Snapshot tooth optical parameter measurement and calculation method

Abstract

The invention discloses a snapshot type tooth optical parameter measurement and calculation method. The method comprises the following steps: controlling a snapshot type optical acquisition unit to synchronously acquire an original optical image data set containing multiple types of light components; pixel-level spectral features are extracted, and spectral response intervals of different tooth tissues are distinguished; dividing and analyzing sub-regions in combination with an anatomical structure, constructing an optical characteristic initial analysis matrix, and obtaining spectral attenuation characteristic data; matching with a preset sample library to obtain a preliminary optical parameter estimation value; and introducing ambient light and equipment imaging deviation factor dynamic correction, and constructing an overall parameter distribution model to output core optical parameters. According to the method, tooth contact is not needed, the efficiency is improved through synchronous imaging, the defects of fuzzy tissue differentiation and poor interference resistance of a traditional method are overcome through spectrum-anatomical structure combined analysis and a dynamic correction mechanism, output parameters are accurate and stable, and the method is suitable for oral clinical diagnosis and tooth restoration scenes.

Description

Technical Field

[0001] This invention relates to the field of dental optical parameter measurement technology, and in particular to a snapshot-type method for calculating dental optical parameters. Background Technology

[0002] Dental optical parameters, including core indicators such as refractive index, scattering coefficient, and absorption coefficient, play a crucial role in the accurate measurement of oral diseases, dental material selection, and dental restoration planning. As oral medical technology advances towards precision and non-invasiveness, traditional contact-based measurement methods, due to their potential to damage tooth tissue and complex procedures, are no longer sufficient to meet clinical needs. Snapshot optical imaging technology, with its advantage of simultaneous acquisition across multiple spectral bands, can rapidly capture the light reflection and scattering characteristics of the tooth surface and interior, providing rich raw data for optical parameter calculations, and has become a research hotspot in the field of dental optical measurement. Against this backdrop, there is an urgent need to develop a snapshot-based method for measuring and calculating dental optical parameters that balances measurement efficiency and data accuracy to meet the practical needs of clinical practice for rapid analysis of tooth characteristics.

[0003] Existing technologies for snapshot-based measurement and calculation of dental optical parameters suffer from two significant drawbacks. Firstly, the correlation between feature extraction and parameter matching is insufficient. Parameter estimation often relies solely on light intensity data from a single spectral range, failing to finely segment the measurement area based on dental anatomy. This results in inadequate differentiation of spectral attenuation characteristics across different tissue types, leading to potential parameter estimation biases. Secondly, the interference factor correction mechanism is incomplete. The impact of ambient light intensity fluctuations and equipment imaging biases is often ignored during parameter calculation. The lack of a dynamic correction model to calibrate preliminary estimates makes the measurement results highly susceptible to external environmental factors and equipment conditions, hindering the stable output of highly reliable optical parameter data. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of the existing technology, the present invention provides a snapshot-type method for measuring and calculating dental optical parameters.

[0005] The technical solution adopted in this invention is a snapshot-type method for measuring and calculating optical parameters of teeth, comprising: Step S1. Controlling a snapshot-type optical acquisition unit to perform multispectral band synchronous imaging of the target tooth region, acquiring a raw optical image dataset including diffuse reflection light, specular reflection light, and subsurface scattered light components from the tooth surface; Step S2. Extracting pixel-level spectral features from the raw optical image dataset, distinguishing the spectral response intervals corresponding to different tooth tissues, and determining the light intensity distribution data corresponding to the characteristic wavelengths within different intervals; Step S3. Constructing an initial analysis matrix of tooth optical properties based on the characteristic wavelength light intensity distribution data, combined with tooth analysis... Step S4: Divide the structure into multiple sub-regions for analysis and obtain spectral attenuation characteristic data for different sub-regions; Step S5: Call the preset tooth tissue optical parameter sample library, and perform feature matching between the spectral attenuation characteristic data of different sub-regions and the standard data in the sample library to determine the preliminary optical parameter estimates; Step S6: Introduce ambient light interference factors and equipment imaging deviation factors to dynamically correct the preliminary optical parameter estimates and generate corrected intermediate optical parameter values; Step S7: Construct an overall tooth optical parameter distribution model based on the intermediate optical parameter values ​​of different sub-regions, and output the core optical parameters of the target tooth, including refractive index, scattering coefficient, and absorption coefficient.

[0006] Furthermore, the following model is used when constructing the initial analysis matrix of tooth optical properties in step S3: Where M represents the initial optical characteristic analysis matrix, n represents the number of characteristic wavelengths, and m represents the number of pixels. This represents the weight coefficient of the j-th pixel under the i-th feature wavelength. This represents the light intensity value of the j-th pixel at the i-th characteristic wavelength. Indicates the reference value of incident light intensity. This represents the spectral attenuation index of the j-th pixel at the i-th characteristic wavelength. This represents the scattering influence coefficient of the j-th pixel at the i-th characteristic wavelength. This represents the light propagation path factor of the j-th pixel at the i-th characteristic wavelength.

[0007] Furthermore, the feature matching process in step S4 adopts the following model: Where S represents the feature matching degree, and p represents the dimension of the matching feature. This represents the spectral attenuation characteristics of the k-th dimension of the sub-region. This represents the k-th dimension of the standard data in the sample library. This represents the average value of the spectral attenuation characteristics data for the sub-region. This represents the average value of the standard data in the sample library. This represents the matching deviation adjustment factor.

[0008] Furthermore, the dynamic correction process in step S5 adopts the following model: ,in, This represents the intermediate value of the corrected optical parameters. This represents a preliminary estimate of optical parameters. This represents the environmental interference correction factor. Indicates the actual ambient light intensity. Indicates standard ambient light intensity. This represents the equipment deviation correction factor. This represents the actual imaging deviation value. This indicates the equipment calibration deviation threshold.

[0009] Furthermore, the overall tooth optical parameter distribution model constructed in step S6 adopts the following model: ,in, Representing three-dimensional spatial coordinates The optical parameter value at the location, where r represents the number of sub-regions analyzed. This represents the spatial weight of the q-th sub-region. This represents the intermediate value of the optical parameters of the q-th sub-region. This represents the spatial decay coefficient of the parameter in the q-th sub-region. Representing coordinates The distance to the center of the q-th sub-region.

