each deformation b is:

$\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1</span>}<span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}$

4) Deformation weight coefficient design: Quantify and evaluate the proximity between the deformation point and the target point by designing the weight matrix, through the following definition:

w _i =w _d (k,d)×w _a (α)

Among them, w _d (k, d) is the distance weight function, k is the current iteration number, d is the distance between the current deformation point g _i and the target deformation point v _i , w _a (α) is the angle weight function, and α is The angle between the normal vector direction of the deformation point g _i and the ray direction l _i from the center of gravity of the scanning model to the deformation point;

5) Model grid reconstruction: introduce the above weight matrix W into the linear equation and energy function, and get:

$\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; L</span>}\\ {\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; m</span>}}\end{array}\right){\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; L</span>}}^{<span\; class="notranslate">\; \&prime;</span>}{\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; \&delta;</span>}\\ \mathrm{<span\; class="notranslate">\; wxya</span>}\end{array}\right)$

$\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; L</span>}}^{<span\; class="notranslate">\; \&prime;</span>}{\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&delta;</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; C</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; w</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}$

When the energy function is minimized, in order to achieve the expected deformation position of the deformed grid, the pre-computation acceleration solution based on Cholesky decomposition is adopted in the solution process, and x′, y′, z′ are obtained respectively;

The constraint modeling modules include:

1) Axis extraction of spatial point cloud data: use the optimization method of spatial straight line fitting to extract

Take the long axis of the tooth, set the direction vector in space as S=(m,n,p) and the equation of the line passing through the point (x ₀ ,y ₀ ,z ₀ ) is

$\frac{\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; p</span>}}$

Among them, x ₀ , y ₀ , z ₀ are the three-dimensional coordinate values of passing points, m, n, p are the coordinate representation of the direction vector,

According to the best square approximation principle, the error equation is $f(S_{t1},S_{t2},S_{t3})={\mathrm{\&Sigma;}}_{i=0}^R({S_{t1}}^2+{S_{t2}}^2{+S_{t3}}^2)$ , where, S _t1 , S _t2 , ·S _t3 are the error components of the points in the X-axis, Y-axis, and Z-axis respectively, and N is the number of points,

Using the optimal gradient method to solve this ternary quadratic nonlinear equation, the optimal direction vector (m, n, p) can be obtained. Using this direction vector and the known midpoint, a straight line can be obtained, which is the tooth axis;

2) Spatial curve projection: use the line segment encryption projection method to discretize the cervical margin line into a data point set, project point by point along the curvature direction onto the three-dimensional volume data model obtained by the CT machine, and connect each projection point to form a projection curve;

3) Discrete model bias: Use the point-based bias algorithm to bias the model. The multi-vectors on the grid vertices are classified by type and given different weights. The calculation equation for the bias direction is as follows:

${\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; V</span>}}}_{\mathrm{<span\; class="notranslate">\; Offset</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; \&CenterDot;</span>{\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; N</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}$

Among them, V _offset is the offset direction of the point, 1...n is the number of triangles in a circle of vertices, W _j is different weight values of different types of triangles, and N _ij is the normal vector corresponding to a circle of triangles in a vertex. The model after the offset point is obtained at each point normal vector direction offset point;

4) Analysis and solution of the model constraint equations: by analyzing the size chain relationship of the constraint model, list the model constraint equations:

$\left\{\begin{array}{c}\frac{\mathrm{<span\; class="notranslate">\; L</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}\\ \mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}\\ \mathrm{<span\; class="notranslate">\; the\; tan</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}}{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}\end{array}\right.$

Among them, L ₁ , L ₂ , and α are three parameterized variable values. For a specified tooth, H, L, and L4 are fixed values. When the parameter values are given, the equation system is an equation system containing three unknowns. The three known equations can uniquely determine a set of values of L _x , Ly _and L ₃ , so that the preparation model can be uniquely determined. By solving this constraint equation, the spatial position of each part of the model can be determined, as shown in Figure 2;

5) Parametric modeling of multi-constraint model: use the method of parametric operation to drive the modeling of the preparation, and use the method based on history to realize the parametric dynamic modification of the model;

The post-processing module includes:

Check the satisfaction of the constraints: By calculating the angle and size constraints of each section of the generated preparation, detect the size of the error of the generated preparation model, and use the color cloud image to visually display the distribution of the error. the

Among them, the history-based method is to record the parameterized operations and parameter values together in the order of model construction, form a model construction tree, and attach a mark to each operation. When the parameter size is modified, find the operation corresponding to the mark And start to rebuild the model with new parameter values according to the construction history, and complete the update of the new model. the

