Face fitting adjustment method for intelligent protective glasses

Abstract

The invention discloses a face fitting adjustment method for intelligent protective glasses, and the method comprises the steps: obtaining the face three-dimensional scanning data of a wearer through a 3D vision module, and determining an initial fitting parameter based on the three-dimensional scanning data; obtaining a relative position relationship between the pressure sensor array and the 3D vision module, and performing space coordinate conversion on the initial fitting parameter according to the relative position relationship to obtain a first adjustment parameter; determining a relative included angle relationship between the pressure sensor distribution plane and the frame installation plane, and performing mechanical compensation on the first adjustment parameter according to the relative included angle relationship to obtain a second adjustment parameter; calculating a proportionality coefficient set of a first mapping relation and a second mapping relation according to the first mapping relation of the wearer in the standard fitting state and the second mapping relation monitored in real time, and optimizing the second adjustment parameter according to the proportionality coefficient set to obtain a target adjustment parameter; according to the invention, through multi-source data fusion, dynamic mechanical compensation and subjective comfort matching, the accuracy and comfort of face fitting adjustment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent protective glasses adjustment, and in particular to a method for adjusting the fit of intelligent protective glasses on a human face. Background Art

[0002] In modern industrial production, medical surgery, outdoor adventures, and various special work environments, protective glasses, as an important piece of personal protective equipment, play an irreplaceable role in ensuring personnel's eye safety. However, traditional protective glasses often face a key problem in actual use: due to the large differences in facial shapes between users, protective glasses are difficult to achieve a perfect fit with the face, resulting in a series of problems such as poor wearing comfort, insufficient sealing, and limited field of view. These problems not only affect the user's work efficiency and experience, but also may allow harmful substances or foreign objects to enter the glasses due to poor sealing, causing potential damage to the eyes.

[0003] Currently, the existing protective eyewear adjustment methods on the market are relatively limited. Some products attempt to adapt to different face shapes through simple mechanical structures, such as adjustable nose pads or temples. However, this adjustment method has significant limitations in flexibility and precision, and cannot meet the diverse needs of facial shapes. Traditional methods also face problems such as fit failure caused by posture changes, and the lack of comfort assessment during the fit adjustment process. These problems are even more prominent, and there is an urgent need for a protective eyewear solution that can intelligently and precisely adjust the fit of the face. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for adjusting the fit of smart protective glasses to the face, so as to solve the above-mentioned problems existing in the prior art.

[0005] The specific application is as follows:

[0006] A method for adjusting the facial fit of smart protective glasses is applied to the smart protective glasses, wherein the smart protective glasses include a pressure sensor array, a 3D vision module, and an inertial measurement unit. The pressure sensor array and the 3D vision module are installed at different positions of the glasses frame. The method comprises:

[0007] S1. Obtaining three-dimensional facial scanning data of the wearer through a 3D vision module, and determining initial fitting parameters based on the three-dimensional scanning data;

[0008] S2. Obtaining a relative positional relationship between the pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters based on the relative positional relationship to obtain a first adjustment parameter; the relative positional relationship includes: a physical positional relationship between the pressure sensor array and the 3D vision module and a proportional relationship between pressure resolution and 3D reconstruction accuracy;

[0009] S3. Determine the relative angle between the pressure sensor distribution plane and the frame mounting plane based on the mounting plane of the eyeglass frame, and perform mechanical compensation on the first adjustment parameter according to the relative angle to obtain a second adjustment parameter;

[0010] S4. Establish a first mapping relationship between contact pressure distribution and facial contour and standard fitting facial contour features based on the wearer's historical data in a standard fitting state, monitor the contact pressure distribution and facial contour between the wearer's face and the eyeglass frame in real time, construct a second mapping relationship and real-time fitting facial contour features between the two, calculate a set of proportional coefficients between the first mapping relationship and the second mapping relationship, dynamically optimize the second adjustment parameter based on the proportional coefficient set to obtain a target adjustment parameter, and control the adjustment mechanism to execute the target adjustment parameter.