[0010] Furthermore, the light intensity distribution data determined in step S2 uses the following model: ,in, Indicates wavelength ,coordinate The light intensity value at that location, Indicates the range of imaging solid angles. Indicates wavelength ,coordinate Location, angle Radiance in direction, Indicates the zenith angle. Indicates the azimuth angle.

[0011] Further, step S2 includes the following sub-steps: S21. Call the spectral analysis module to perform channel separation on the original optical image dataset, extract the wavelength information corresponding to each imaging channel, establish the mapping relationship between wavelength and image channel, and obtain a single-wavelength image sequence; S22. Perform pixel traversal on the single-wavelength image sequence, record the gray value of each pixel at different wavelengths, convert the gray value into the corresponding light intensity response value, and form a pixel-wavelength-light intensity three-dimensional data array; S23. Based on the tooth tissue spectral database, determine the characteristic spectral intervals corresponding to enamel, dentin, and cementum, divide the three-dimensional data array according to the characteristic spectral intervals, and obtain light intensity distribution subsets corresponding to different tissue types; S24. Remove edge pixels from different light intensity distribution subsets, retain the light intensity data in the effective tooth area, and organize them to form light intensity distribution data corresponding to the characteristic wavelength.

[0012] Further, step S3 includes the following sub-steps: S31. Based on the characteristic wavelength light intensity distribution data, construct an initial data matrix with row vectors representing characteristic wavelengths and column vectors representing pixels, and normalize the matrix elements to obtain a standardized data matrix; S32. Import a three-dimensional model of the tooth anatomy, project the model onto the image plane, divide the image regions corresponding to the crown, neck, and root, and determine the boundary pixel coordinates of different regions; S33. Divide the standardized data matrix into sub-matrices according to the boundary pixel coordinates to obtain sub-matrices corresponding to different anatomical regions, calculate the row mean and column variance of each sub-matrix, and obtain the basic data of spectral attenuation characteristics; S34. Combine the propagation characteristics of light in tooth tissue to perform gradient calculation on the basic data of spectral attenuation characteristics to obtain the spectral attenuation characteristic data of different sub-regions.

[0013] Further, step S4 includes the following sub-steps: S41. Access a preset dental tissue optical parameter sample library, which includes standard spectral attenuation data and optical parameter data corresponding to different ages, genders, and dental health states, and establish a sample index table; S42. Extract feature vectors from the spectral attenuation characteristic data of different sub-regions, including peak wavelength, attenuation slope, and full width at half maximum (FWHM) feature parameters, to form a feature set to be matched; S43. Use the K-nearest neighbor matching algorithm to calculate the similarity between the feature set to be matched and the standard feature vectors in the sample library, and select the multiple sample data with the highest similarity; S44. Perform weighted calculations on the optical parameters corresponding to the selected sample data, with the weights determined based on the similarity values, to obtain preliminary optical parameter estimates for different sub-regions.

[0014] Further, step S5 includes the following sub-steps: S51. Collect ambient light spectral data of the measurement site through an ambient light sensor, calculate the difference with standard ambient light spectral data to obtain the ambient light interference factor, which includes the interference intensity and phase deviation at different wavelengths; S52. Retrieve the equipment calibration log to obtain the lens distortion and sensor response deviation parameters of the current imaging equipment, and calculate the equipment imaging deviation factor by combining the exposure time and gain value during imaging; S53. Establish a correlation model between the correction factor and optical parameters, input the ambient light interference factor and the equipment imaging deviation factor into the model, and generate correction coefficients corresponding to different optical parameters; S54. Use a multiplicative correction method to multiply the preliminary optical parameter estimate by the corresponding correction coefficient to obtain the corrected intermediate value of the optical parameters.

[0015] Beneficial Effects: This invention proposes a snapshot-type method for measuring and calculating dental optical parameters. This method achieves simultaneous multi-spectral imaging through a snapshot-type optical acquisition unit, rapidly acquiring raw data including multiple light components without contacting the tooth, significantly shortening the measurement cycle. Addressing the issue of insufficient correlation between feature extraction and parameter matching, this method first extracts pixel-level spectral features and distinguishes the spectral response ranges of different dental tissues. Then, it divides the tooth into multiple analysis sub-regions based on the anatomical structure of the tooth, accurately matching the spectral attenuation characteristics data of each sub-region with standard data from the sample library. This ensures sufficient differentiation of the optical characteristics of different tissue types and reduces parameter estimation deviations. Regarding the imperfect interference factor correction mechanism, it introduces ambient light interference factors and equipment imaging deviation factors, constructing a dynamic correction model to calibrate the initial estimates. This effectively offsets the influence of external environmental fluctuations and equipment status deviations, ensuring high stability and reliability of core parameters such as refractive index and scattering coefficient. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the overall steps of the method of the present invention. Detailed Implementation

[0017] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] like Figure 1 As shown, a snapshot-based method for measuring and calculating dental optical parameters includes: Step S1. Control the snapshot optical acquisition unit to perform multispectral band synchronous imaging of the target tooth area to acquire the original optical image dataset including the diffuse reflection light, specular reflection light and subsurface scattering light components of the tooth surface; Specifically, step S1 revolves around the precise control of the snapshot-type optical acquisition unit. This unit integrates a multispectral filter array and a high-resolution image sensor with a pixel size of 3.45 micrometers and a frame rate of 30 frames per second, enabling simultaneous imaging of 16 consecutive spectral bands within the 400-1100 nanometer wavelength range. During implementation, the distance between the acquisition unit and the target tooth area is first adjusted to 15-20 centimeters, ensuring the lens optical axis forms a 45-degree angle with the occlusal surface of the tooth. Then, the light source module is activated, using a 50-watt halogen lamp to provide stable incident light. A trigger signal controls a single exposure of the acquisition unit, with the exposure time set to 1 / 100 second. A single snapshot acquires a raw optical image dataset including diffuse reflection, specular reflection, and subsurface scattering components from the tooth surface. The dataset resolution is 2048×1536 pixels, with one frame corresponding to each spectral band, generating a total of 16 single-band images. The significance of this step lies in avoiding measurement errors caused by traditional step-by-step acquisition through synchronous multispectral imaging, providing complete and synchronous raw data support for subsequent spectral feature extraction, and ensuring that the data can truly reflect the optical properties of tooth tissues at different depths.