See Fig. 3, a kind of numerical control laser automatic tooth preparation equipment of the present invention, comprises intraoral three-dimensional scanner, dental laser, numerical control laser tooth preparation control system intraoral working end, mouth and maxillofacial cone beam CT machine, computer, Tooth fixator, negative pressure suction device, the computer is respectively connected with intraoral three-dimensional scanner, dental laser, oral and maxillofacial cone beam CT machine, negative pressure suction device, dental laser and numerical control laser tooth preparation control system in oral cavity The working end is connected, and the working end in the oral cavity of the CNC laser tooth preparation control system is connected with the tooth retainer. the

As shown in Figure 4, the oral working end of the numerical control laser tooth preparation control system includes a light guide arm, a reflector cover, a positioner interface, a base, a motor seat, a swing motor 1, a swing motor 2, a double vibrating mirror system, and a linear motor , focusing lens seat, linear guide rail, grating sensor, the light guide arm is fixed on the left side of the base, the reflector cover is located at the end of the light guide arm, the locator interface is located under the reflector cover, and the linear guide rail is located on the base The focus lens seat is located on the linear guide rail, the grating sensor is located under the focus lens seat, the linear motor is fixed on the base, the swing motor 1 and swing motor 2 are fixed on the motor base, and the double vibrating mirror system and swing motor 1 and 2 are fixed on the motor base. The swing motor two is connected. the

See Fig. 5, described double vibrating mirror system comprises vibrating mirror one, vibrating mirror two, focus lens, mirror, described vibrating mirror one is positioned at the below of vibrating mirror two, focusing lens is positioned between vibrating mirror two and reflecting mirror, vibrating mirror Mirror one and vibrating mirror two are respectively driven to rotate by swing motor one and swing motor two, and the focusing lens is driven by a linear motor. the

Apparently, the above-mentioned implementation of the present invention is only an example for clearly illustrating the present invention, rather than limiting the implementation of the present invention. For those of ordinary skill in the art, other changes or changes in different forms can be made on the basis of the above description. All the implementation manners cannot be exhaustively listed here. All obvious changes or changes that are derived from the technical solutions of the present invention are still within the protection scope of the present invention. the

Claims (7)