[0011] Furthermore, the S1 includes:

[0012] Acquire three-dimensional facial scanning data of the wearer, and establish a three-dimensional facial point cloud model based on the three-dimensional facial scanning data;

[0013] Extract the nose bridge curvature features and ear pinna positioning features through the facial 3D point cloud model;

[0014] Calculate the spatial geometric relationship parameters between the nose bridge curvature features and the auricle positioning features;

[0015] The initial fitting parameters are determined by spatial geometric relationship parameters, and the initial fitting parameters include nose pad fitting parameters, temple opening and closing parameters, frame inclination parameters and pressure pre-distribution parameters; the nose pad fitting parameters characterize the curvature matching degree of the nose bridge contact surface; the temple opening and closing parameters characterize the matching of the auricle distance and the temple length; the frame inclination parameters characterize the angle between the frame plane and the facial coronal plane; and the pressure pre-distribution parameters characterize the expected pressure distribution of each contact area.

[0016] Furthermore, the acquiring of the relative positional relationship between the pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters according to the relative positional relationship to obtain the first adjustment parameters, includes:

[0017] Obtaining a physical position offset between the pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters based on the physical position offset;

[0018] A proportional relationship between pressure resolution and three-dimensional reconstruction accuracy is obtained, and initial fitting parameters after spatial coordinate conversion are accurately matched according to the proportional relationship to obtain first adjustment parameters.

[0019] Furthermore, the obtaining of the physical position offset between the pressure sensor array and the 3D vision module, and performing spatial coordinate conversion on the initial fitting parameters based on the physical position offset, includes:

[0020] Establish the global coordinate system of the eyeglass frame and the local coordinate system of the pressure sensor;

[0021] Determine the transformation matrix between the two coordinate systems based on the calibration parameters of the 3D vision module;

[0022] The initial fitting parameters are transformed from the global coordinate system to the local coordinate system through the transformation matrix.

[0023] Furthermore, the obtaining of the proportional relationship between the pressure resolution and the three-dimensional reconstruction accuracy, and performing precision matching on the initial fitting parameters after the spatial coordinate conversion to obtain the first adjustment parameter according to the proportional relationship, includes:

[0024] Calculate the ratio of the pressure sensor spatial resolution to the 3D vision module point cloud density;

[0025] Performing Newton interpolation algorithm compensation on the converted fitting parameters according to the ratio;

[0026] A Kalman filter is applied to smooth the compensation result of the Newton interpolation algorithm to obtain the first adjustment parameter.

[0027] Furthermore, the determining of the relative angle relationship between the pressure sensor distribution plane and the frame mounting plane based on the mounting plane of the eyeglass frame, and performing mechanical compensation on the first adjustment parameter according to the relative angle relationship to obtain the second adjustment parameter, includes:

[0028] The spatial posture of the glasses frame is obtained through the inertial measurement unit to generate posture data;

[0029] Calculate the actual angle between the normal vector of the pressure sensor plane and the normal vector of the facial contact plane through the posture data;

[0030] performing vector decomposition on the pressure measurement value according to the actual angle;

[0031] Based on the decomposition result, mechanical compensation is performed on the first adjustment parameter to obtain the second adjustment parameter.

[0032] Furthermore, establishing a first mapping relationship between contact pressure distribution and facial contour based on historical data of the wearer in a standard fitting state includes:

[0033] In a standard fitting state, the fit of the smart protective glasses is controlled to gradually change from loose to tight according to a preset adjustment step length, an adjustment change point is generated according to the change, and a plurality of facial contour features corresponding to a plurality of contact pressure distributions at the adjustment change points when the wearer wears the smart protective glasses are simultaneously obtained;

[0034] Generate a sequence by corresponding the plurality of contact pressure distributions to the plurality of facial contour features one by one, and establish a first mapping relationship between standard fitting facial contour features, contact pressure, and facial contour based on the sequence;

[0035] A multivariate regression analysis is performed on the contact pressure distributions and the corresponding facial contour features to obtain a sensitivity parameter matrix of the facial contour features as the contact pressure changes.

[0036] Furthermore, the calculating of the proportional coefficient set of the first mapping relationship and the second mapping relationship includes:

[0037] The deviation between the first mapping relationship and the second mapping relationship is multiplied by the sensitivity parameter matrix to obtain a set of proportional coefficients, and the set of proportional coefficients is input into the facial comfort mapping model to obtain a final set of proportional coefficients. The set of proportional coefficients includes proportional coefficients corresponding to nose pad adaptation parameters, proportional coefficients corresponding to temple opening and closing parameters, proportional coefficients corresponding to frame inclination parameters, and proportional coefficients corresponding to pressure pre-distribution parameters.