[0019] Step S2. Extract pixel-level spectral features from the original optical image dataset, distinguish the spectral response intervals corresponding to different tooth tissues, and determine the light intensity distribution data corresponding to the feature wavelengths in different intervals; Specifically, step S2 extracts pixel-level spectral features from the original optical image dataset. First, the image preprocessing module is called to repair bad pixels in 16 frames of single-band images, using neighborhood interpolation to fill in abnormal pixels caused by sensor defects. Then, the spectral feature extraction algorithm is started, traversing all single-band images pixel by pixel, recording the grayscale value of each pixel in 16 spectral bands, with a grayscale value range of 0-4095. The grayscale values ​​are then converted into corresponding light intensity values ​​using the sensor response curve. Based on the dental tissue spectral characteristic database, the spectral response range for enamel is determined to be 400-550 nm, for dentin 550-800 nm, and for cementum 800-1100 nm. Based on this range division, the pixel light intensity data is categorized, forming subsets of light intensity data corresponding to the three tissue types. Further edge detection was performed on each subset, and the Canny operator was used to extract the tooth contours. Background pixels outside the contours and noise pixels at the contour edges were removed. Finally, the light intensity distribution data of the effective tooth region at 16 feature wavelengths were obtained, with each feature wavelength corresponding to a 2048×1536 pixel light intensity distribution image. The significance of this step is to accurately separate the spectral signals of different tooth tissues, laying the foundation for subsequent targeted analysis of the optical parameters of each tissue and improving the tissue specificity of parameter calculation.

[0020] Step S3. Construct an initial analysis matrix of tooth optical properties based on characteristic wavelength light intensity distribution data, divide the tooth into multiple analysis sub-regions in combination with tooth anatomical structure features, and obtain spectral attenuation characteristic data of different sub-regions; Specifically, step S3 constructs an initial analysis matrix of tooth optical properties and acquires spectral attenuation characteristic data for sub-regions. In practice, based on the intensity distribution data of the 16 characteristic wavelengths output in S2, an initial data matrix with a row dimension of 16 (characteristic wavelengths) and a column dimension of 3,145,728 (total number of pixels) is constructed, with matrix elements representing the intensity values ​​of each pixel. Simultaneously, a three-dimensional model of the tooth anatomy is imported. This model, acquired through CBCT scanning with a resolution of 0.1 mm, is projected onto the image plane along the optical acquisition direction. Threshold segmentation is used to divide the tooth into three analysis sub-regions: crown, neck, and root. The boundary pixel coordinates of each sub-region are determined, with the crown region comprising 1.2 million pixels, the neck region comprising 0.8 million pixels, and the root region comprising 1.14 million pixels. The initial data matrix is ​​partitioned according to the boundary coordinates to obtain three sub-matrices. The row mean and column variance corresponding to each characteristic wavelength in each sub-matrix are calculated. The row mean reflects the average intensity of the sub-region at that wavelength, and the column variance reflects the uniformity of the intensity distribution. By leveraging the characteristic of light attenuation with depth in tooth tissue, the intensity difference between adjacent characteristic wavelengths is calculated to obtain spectral attenuation characteristic data for each sub-region. The data is presented in tabular form, including sub-region name, mean intensity of each characteristic wavelength, intensity variance, and attenuation difference between adjacent wavelengths. The significance of this step lies in correlating spectral data with tooth anatomy, enabling zonal analysis of optical properties, and avoiding the ambiguity of local characteristics caused by overall analysis.

[0021] Step S4. Call the preset dental tissue optical parameter sample library, perform feature matching between the spectral attenuation characteristic data of different sub-regions and the standard data in the sample library, and determine the preliminary optical parameter estimates; Specifically, step S4 relies on a pre-set optical parameter sample library of dental tissues. This library contains 1000 different sample data sets, representing different age groups (20-70 years old), genders, and various dental health conditions such as health status, caries, and periodontitis. Each data set includes corresponding spectral attenuation characteristic data and optical parameter data calibrated using standard instruments. During implementation, the core features of the spectral attenuation characteristic data of the three sub-regions obtained in S3 are first extracted, including the wavelength corresponding to the peak light intensity, the slope of the light intensity change with wavelength, and the half-width at half-maximum (WHM) of the light intensity distribution, forming a feature vector to be matched. The feature matching algorithm is then activated, using Euclidean distance to calculate the similarity between the feature vector to be matched and each set of standard feature vectors in the sample library. A similarity threshold of 0.85 is set, and 20 sample data sets with similarities higher than the threshold are selected. Weights are assigned based on similarity values, with higher similarity values ​​carrying greater weights, and the total weights are 1. The optical parameters corresponding to the 20 sets of sample data are weighted and summed to obtain preliminary estimates of the optical parameters for three sub-regions: crown, neck, and root. These estimates include three core parameters: refractive index, scattering coefficient, and absorption coefficient. The significance of this step lies in leveraging the rich data in the sample database to provide reliable preliminary estimates of the optical parameters for each sub-region, reducing the randomness of parameter calculations.

[0022] S5. Introduce ambient light interference factor and equipment imaging deviation factor to dynamically correct the preliminary optical parameter estimates and generate corrected intermediate optical parameter values; Specifically, step S5 performs dynamic correction of the preliminary optical parameter estimates. During implementation, ambient light data from the measurement site is first collected using an ambient light sensor integrated into the acquisition unit. The sensor's measurement range is 0-10000 lux, and the sampling frequency is 10 Hz. Ten sets of data are continuously collected, and the average value is taken. This average is compared with standard ambient light data (500 lux) in the sample library to calculate the ambient light interference factor. This factor includes the difference in interference intensity across different spectral bands. Simultaneously, the calibration log of the imaging equipment is retrieved to obtain parameters such as lens distortion coefficient (less than 1%) and sensor response deviation (less than 2%). Combined with the current imaging exposure time and gain value, the equipment imaging deviation factor is calculated using the equipment error model. A correction model is constructed based on these two types of factors, converting the ambient light interference factor and the equipment imaging deviation factor into corresponding correction coefficients. The ambient light correction coefficient ranges from 0.95 to 1.05, and the equipment deviation correction coefficient ranges from 0.98 to 1.02. A multiplicative correction method is used, where the preliminary optical parameter estimates for each sub-region obtained in S4 are multiplied by their corresponding correction coefficients to generate intermediate corrected optical parameter values. This ensures that the parameter values ​​are unaffected by environmental and equipment condition fluctuations. The significance of this step lies in eliminating the influence of external interference and equipment errors on the parameter results, thereby improving the accuracy and reliability of the optical parameters.