1. A numerically controlled laser automatic tooth body preparation method is characterized in that comprising the following steps: 1) Use an intraoral 3D scanner to obtain the 3D surface scanning data of the crown of the target tooth, use a cone beam CT machine in the mouth and maxillofacial region to obtain the 3D data of the crown of the target tooth, and store the above two sets of data separately for subsequent reading operations ; 2) Use the tooth preparation CAD software to register the above two sets of data, and unify them in the same coordinate system, extract the edge of the preparation on the computer screen, define the design parameters of the tooth preparation, and segment the enamel, dentin and pulp Cavity, complete the virtual modeling of the tooth preparation, and finally store the result as STL format data; 3) According to the tooth preparation virtual model data obtained in step 2), the laser tooth preparation CAM software automatically generates the laser focus spot diameter, spot movement path, and cutting process related parameters of speed during the tooth preparation process, and outputs them to the CNC laser In the tooth preparation control system; 4) Fix the tooth position locator on the target tooth and the mesial and distal adjacent teeth with silicone rubber, remove the silicone rubber around the crown of the target tooth, and obtain the overall three-dimensional image of the tooth position locator and the target tooth again with an intraoral 3D scanner Data, use the software in step 2) to register the virtual model data of the preparation with the overall scan data, and through the positioning column on the tooth position locator, connect the intraoral working end of the CNC laser tooth preparation control system to the tooth position locator The opening end of the occlusal surface of the occlusal surface is rigidly connected to realize the uniform fixation of the spatial position relationship of the target tooth crown, tooth position locator and the working end of the oral cavity of the CNC laser tooth preparation control system; 5) In the coordinate system formed in step 2), the laser tooth preparation CAM software is used to accurately focus the laser on the initial position of the target tooth, and the laser spot is controlled to automatically complete the tooth preparation process according to the set scanning path and scanning speed. Install a negative pressure suction device on the tooth position locator to remove tooth debris in real time; 6) Complete the tooth preparation, and sequentially remove the working end of the CNC laser tooth preparation control system and the tooth position locator from the patient's mouth. 2. A kind of numerical control laser automatic tooth body preparation method as claimed in claim 1, is characterized in that: described tooth body preparation CAD software comprises data reading module, preprocessing module, data fusion module, constraint modeling module, post-processing processing module. 3. A kind of numerical control laser automatic tooth body preparation method as claimed in claim 2, is characterized in that: The data reading module includes: 1) CT data reconstruction: Read the 3D data of CT image reconstruction into the tooth preparation CAD software for illumination and rendering. 2) Scanning data reconstruction: Obtain 3D surface scanning data and 3D volume data obtained by CT machine by scanning the target teeth, and read them into the software for storage to facilitate subsequent reading operations; The preprocessing module includes: 1) Acquisition of fusion area surfaces: Firstly, the 3D surface scanning data model of the target teeth is divided into areas, which are divided into fusion areas and non-fusion areas. The point set in the non-fusion area is fixed, and the cervical margin line extraction based on a heuristic search strategy is adopted. Algorithm performs fused surface clipping and extraction on the 3D surface scanning data model; 2) Initial registration based on manual interaction: According to the principle of three-point relocation, three points can establish a coordinate relationship, pick up no less than three corresponding feature points on the model obtained by the CT machine and the intraoral 3D scanner, and pass the feature points Alignment realizes the alignment of the model, respectively calibrates the feature points s _{i and s} ′ i on the model acquired by the CT machine and the intraoral 3D scanner, calculates the geometric transformation matrix R, T between the feature points by s′ _i = _{s i R} +T, and sets The matrix acts on the model acquired by the CT machine and the intraoral 3D scanner for initial transformation. This method can greatly reduce the translation error and rotation error between the closest point model registration of subsequent iterations, and provide a good initial value for accurate registration; 3) Precise registration based on the ICP algorithm: through the initial registration, the ICP algorithm is used for precise registration, and the position registration between the models is between the model obtained by scanning the intraoral 3D scanner and the coordinate system of the 3D volume data model obtained by the CT machine Rotation and translation to minimize the distance between homologous points. By calculating the optimal rotation matrix R ^k and translation vector T ^k of the k-th iteration, the CT model N is approximated by the least squares method of the scanning model M after spatial changes The objective function reaches f(R,T) to a minimum: $\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; Rv</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&Right\; Arrow;</span>\mathrm{<span\; class="notranslate">\; min</span>}$ The data fusion module includes: 1) Establishment of differential coordinates: introduce the Laplacian coefficient matrix L to ensure that the conversion between Cartesian coordinates and differential coordinates can be performed in the form of matrix multiplication: Δ=LV, L=ID ^-1 A Where Δ={δ _i } is the grid differential coordinate, D is a diagonal matrix, where D _ii =d _i , and A is the adjacency matrix of the grid; 2) Constraint condition establishment: after the deformed surface N of the 3D volume data model acquired by the CT machine is registered with the fixed surface M of the model scanned by the intraoral 3D scanner, the scanning model M is realized by combining the grid saliency with Morse theory The extraction of the salient feature points, the feature points extracted by searching M correspond to the nearest point in N as a constraint; 3) Fusion deformation iterative processing: In the process of deforming the deformed surface N to the fixed surface M, a single deformation is likely to cause mesh distortion. Therefore, a strategy of multiple iterative deformations is adopted, which decomposes a single deformation into multiple operations. The iterative strategy not only makes the deformed The grid is naturally smooth and avoids the grid self-intersection phenomenon. The amount of deformation b each time is: $\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1</span>}<span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}$ 4) Deformation weight coefficient design: Quantify and evaluate the proximity between the deformation point and the target point by designing the weight matrix, defined by the following: w _i =w _d (k,d)×w _a (α) Among them, w _d (k, d) is the distance weight function, k is the current iteration number, d is the distance between the current deformation point g _i and the target deformation point v _i , w _a (α) is the angle weight function, and α is The angle between the normal vector direction of the deformation point g _i and the ray direction l _i from the center of gravity of the scanning model to the deformation point; 5) Model grid reconstruction: introduce the above weight matrix W into the linear equation and energy function, and get: $\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; L</span>}\\ {\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; m</span>}}\end{array}\right){\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; L</span>}}^{<span\; class="notranslate">\; \&prime;</span>}{\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; \&delta;</span>}\\ \mathrm{<span\; class="notranslate">\; wxya</span>}\end{array}\right)$ $\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; L</span>}}^{<span\; class="notranslate">\; \&prime;</span>}{\mathrm{<span\; class="notranslate">\; V</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&delta;</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; C</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; w</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}$ When the energy function is minimized, in order to achieve the expected deformation position of the deformed grid, the pre-computation acceleration solution based on Cholesky decomposition is adopted in the solution process, and x′, y′, and z′ are obtained respectively; The constraint modeling module includes: 1) Spatial point cloud data axis extraction: use the optimization method of spatial straight line fitting to extract the tooth long axis of the tooth, set the direction vector in space as S=(m,n,p) and pass through the point (x ₀ ,y ₀ , z ₀ ) The equation of the straight line is: $\frac{\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; z</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{\mathrm{<span\; class="notranslate">\; p</span>}}$ Among them, x ₀ , y ₀ , z ₀ are the three-dimensional coordinate values of passing points, m, n, p are the coordinate representation of the direction vector, According to the best square approximation principle, the error equation is $f(S_{t1},S_{t2},S_{t3})={\mathrm{\&Sigma;}}_{i=0}^R({S_{t1}}^2+{S_{t2}}^2{+S_{t3}}^2)$ , where, S _t1 , S _t2 , S _t3 are the error components of the points in the X-axis, Y-axis, and Z-axis respectively, and N is the number of points, Using the optimal gradient method to solve this ternary quadratic nonlinear equation, the optimal direction vector (m, n, p) can be obtained. Using this direction vector and the known midpoint, a straight line can be obtained, which is the tooth axis; 2) Spatial curve projection: Use the line segment encryption projection method to discretize the cervical line into a data point set, project point by point along the curvature direction onto the three-dimensional data model obtained by the CT machine, and connect the projection points to form a projection curve; 3) Discrete model bias: use the point-based bias algorithm to bias the model, and the multi-vectors on the grid vertices are classified by type and given different weights. The calculation equation of the bias direction is as follows: ${\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; V</span>}}}_{\mathrm{<span\; class="notranslate">\; Offset</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>{\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; N</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}$ Among them, V _offset is the offset direction of the point, 1...n is the number of triangles in a circle of vertices, W _j is different weight values of different types of triangles, and N _ij is the normal vector corresponding to a circle of triangles in a vertex. The model after the offset point is obtained at each point normal vector direction offset point; 4) Analysis and solution of the model constraint equations: by analyzing the size chain relationship of the constraint model, list the model constraint equations: $\left\{\begin{array}{c}\frac{\mathrm{<span\; class="notranslate">\; L</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}\\ \mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}\\ \mathrm{<span\; class="notranslate">\; the\; tan</span>}\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}}{{\mathrm{<span\; class="notranslate">\; L</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}\end{array}\right.$ Among them, L ₁ , L ₂ , and α are three parameterized variable values. For a specified tooth, H, L, and L4 are fixed values. When the parameter values are given, the equation system is an equation system containing three unknowns. The three known equations can uniquely determine a set of values of L _x , Ly _and L ₃ , so that the preparation model can be uniquely determined. By solving this constraint equation, the spatial position of each part of the model can be determined; 5) Parametric modeling of multi-constraint model: use the method of parametric operation to drive the modeling of the preparation, and use the method based on history to realize the parametric dynamic modification of the model; The post-processing module includes: Check the satisfaction of the constraints: By calculating the angle and size constraints of each section of the generated preparation, detect the size of the error of the generated preparation model, and use the color cloud image to visually display the distribution of the error. 4. A kind of numerical control laser automatic tooth body preparation method as claimed in claim 3, it is characterized in that: described method based on history is that parametric operation and parameter value are recorded together according to model construction order, form model construction tree , and attach a mark to each operation. When the parameter size is modified, find the operation corresponding to the mark and start to reconstruct the model with new parameter values according to the construction history to complete the update of the new model. 5. A numerically controlled laser automatic tooth preparation equipment, characterized in that it includes an intraoral three-dimensional scanner, a dental laser, a numerically controlled laser tooth preparation control system, an oral cavity working end, an oral and maxillofacial cone beam CT machine, a computer, a dental Fixer, negative pressure suction device, the computer is respectively connected with intraoral three-dimensional scanner, dental laser, oral and maxillofacial cone beam CT machine, negative pressure suction device, and the dental laser works in the oral cavity with the numerical control laser tooth preparation control system The end is connected, and the working end of the CNC laser tooth preparation control system in the oral cavity is connected with the tooth retainer. 6. A CNC laser automatic tooth preparation equipment according to claim 5, characterized in that: the oral cavity working end of the CNC laser tooth preparation control system includes a light guide arm, a reflector cover, a locator interface, and a base , motor base, swing motor one, swing motor two, double vibrating mirror system, linear motor, focusing lens seat, linear guide rail, grating sensor, the light guide arm is fixed on the left side of the base, and the reflector cover is located on the light guide arm At the end, the locator interface is located under the reflector cover, the linear guide is located on the base, the focus lens seat is located on the linear guide, the grating sensor is located under the focus lens seat, the linear motor is fixed on the base, and the swing motor The second swing motor is fixed on the motor base, and the double vibrating mirror system is connected with the first swing motor and the second swing motor. 7. A numerically controlled laser automatic tooth preparation equipment as claimed in claim 6, characterized in that: the double oscillating mirror system includes vibrating mirror one, vibrating mirror two, focusing lens, and reflecting mirror, and the vibrating mirror one is located at Below the vibrating mirror 2, the focusing lens is located between the vibrating mirror 2 and the reflecting mirror. The vibrating mirror 1 and the vibrating mirror 2 are respectively driven to rotate by the swing motor 1 and the swing motor 2, and the focusing lens is driven by a linear motor.