[0038] Furthermore, the facial comfort mapping model includes:

[0039] Collect historical data on the target users' proportional coefficient sets and subjective comfort evaluations;

[0040] The historical data of subjective comfort evaluation is divided by the first algorithm. The division result is a dataset with strong correlation with subjective comfort evaluation and a dataset with weak correlation with subjective comfort evaluation. Only the dataset with strong correlation with subjective comfort evaluation is retained.

[0041] Perform feature extraction on historical data through feature engineering to generate proportional coefficient set features and subjective comfort evaluation features, with part of the proportional coefficient set features and subjective comfort evaluation features used as a training set and part as a test set;

[0042] A deep learning network model is used to train a training set, wherein the input of the deep learning network model is the training set, and the output of the deep learning network model is a set of proportional coefficients;

[0043] Regularly update the parameters of the deep learning network model through user feedback data.

[0044] Furthermore, the first algorithm includes:

[0045] The subjective comfort evaluation strong correlation dataset and the subjective comfort evaluation weak correlation dataset are determined according to the following formula:

[0046] G(u,i)=∑ v∈S(u,K)∩N(i) t vi ,

[0047] Where G(u,i) represents the subjective comfort evaluation degree of target user u on subjective comfort evaluation history data i; S(u,K) represents the K related evaluation contents that are most similar to the target user's subjective comfort evaluation; N(i) represents the subjective comfort evaluation degree of target user that is most similar to the target user's subjective comfort evaluation history data i; t vi It represents the subjective comfort evaluation degree of the target user v on the subjective comfort evaluation history data i, that is, the score of the user target v on the visual comfort evaluation history data i; and when G(u,i)∈Q, it is determined that the visual comfort evaluation history data i belongs to the subjective comfort evaluation strong correlation data set; when When , it is determined that the visual comfort evaluation historical data i belongs to the subjective comfort evaluation weak correlation data set; wherein Q is the subjective comfort evaluation standard related to the threshold range of the first algorithm.

[0048] Compared with the prior art, the present invention achieves the following beneficial effects:

[0049] The present invention provides a method for obtaining three-dimensional scanning data of a wearer's face through a 3D vision module, determining initial fitting parameters based on the three-dimensional scanning data; obtaining a relative positional relationship between a pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters according to the relative positional relationship to obtain a first adjustment parameter; the relative positional relationship includes: a physical positional relationship between the pressure sensor array and the 3D vision module and a proportional relationship between pressure resolution and three-dimensional reconstruction accuracy; determining a relative angle relationship between a pressure sensor distribution plane and a frame mounting plane based on the mounting plane of the eyeglass frame, and performing mechanical compensation for the first adjustment parameter according to the relative angle relationship. Obtain a second adjustment parameter; establish a first mapping relationship between contact pressure distribution and facial contour and standard fitting facial contour features based on the wearer's historical data in a standard fitting state, monitor the contact pressure distribution and facial contour of the wearer's face and the eyeglass frame in real time, construct a second mapping relationship and real-time fitting facial contour features between the two, calculate a set of proportional coefficients between the first mapping relationship and the second mapping relationship, dynamically optimize the second adjustment parameter based on the proportional coefficient set to obtain a target adjustment parameter, and control the adjustment mechanism to execute the target adjustment parameter; the present invention improves the accuracy and comfort of facial fitting adjustment through multi-source data fusion, dynamic mechanical compensation and subjective comfort matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1The present invention provides a flow chart of a method for adjusting the fit of smart protective glasses to the face. DETAILED DESCRIPTION

[0051] The present invention will be described in detail below with reference to the accompanying drawings.