[0023] Step S6. Construct an overall tooth optical parameter distribution model based on the intermediate values ​​of optical parameters in different sub-regions, and output the core optical parameters of the target tooth, including refractive index, scattering coefficient, and absorption coefficient.

[0024] Specifically, step S6 aims to construct a comprehensive optical parameter distribution model of the teeth and output core parameters. During implementation, the intermediate values ​​of optical parameters from the three sub-regions obtained in S5 are used as a basis. Combined with the spatial coordinate information of the three-dimensional model of the tooth anatomy, a spatial interpolation algorithm is employed to smooth the parameter values ​​at the boundaries of each sub-region, avoiding abrupt parameter changes. Based on the interpolated parameter data, a comprehensive distribution model including three-dimensional spatial coordinate information and corresponding optical parameter values ​​is constructed. The model's spatial resolution is 0.1 mm, consistent with the resolution of the CBCT three-dimensional model, accurately corresponding to the optical characteristics of each spatial location of the tooth. Further processing of the model data extracts the average optical parameter values ​​from different tissue regions such as enamel, dentin, and cementum. The refractive index and scattering coefficient are output primarily for the enamel region, while the absorption and scattering coefficients are output primarily for the dentin and cementum regions. Finally, the results are output in two forms: a data table and a three-dimensional model visualization. The table includes information such as tissue type, refractive index, scattering coefficient, and absorption coefficient, while the three-dimensional model uses different colors to indicate the distribution differences of each parameter. The significance of this step lies in integrating the regional parameter data to form a complete spatial distribution of tooth optical parameters, providing comprehensive and intuitive optical characteristic data for clinical diagnosis.

[0025] Preferably, the following model is used when constructing the initial analysis matrix of tooth optical properties in step S3: Where M represents the initial optical characteristic analysis matrix, n represents the number of characteristic wavelengths, and m represents the number of pixels. This represents the weight coefficient of the j-th pixel under the i-th feature wavelength. This represents the light intensity value of the j-th pixel at the i-th characteristic wavelength. Indicates the reference value of incident light intensity. This represents the spectral attenuation index of the j-th pixel at the i-th characteristic wavelength. This represents the scattering influence coefficient of the j-th pixel at the i-th characteristic wavelength. This represents the light propagation path factor of the j-th pixel at the i-th characteristic wavelength.

[0026] Specifically, in step S3, the initial analysis matrix of tooth optical properties is constructed. This model is used to integrate characteristic wavelength light intensity data and pixel characteristics to achieve a structured expression of optical information. During implementation, the specific values ​​and determination methods of each parameter in the model are first clarified: the number of feature wavelengths is set to 16 based on the multispectral configuration of the acquisition unit, and the number of pixels corresponds to an image resolution of 3,145,728; the weight coefficient of the j-th pixel under the i-th feature wavelength is determined according to the type of tooth tissue in which the pixel is located, with a value of 0.35-0.45 for the enamel region, 0.3-0.4 for the dentin region, and 0.25-0.35 for the cementum region, ensuring that the light intensity contribution of different tissues is reasonably allocated; the incident light intensity benchmark value is obtained through the power calibration of the light source module and is set as the standard light intensity value of a 50-watt halogen lamp at a distance of 15-20 cm; the spectral attenuation index is determined based on the measured data of the spectral characteristics of similar tooth tissues, ranging from 0.8 to 1.2; the scattering influence coefficient is set to 0.15-0.25 based on the scattering law of light in tooth tissue; and the light propagation path factor is calculated based on the tissue depth corresponding to the pixel in the three-dimensional model of tooth anatomy, with a larger value for the greater depth, ranging from 0.5 to 1.5. During calculation, the data is first traversed along the wavelength and pixel dimensions, and the logarithmic and exponential terms are calculated point by point. Then, the weighting coefficients and scattering influence coefficients are added to adjust the data, and finally, the summation is performed to generate the initial analysis matrix. The significance of this model lies in transforming scattered light intensity data into a matrix form with anatomical structural correlation, providing a structured data foundation for subsequent sub-region division and characteristic analysis. At the same time, by adjusting multiple parameters, the effective signal in the light intensity data is enhanced and noise is suppressed, thereby improving the accuracy of optical characteristic expression.

[0027] Preferably, the feature matching process in step S4 adopts the following model: Where S represents the feature matching degree, and p represents the dimension of the matching feature. This represents the spectral attenuation characteristics of the k-th dimension of the sub-region. This represents the k-th dimension of the standard data in the sample library. This represents the average value of the spectral attenuation characteristics data for the sub-region. This represents the average value of the standard data in the sample library. This represents the matching deviation adjustment factor.

[0028] Specifically, in step S4, the feature matching process uses a model to quantify the similarity between the sub-region spectral attenuation characteristic data and the standard data in the sample library, providing a basis for preliminary optical parameter estimation. During implementation, the model parameter setting rules are first clarified: the matching feature dimension is determined to be 3 based on the core indicators of spectral attenuation characteristics, including peak wavelength, attenuation slope, and half-width at half-maximum (WHM); both the k-th dimension spectral attenuation characteristic data of the sub-region and the k-th dimension standard data of the sample library come from a preprocessed structured dataset. The sub-region data is generated by calculation in step S3, and the sample library data is the mean of 1000 samples calibrated with standard instruments; the average value of the sub-region spectral attenuation characteristic data and the average value of the standard data in the sample library are obtained by calculating the arithmetic mean of the data in each dimension; the matching deviation adjustment factor is set to 0.05-0.15 based on the dispersion of the sample library, with a larger value for higher dispersion, used to correct the influence of mean deviation on the matching results. The calculation process is as follows: first, calculate the weighted sum of the numerator and the two square roots of the denominator to obtain the basic matching degree; then calculate the correction term using the mean deviation and the adjustment factor, and multiply the two to obtain the final matching degree. The significance of this model lies in avoiding the limitations of single-feature matching by comprehensively comparing and correcting for biases in multi-dimensional features, improving the reliability of matching results, ensuring that the selected sample data can accurately reflect the current dental tissue status of the sub-region, and providing a high-quality reference for preliminary parameter estimation.