[0052] Example 1

[0053] The embodiment of the present invention provides a method for adjusting the fit of smart protective glasses to the face. Figure 1 , applied to smart protective glasses, the smart protective glasses including a pressure sensor array, a 3D vision module and an inertial measurement unit, the pressure sensor array and the 3D vision module being installed at different positions of the glasses frame; the method comprising:

[0054] S1. Obtaining three-dimensional facial scanning data of the wearer through a 3D vision module, and determining initial fitting parameters based on the three-dimensional scanning data;

[0055] S2. Obtaining a relative positional relationship between the pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters based on the relative positional relationship to obtain a first adjustment parameter; the relative positional relationship includes: a physical positional relationship between the pressure sensor array and the 3D vision module and a proportional relationship between pressure resolution and 3D reconstruction accuracy;

[0056] S3. Determine the relative angle between the pressure sensor distribution plane and the frame mounting plane based on the mounting plane of the eyeglass frame, and perform mechanical compensation on the first adjustment parameter according to the relative angle to obtain a second adjustment parameter;

[0057] S4. Establish a first mapping relationship between contact pressure distribution and facial contour and standard fitting facial contour features based on the wearer's historical data in a standard fitting state, monitor the contact pressure distribution and facial contour between the wearer's face and the eyeglass frame in real time, construct a second mapping relationship and real-time fitting facial contour features between the two, calculate a set of proportional coefficients between the first mapping relationship and the second mapping relationship, dynamically optimize the second adjustment parameter based on the proportional coefficient set to obtain a target adjustment parameter, and control the adjustment mechanism to execute the target adjustment parameter.

[0058] Specifically, the three-dimensional scanning data of the wearer's face is obtained through the 3D vision module, and the initial fitting parameters are determined based on the three-dimensional scanning data; the relative position relationship between the pressure sensor array and the 3D vision module is obtained, and the spatial coordinate transformation of the initial fitting parameters is performed on the relative position relationship to obtain the first adjustment parameter; the relative position relationship includes: the physical position relationship between the pressure sensor array and the 3D vision module and the proportional relationship between the pressure resolution and the three-dimensional reconstruction accuracy; the relative angle relationship between the pressure sensor distribution plane and the frame mounting plane is determined based on the mounting plane of the glasses frame, and the first adjustment parameter is mechanically compensated according to the relative angle relationship to obtain The invention obtains a second adjustment parameter; establishes a first mapping relationship between contact pressure distribution and facial contour and standard fitting facial contour features based on the wearer's historical data in a standard fitting state, monitors the contact pressure distribution and facial contour of the wearer's face and the eyeglass frame in real time, constructs a second mapping relationship and real-time fitting facial contour features between the two, calculates a set of proportional coefficients between the first mapping relationship and the second mapping relationship, dynamically optimizes the second adjustment parameter according to the set of proportional coefficients to obtain a target adjustment parameter, and controls the adjustment mechanism to execute the target adjustment parameter; the present invention improves the accuracy and comfort of facial fitting adjustment through multi-source data fusion, dynamic mechanical compensation and subjective comfort matching.

[0059] In the above embodiment, specifically, S1 includes:

[0060] Acquire three-dimensional facial scanning data of the wearer, and establish a three-dimensional facial point cloud model based on the three-dimensional facial scanning data;

[0061] Extract the nose bridge curvature features and ear pinna positioning features through the facial 3D point cloud model;

[0062] Calculate the spatial geometric relationship parameters between the nose bridge curvature features and the auricle positioning features;

[0063] The initial fitting parameters are determined by spatial geometric relationship parameters, and the initial fitting parameters include nose pad fitting parameters, temple opening and closing parameters, frame inclination parameters and pressure pre-distribution parameters; the nose pad fitting parameters characterize the curvature matching degree of the nose bridge contact surface; the temple opening and closing parameters characterize the matching of the auricle distance and the temple length; the frame inclination parameters characterize the angle between the frame plane and the facial coronal plane; and the pressure pre-distribution parameters characterize the expected pressure distribution of each contact area.

[0064] In the above embodiment, specifically, acquiring the relative positional relationship between the pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters according to the relative positional relationship to obtain the first adjustment parameters includes:

[0065] Obtaining a physical position offset between the pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters based on the physical position offset;

[0066] A proportional relationship between pressure resolution and three-dimensional reconstruction accuracy is obtained, and initial fitting parameters after spatial coordinate conversion are accurately matched according to the proportional relationship to obtain first adjustment parameters.

[0067] It should be noted that in this embodiment, the first adjustment parameter includes at least a nose pad displacement parameter and a temple opening and closing parameter;

[0068] It should be noted that the relative position relationship processing and parameter conversion steps between the pressure sensor array and the 3D vision module include:

[0069] Step 1: Get the physical position offset and perform space coordinate conversion

[0070] a. Calibrate the hardware coordinate system

[0071] Establish a world coordinate system with the optical center of the 3D vision module as the origin, and determine the local coordinate system of the pressure sensor array through the calibration plate;

[0072] Use a high-precision laser tracker or checkerboard calibration method to measure the rigid transformation parameters between the two coordinate systems: the translation vector is represented by t, and the rotation matrix is ​​represented by R;

[0073] b. Coordinate transformation calculation

[0074] The initial fitting parameters (such as the target fitting point coordinates p world ) to the sensor coordinate system:

[0075] p sensor =R -1 (p world -t),

[0076] If there is nonlinear distortion (such as lens distortion), it is necessary to perform interpolation correction using pre-calibrated distortion coefficients;

[0077] Step 2: Accuracy matching based on resolution ratio

[0078] a1. Resolution scale modeling

[0079] The resolution b1 of the pressure sensor array is expressed as force / unit area, and the 3D reconstruction accuracy b2 of the 3D vision is expressed as length units;

[0080] Fitting the scaling factor k through experimental data:

[0081] Under a known load, record the change in sensor reading ΔF and the change in displacement detected by visual inspection ΔD, and calculate:

[0082]

[0083] a2. Parameter accuracy matching

[0084] The transformed coordinate parameter p sensor Adjust by scale factor:

[0085] Force-displacement coupling: To map the visual positioning error to pressure control, adjust the fitting force F:

[0086]

[0087] Among them, F initial Indicates the initial setting of the fit strength;

[0088] Multimodal filtering: Use Kalman filter to fuse the data of two sensors and optimize the final output parameters.

[0089] In the above embodiment, specifically, obtaining the physical position offset between the pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters based on the physical position offset, includes:

[0090] Establish the global coordinate system of the eyeglass frame and the local coordinate system of the pressure sensor;

[0091] Determine the transformation matrix between the two coordinate systems based on the calibration parameters of the 3D vision module;

[0092] The initial fitting parameters are transformed from the global coordinate system to the local coordinate system through the transformation matrix.

[0093] It should be noted that the global coordinate system includes:

[0094] The origin is the center point of the nose bridge of the glasses frame; the X-axis is horizontal to the right (parallel to the plane of the frame); the Y-axis is vertically upward (perpendicular to the plane of the frame); and the Z-axis points directly in front of the wearer. The global coordinate system is calibrated using a laser tracker.

[0095] The local coordinate system includes:

[0096] Origin, the geometric center of the nose pad pressure sensor group; the axis is parallel to the X, Y, and Z of the global coordinate system;

[0097] In the above embodiment, specifically, obtaining a proportional relationship between pressure resolution and 3D reconstruction accuracy, and performing precision matching on the initial fitting parameters after spatial coordinate conversion to obtain the first adjustment parameter according to the proportional relationship, includes:

[0098] Calculate the ratio of the pressure sensor spatial resolution to the 3D vision module point cloud density;

[0099] Performing Newton interpolation algorithm compensation on the converted fitting parameters according to the ratio;

[0100] A Kalman filter is applied to smooth the compensation result of the Newton interpolation algorithm to obtain the first adjustment parameter.

[0101] In the above embodiment, specifically, determining the relative angle relationship between the pressure sensor distribution plane and the frame mounting plane based on the mounting plane of the eyeglass frame, and performing mechanical compensation on the first adjustment parameter according to the relative angle relationship to obtain the second adjustment parameter includes:

[0102] The spatial posture of the glasses frame is obtained through the inertial measurement unit to generate posture data;

[0103] Calculate the actual angle between the normal vector of the pressure sensor plane and the normal vector of the facial contact plane through the posture data;

[0104] performing vector decomposition on the pressure measurement value according to the actual angle;

[0105] Performing mechanical compensation on the first adjustment parameter based on the decomposition result to obtain a second adjustment parameter;

[0106] It should be noted that the second adjustment parameters include the nose pad curvature radius, the temple clamping force and the local skin elastic modulus.