[0029] Preferably, the dynamic correction process in step S5 adopts the following model: ,in, This represents the intermediate value of the corrected optical parameters. This represents a preliminary estimate of optical parameters. This represents the environmental interference correction factor. Indicates the actual ambient light intensity. Indicates standard ambient light intensity. This represents the equipment deviation correction factor. This represents the actual imaging deviation value. This indicates the equipment calibration deviation threshold.

[0030] Specifically, the dynamic correction process in step S5 uses a model to eliminate the influence of ambient light and equipment imaging deviation on the preliminary optical parameter estimates, improving parameter accuracy. During implementation, parameter determination needs to combine actual measurements with equipment characteristics: the preliminary optical parameter estimates are the refractive index, scattering coefficient, and absorption coefficient output in step S4; the environmental interference correction coefficient is calibrated through numerous ambient light interference experiments, ranging from 0.02 to 0.08, with larger values ​​for stronger interference; the actual ambient light intensity is measured by a sensor integrated on the acquisition unit, with a sampling frequency of 10 Hz, continuously collecting 10 sets and averaging them; the standard ambient light intensity is set to the ideal ambient light intensity of 500 lux for clinical measurement; the equipment deviation correction coefficient is determined based on the equipment calibration log, ranging from 0.01 to 0.04, with larger values ​​for larger deviations; the actual imaging deviation value is detected in real time by the equipment's built-in calibration module, and the equipment calibration deviation threshold is the maximum allowable deviation value set at the factory; during calculation, the ambient light interference correction term and the equipment deviation correction term are calculated separately, both using the form of 1 plus a correction function, and then the preliminary estimate is multiplied by the two correction terms to obtain the intermediate value. The significance of this model lies in the construction of a two-factor dynamic correction mechanism, which can respond in real time to fluctuations in ambient light and changes in equipment status. It solves the problem that parameter estimation in traditional methods is easily affected by external interference, and ensures the stability and reliability of parameter results under different measurement conditions.

[0031] Preferably, the overall tooth optical parameter distribution model constructed in step S6 adopts the following model: ,in, Representing three-dimensional spatial coordinates The optical parameter value at the location, where r represents the number of sub-regions analyzed. This represents the spatial weight of the q-th sub-region. This represents the intermediate value of the optical parameters of the q-th sub-region. This represents the spatial decay coefficient of the parameter in the q-th sub-region. Representing coordinates The distance to the center of the q-th sub-region.

[0032] Specifically, step S6 involves constructing a model of the overall tooth optical parameter distribution. This model integrates sub-region parameter data to generate a three-dimensional parameter distribution. During implementation, parameter settings must consider both anatomical structure and spatial characteristics: the number of sub-regions is divided into three based on the tooth's anatomical structure: crown, neck, and root. The spatial weight of the q-th sub-region is determined by its volume percentage within the overall tooth structure: 0.45-0.55 for the crown, 0.2-0.3 for the neck, and 0.25-0.35 for the root. The spatial attenuation coefficient is set based on the spatial variation of optical parameters within the tooth tissue: 0.1-0.2 for the crown, 0.2-0.3 for the neck, and 0.3-0.4 for the root, reflecting the attenuation characteristics of parameters with tissue depth. The distance from the coordinates to the center of the sub-region is calculated using the spatial coordinates of the three-dimensional model of the tooth's anatomical structure, with the unit consistent with the model resolution at 0.1 mm. During calculation, each coordinate point in the three-dimensional space is traversed, and the contribution value of each sub-region parameter is calculated point by point. The summation yields the optical parameter value for that coordinate. The significance of this model lies in realizing the extension of sub-region parameters into three-dimensional space. By adjusting the spatial weights and attenuation coefficients, it ensures that the parameter distribution conforms to the physiological structure and optical properties of tooth tissue, solves the problem of boundary discontinuity when splicing partition parameters, and provides intuitive and complete spatial distribution information of optical parameters for clinical practice.

[0033] Preferably, the following model is used to determine the light intensity distribution data in step S2: ,in, Indicates wavelength ,coordinate The light intensity value at that location, Indicates the range of imaging solid angles. Indicates wavelength ,coordinate Location, angle Radiance in direction, Indicates the zenith angle. Indicates the azimuth angle.

[0034] Specifically, in step S2, the process of determining the light intensity distribution data involves using a model to convert radiance data into pixel light intensity values, accurately quantifying light intensity information at different locations. During implementation, parameter determination must consider the imaging optical characteristics: the wavelength range is consistent with the multispectral configuration of the acquisition unit, ranging from 400-1100 nanometers; the coordinates are pixel coordinates on the image plane, corresponding to the sensor pixel size of 3.45 micrometers; the imaging solid angle range is determined by the lens field of view of the acquisition unit, set to 60 degrees to ensure complete coverage of the target tooth area; the radiance data is calculated through the sensor's response values ​​to light at different angles, with a zenith angle range of 0-30 degrees and an azimuth angle range of 0-360 degrees, reflecting the radiation characteristics of light in different directions. During calculation, the radiance is first integrated within the set solid angle range. The integration process incorporates the adjustment effect of the zenith angle cosine value to correct the influence of the incident angle on the light intensity measurement, ultimately obtaining the light intensity value at each wavelength and each pixel coordinate. The significance of this model lies in establishing a quantitative relationship between radiance and pixel light intensity, taking into account the influence of imaging solid angle and light angle, avoiding the light intensity quantization deviation caused by directly using sensor grayscale values, providing accurate light intensity data support for subsequent spectral feature extraction, and improving the source data quality of the entire parameter calculation process.

[0035] Preferably, step S2 includes the following sub-steps: S21. Call the spectral analysis module to perform channel separation on the original optical image dataset, extract the wavelength information corresponding to each imaging channel, establish the mapping relationship between wavelength and image channel, and obtain a single-wavelength image sequence; S22. Perform pixel traversal on the single-wavelength image sequence, record the gray value of each pixel at different wavelengths, convert the gray value into the corresponding light intensity response value, and form a pixel-wavelength-light intensity three-dimensional data array; S23. Based on the tooth tissue spectral database, determine the characteristic spectral intervals corresponding to enamel, dentin, and cementum, divide the three-dimensional data array according to the characteristic spectral intervals, and obtain light intensity distribution subsets corresponding to different tissue types; S24. Remove edge pixels from different light intensity distribution subsets, retain the light intensity data in the effective tooth area, and organize them to form light intensity distribution data corresponding to the characteristic wavelength.

[0036] Specifically, step S2 achieves accurate extraction of characteristic wavelength light intensity distribution data through multi-stage processing, laying the foundation for subsequent spectral analysis. The implementation proceeds in four steps: Step S21 activates the spectral analysis module, which integrates a band recognition algorithm to quickly separate the 16 imaging channels of the original optical image dataset. Each channel corresponds to a specific wavelength within the range of 400-1100 nanometers. After separation, a one-to-one mapping relationship between wavelength and channel is automatically established, generating a single-wavelength image sequence comprising 16 frames, each with a resolution of 2048×1536 pixels. Step S22 performs pixel traversal on the single-wavelength image sequence, using a row-by-row, column-by-column scanning method to record the grayscale value of each pixel at 16 wavelengths, with a grayscale value range of 0-4095. Then, based on the sensor's factory-calibrated response curve, the grayscale values ​​are converted into corresponding light intensity response values, ultimately forming a pixel-wavelength-light intensity three-dimensional data array with dimensions of 3145728×16. In stage S23, a dental tissue spectral database is accessed, containing 1000 sets of clinically measured data. The characteristic spectral ranges for enamel, dentin, and cementum are clearly defined as 400-550 nm, 550-800 nm, and 800-1100 nm, respectively. The three-dimensional data array is segmented according to these range boundaries to obtain subsets of light intensity distribution for each of the three tissue types. In stage S24, the Canny edge detection algorithm is used to extract the tooth contour. The contour extraction threshold is set to a grayscale value of 120. Background pixels outside the contour and noise regions with a contour edge width of 5 pixels are removed. The light intensity data within the effective area is retained and sorted by wavelength to form characteristic wavelength light intensity distribution data. This step, through phased and refined processing, achieves a precise correlation between spectral data and dental tissue, ensuring the tissue specificity and reliability of the light intensity distribution data.

[0037] Preferably, step S3 includes the following sub-steps: S31. Based on the characteristic wavelength light intensity distribution data, construct an initial data matrix with row vectors representing characteristic wavelengths and column vectors representing pixels, and normalize the matrix elements to obtain a standardized data matrix; S32. Import a three-dimensional model of the tooth anatomy, project the model onto the image plane, divide the image regions corresponding to the crown, neck, and root, and determine the boundary pixel coordinates of different regions; S33. Divide the standardized data matrix into sub-matrices according to the boundary pixel coordinates to obtain sub-matrices corresponding to different anatomical regions, calculate the row mean and column variance of each sub-matrix, and obtain the basic data of spectral attenuation characteristics; S34. Combine the propagation characteristics of light in tooth tissue to perform gradient calculation on the basic data of spectral attenuation characteristics to obtain the spectral attenuation characteristic data of different sub-regions.

[0038] Specifically, step S3 obtains spectral attenuation characteristic data for each sub-region through anatomical structure association and data partitioning. The implementation strictly follows four steps: In stage S31, based on the 16 characteristic wavelength light intensity distribution data output from S2, an initial data matrix is ​​constructed where rows represent characteristic wavelengths and columns represent pixels. The matrix elements are light intensity values, and linear scaling is used to normalize the element values ​​to the 0-1 range, eliminating differences in light intensity magnitude between different wavelengths, resulting in a standardized data matrix. In stage S32, a three-dimensional model of the tooth anatomy obtained through CBCT scanning is imported. The model has a spatial resolution of 0.1 mm and includes complete structural information of the crown, neck, and root. The model is projected onto the image plane along the optical acquisition direction, with the projection scaling ratio consistent with the imaging magnification. Threshold segmentation is used to divide each anatomical region. The threshold for the crown region is set to a grayscale value of 180-255, for the neck 120-180, and for the root 50-120, determining the boundary pixel coordinates of each region. In stage S33, the standardized data matrix is ​​partitioned based on boundary coordinates, resulting in three sub-matrices. The row mean (reflecting the average light intensity at each wavelength) and column variance (reflecting the uniformity of light intensity distribution) are calculated for each sub-matrix. The row mean is calculated using the arithmetic mean method, and the column variance is calculated using the sample variance formula, yielding the basic data for spectral attenuation characteristics. In stage S34, considering the exponential attenuation of light with depth in tooth tissue, the light intensity difference between adjacent characteristic wavelengths is calculated. This difference is the difference between the average light intensity of the subsequent wavelength and the average light intensity of the preceding wavelength, obtaining the spectral attenuation characteristic data for each sub-region. The data includes four dimensions: wavelength, average light intensity, light intensity variance, and attenuation difference. This step, through deep fusion of anatomical structure and spectral data, achieves the zonal quantification of spectral attenuation characteristics, providing structured, region-specific data for subsequent parameter matching.

[0039] Preferably, step S4 includes the following sub-steps: S41. Access a preset dental tissue optical parameter sample library, which includes standard spectral attenuation data and optical parameter data corresponding to different ages, genders, and dental health states, and establish a sample index table; S42. Extract feature vectors from the spectral attenuation characteristic data of different sub-regions, including peak wavelength, attenuation slope, and full width at half maximum (FWHM) feature parameters, to form a feature set to be matched; S43. Use the K-nearest neighbor matching algorithm to calculate the similarity between the feature set to be matched and the standard feature vectors in the sample library, and select the multiple sample data with the highest similarity; S44. Perform weighted calculations on the optical parameters corresponding to the selected sample data, with the weights determined based on the similarity values, to obtain preliminary optical parameter estimates for different sub-regions.

[0040] Specifically, step S4 provides preliminary optical parameter estimates for each sub-region through sample library matching. This is implemented in four sequential steps: Step S41 accesses a pre-defined dental tissue optical parameter sample library, which stores 1000 sets of sample data, covering different ages (20-70 years), different genders, and six health states including healthy, caries, and periodontitis. Each set of data includes spectral attenuation data and optical parameters calibrated with standard instruments. A sample index table is established, indexed by tissue type, age, and health state, with an index response time of less than 0.5 seconds. Step S42 extracts the core features of the spectral attenuation characteristics of each sub-region, including peak wavelength (the wavelength corresponding to the maximum light intensity), attenuation slope (the average difference in light intensity between adjacent wavelengths), and half-width at half-maximum (the width of the wavelength interval at half the peak light intensity), forming a 3D feature set to be matched. Feature values ​​are retained to two decimal places. In stage S43, the K-nearest neighbor matching algorithm is used, with K set to 20. The Euclidean distance between the feature set to be matched and the standard feature vectors in the sample library is calculated; the smaller the distance, the higher the similarity. The 20 sets of sample data with the smallest distance are selected, and the similarity threshold is set to 0.85. In stage S44, weights are calculated based on similarity, which is positively correlated with weights. Normalization is used to make the sum of weights equal to 1. The optical parameters of the 20 sets of samples are weighted and summed to obtain preliminary optical parameter estimates for the crown, neck, and root of the tooth. The estimated values ​​include three core parameters: refractive index, scattering coefficient, and absorption coefficient. This step, supported by rich data from the sample library and multi-dimensional matching, ensures the accuracy and relevance of the preliminary parameter estimates.

[0041] Preferably, step S5 includes the following sub-steps: S51. Collect ambient light spectral data of the measurement site through an ambient light sensor, calculate the difference with standard ambient light spectral data to obtain an ambient light interference factor, which includes the interference intensity and phase deviation at different wavelengths; S52. Retrieve the equipment calibration log to obtain the lens distortion and sensor response deviation parameters of the current imaging equipment, and calculate the equipment imaging deviation factor by combining the exposure time and gain value during imaging; S53. Establish a correlation model between the correction factor and optical parameters, input the ambient light interference factor and the equipment imaging deviation factor into the model, and generate correction coefficients corresponding to different optical parameters; S54. Use a multiplicative correction method to multiply the preliminary optical parameter estimate by the corresponding correction coefficient to obtain the corrected intermediate value of the optical parameters.

[0042] Specifically, step S5 improves the reliability of optical parameters through dual-factor dynamic correction. This is implemented in four strict steps: In step S51, the ambient light sensor integrated into the acquisition unit is activated. The sensor has a measurement range of 0-10000 lux, an accuracy of ±5 lux, and a sampling frequency of 10 Hz. Ten sets of ambient light spectral data are continuously acquired, and the average value is taken as the measured value. This average value is then compared with the standard ambient light spectral data of 500 lux in the sample library, and wavelength-by-wavelength difference calculations are performed to obtain the ambient light interference factor, which includes the interference intensity and phase deviation of 16 wavelengths. In step S52, the calibration log of the imaging equipment is retrieved. The log stores nearly 30 days of equipment status data. The current lens distortion coefficient (less than 1%) and sensor response deviation (less than 2%) are extracted. Combined with the 1 / 100-second exposure time and 1.2 times gain value during imaging, these are substituted into the equipment error model to calculate the equipment imaging deviation factor, which includes systematic error and random error components. In stage S53, a correlation model between the correction factor and optical parameters is established. This model is calibrated through 100 sets of comparative experiments. Ambient light interference factors and equipment imaging deviation factors are input into the model, which outputs correction coefficients for refractive index, scattering coefficient, and absorption coefficient, ranging from 0.95 to 1.05. In stage S54, a multiplicative correction method is used. The preliminary optical parameter estimates for each sub-region obtained in S4 are multiplied by their corresponding correction coefficients, with the calculations rounded to four decimal places, yielding the corrected intermediate values ​​of the optical parameters. This step, through real-time interference detection and dynamic correction, effectively offsets the influence of environmental and equipment factors, ensuring the stability and accuracy of the intermediate parameter values.

[0043] A snapshot-based method for measuring and calculating dental optical parameters offers significant advantages in measurement efficiency, enabling rapid acquisition of dental optical parameters. Its core relies on the multispectral band synchronous imaging capability of a snapshot-based optical acquisition unit. This method allows for the simultaneous capture of raw optical image datasets, including diffuse surface reflection, specular reflection, and subsurface scattering, without requiring contact with the teeth. This eliminates the cumbersome process of step-by-step acquisition and multiple calibrations required in traditional methods, significantly shortening the overall cycle from data acquisition to parameter output. This method is perfectly suited to the needs of rapid dental characteristic analysis in clinical settings.

[0044] In terms of parameter measurement accuracy, this method demonstrates significant advantages, outputting highly reliable core optical parameters. Through pixel-level spectral feature extraction, it can accurately distinguish the spectral response ranges of different tissues such as enamel and dentin. Combined with the sub-regional analysis based on tooth anatomical structure, the spectral attenuation characteristics of each region can be analyzed in a targeted manner. Simultaneously, it utilizes a pre-defined sample library for feature matching, providing a reliable reference for parameter estimation. This improves the accuracy of parameter calculation from the data processing source, resolving the estimation bias problem caused by the ambiguity in tissue characteristic differentiation in traditional methods.

[0045] This method addresses the problem of insufficient correlation between feature extraction and parameter matching by combining spectral feature analysis with anatomical structural partitioning, achieving a precise correspondence between data and the actual physiological structure of teeth. To address the problem of an imperfect interference factor correction mechanism, it specifically introduces two types of factors: ambient light and equipment imaging deviation, and constructs a dynamic correction model to calibrate the preliminary results. This effectively offsets the impact of external environmental fluctuations and equipment status differences, ensuring the stability and consistency of output parameters under different measurement conditions.

[0046] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A snapshot method of measuring optical parameters of teeth, characterized in that, Comprise: Step S1. Control snapshot optical acquisition unit to carry out multispectral waveband synchronous imaging to target tooth area, obtain original optical image data set including tooth surface diffuse reflection light, specular reflection light and subsurface scattering light component; Step S2. Carry out pixel level spectral feature extraction to original optical image data set, distinguish spectral response interval corresponding to different tooth tissues, determine light intensity distribution data corresponding to feature wavelength in different interval; Step S3. Construct tooth optical property initial analysis matrix based on feature wavelength light intensity distribution data, divide multiple analysis subareas in combination with tooth anatomical structure feature, obtain spectral attenuation characteristic data of different subareas; Step S4. Call preset tooth tissue optical parameter sample library, carry out feature matching to spectral attenuation characteristic data of different subareas and standard data in sample library, determine preliminary optical parameter estimated value; Step S5. Introduce environmental light interference factor and equipment imaging deviation factor, carry out dynamic correction to preliminary optical parameter estimated value, generate corrected optical parameter intermediate value; Step S6. Construct overall tooth optical parameter distribution model based on optical parameter intermediate value of different subareas, output refractive index, scattering coefficient and absorption coefficient core optical parameter of target tooth.

2. A snapshot method of measuring and calculating optical parameters of teeth according to claim 1, characterized in that, The following model is used in step S3 to construct the initial analysis matrix of tooth optical properties: wherein M represents the initial analysis matrix of optical properties, n represents the number of characteristic wavelengths, m represents the number of pixel points, represents the weight coefficient of the jth pixel at the ith characteristic wavelength, represents the light intensity value of the jth pixel at the ith characteristic wavelength, represents the reference value of incident light intensity, represents the spectral attenuation index of the jth pixel at the ith characteristic wavelength, represents the scattering influence coefficient of the jth pixel at the ith characteristic wavelength, represents the light propagation path factor of the jth pixel at the ith characteristic wavelength.

3. A snapshot method of measuring and calculating optical parameters of teeth according to claim 1, characterized in that, The feature matching process in step S4 adopts the following model: wherein S represents the feature matching degree, p represents the matching feature dimension, represents the kth dimension spectral attenuation characteristic data of the sub-region, represents the kth dimension standard data of the sample library, represents the average value of the spectral attenuation characteristic data of the sub-region, represents the average value of the standard data of the sample library, represents the matching deviation adjustment factor.

4. The snapshot method of measuring and calculating optical parameters of teeth according to claim 1, wherein, The dynamic correction process in step S5 uses the following model: wherein, represents the corrected optical parameter intermediate value, represents the preliminary optical parameter estimate, represents the environmental interference correction factor, represents the actual ambient light intensity, represents the standard ambient light intensity, represents the device bias correction factor, represents the actual imaging bias value, represents the device calibration bias threshold.

5. The snapshot method of measuring optical parameters of teeth according to claim 1, wherein, The model of the whole tooth optical parameter distribution is constructed in step S6, using the following model: wherein, represents the optical parameter value at the three-dimensional space coordinate , r represents the number of analysis sub-regions, represents the spatial weight of the qth sub-region, represents the optical parameter intermediate value of the qth sub-region, represents the parameter space attenuation coefficient of the qth sub-region, represents the distance from the coordinate to the center of the qth sub-region.

6. The snapshot method of measuring optical parameters of teeth according to claim 1, wherein, The light intensity distribution data determined in step S2 adopts the following model: wherein, denotes the wavelength , the coordinate , the light intensity value at the coordinate denotes the imaging solid angle range, denotes the wavelength , the coordinate , the angle , the radiance in the direction of the coordinate denotes the zenith angle, denotes the azimuth angle.

7. The snapshot method of measuring optical parameters of teeth according to claim 1, wherein, Step S2 comprises the following steps: S21. Call spectral analysis module to carry out channel separation to original optical image data set, extract wavelength information corresponding to each imaging channel, establish mapping relationship between wavelength and image channel, obtain single wavelength image sequence; S22. Carry out pixel traversal to single wavelength image sequence, record gray value of each pixel under different wavelengths, convert gray value into corresponding light intensity response value, form pixel-wavelength-light intensity three-dimensional data array; S23. Determine feature spectral interval corresponding to enamel, dentin and cementum based on tooth tissue spectral database, segment three-dimensional data array according to feature spectral interval, obtain light intensity distribution subset corresponding to different tissue types; S24. Carry out edge pixel elimination to different light intensity distribution subsets, retain light intensity data in effective tooth area, organize and form light intensity distribution data corresponding to feature wavelength.

8. The snapshot method of measuring optical parameters of teeth according to claim 1, wherein, Step S3 comprises the following steps: S31. Based on feature wavelength light intensity distribution data, construct initial data matrix with row vector as feature wavelength and column vector as pixel point, carry out normalization processing to matrix element, obtain standardized data matrix; S32. Import tooth anatomical structure three-dimensional model, project model to image plane, divide out image area corresponding to tooth crown, tooth neck and tooth root, determine boundary pixel coordinates of different areas; S33. According to boundary pixel coordinates, carry out partition cutting to standardized data matrix, obtain submatrix corresponding to different anatomical areas, calculate row mean and column variance of each submatrix, obtain spectral attenuation characteristic basic data; S34. In combination with propagation characteristic of light in tooth tissue, carry out gradient calculation to spectral attenuation characteristic basic data, obtain spectral attenuation characteristic data of different subareas.

9. The snapshot method of measuring optical parameters of teeth according to claim 1, wherein, Step S4 comprises the following sub-steps: S41. Accessing a preset tooth tissue optical parameter sample library, the sample library comprising standard spectral attenuation data and optical parameter data corresponding to different ages, genders and tooth health statuses, and establishing a sample index table; S42. Extracting feature vectors of spectral attenuation characteristic data of different sub-regions, including peak wavelength, attenuation slope and half-width characteristic parameters, to form a to-be-matched feature set; S43. Using a K-nearest neighbor matching algorithm, performing similarity calculation on the to-be-matched feature set and standard feature vectors in the sample library, and screening a plurality of sample data with the highest similarity; S44. Performing weighted calculation on optical parameters corresponding to the screened sample data, and determining weights based on similarity values, to obtain preliminary optical parameter estimation values of different sub-regions.

10. The snapshot method of measuring optical parameters of teeth according to claim 1, wherein, Step S5 comprises the following sub-steps: S51. Collecting environmental light spectrum data of a measurement site through an environmental light sensor, and performing difference calculation on the environmental light spectrum data and standard environmental light spectrum data to obtain an environmental light interference factor, the factor comprising interference intensity and phase deviation at different wavelengths; S52. Retrieving a device calibration log to obtain lens distortion and sensor response deviation parameters of a current imaging device, and combining exposure time and gain value during imaging to calculate a device imaging deviation factor; S53. Establishing a correlation model of correction factors and optical parameters, inputting the environmental light interference factor and the device imaging deviation factor into the model, and generating correction coefficients corresponding to different optical parameters; S54. Using a multiplication correction method, multiplying the preliminary optical parameter estimation values and the corresponding correction coefficients to obtain corrected optical parameter intermediate values.

Images