[0107] In the above embodiment, specifically, establishing a first mapping relationship between contact pressure distribution and facial contour based on historical data of the wearer in a standard fitting state includes:

[0108] In a standard fitting state, the fit of the smart protective glasses is controlled to gradually change from loose to tight according to a preset adjustment step length, an adjustment change point is generated according to the change, and a plurality of facial contour features corresponding to a plurality of contact pressure distributions at the adjustment change points when the wearer wears the smart protective glasses are simultaneously obtained;

[0109] Generate a sequence by corresponding the plurality of contact pressure distributions to the plurality of facial contour features one by one, and establish a first mapping relationship between standard fitting facial contour features, contact pressure, and facial contour based on the sequence;

[0110] A multivariate regression analysis is performed on the contact pressure distributions and the corresponding facial contour features to obtain a sensitivity parameter matrix of the facial contour features as the contact pressure changes.

[0111] In the above embodiment, specifically, the calculating of the proportional coefficient set of the first mapping relationship and the second mapping relationship includes:

[0112] The deviation between the first mapping relationship and the second mapping relationship is multiplied by the sensitivity parameter matrix to obtain a set of proportional coefficients, and the set of proportional coefficients is input into the facial comfort mapping model to obtain a final set of proportional coefficients. The set of proportional coefficients includes proportional coefficients corresponding to nose pad adaptation parameters, proportional coefficients corresponding to temple opening and closing parameters, proportional coefficients corresponding to frame inclination parameters, and proportional coefficients corresponding to pressure pre-distribution parameters.

[0113] In the above embodiment, specifically, the facial comfort mapping model includes:

[0114] Collect historical data on the target users' proportional coefficient sets and subjective comfort evaluations;

[0115] The historical data of subjective comfort evaluation is divided by the first algorithm. The division result is a dataset with strong correlation with subjective comfort evaluation and a dataset with weak correlation with subjective comfort evaluation. Only the dataset with strong correlation with subjective comfort evaluation is retained.

[0116] Perform feature extraction on historical data through feature engineering to generate proportional coefficient set features and subjective comfort evaluation features, with part of the proportional coefficient set features and subjective comfort evaluation features used as a training set and part as a test set;

[0117] A deep learning network model is used to train a training set, wherein the input of the deep learning network model is the training set, and the output of the deep learning network model is a set of proportional coefficients;

[0118] Regularly update the parameters of the deep learning network model through user feedback data.

[0119] In the above embodiment, specifically, the first algorithm includes:

[0120] The subjective comfort evaluation strong correlation dataset and the subjective comfort evaluation weak correlation dataset are determined according to the following formula:

[0121] G(u,i)=∑ v∈S(u,K)∩N(i) t vi ,

[0122] Where G(u,i) represents the subjective comfort evaluation degree of target user u on subjective comfort evaluation history data i; S(u,K) represents the K related evaluation contents that are most similar to the target user's subjective comfort evaluation; N(i) represents the subjective comfort evaluation degree of target user that is most similar to the target user's subjective comfort evaluation history data i; t viIt represents the subjective comfort evaluation degree of the target user v on the subjective comfort evaluation history data i, that is, the score of the user target v on the visual comfort evaluation history data i; and when G(u,i)∈Q, it is determined that the visual comfort evaluation history data i belongs to the subjective comfort evaluation strong correlation data set; when When , it is determined that the visual comfort evaluation historical data i belongs to the subjective comfort evaluation weak correlation data set; wherein Q is the subjective comfort evaluation standard related to the threshold range of the first algorithm.

[0123] It should be noted that, through the comparison of technical effects, the effects of the initial fitting parameters, the first adjustment parameters, and the second adjustment parameters are compared as follows:

[0124]

[0125] Through two-level parameter conversion, progressive optimization from "geometric matching" to "mechanical adaptation" is achieved, which solves the problem of fitting failure caused by posture changes in traditional methods. This is also one of the technical highlights of the present invention.

[0126] It should be understood that the above embodiments are one or more embodiments of the present invention, and there are many other embodiments and variations thereof based on the present invention; the variations and modifications made by ordinary technicians in this industry through the present invention without making groundbreaking innovations all fall within the scope of protection of the present invention.

Claims

1. A method for adjusting the fit of smart protective glasses to the face, applied to the smart protective glasses, wherein the smart protective glasses include a pressure sensor array, a 3D vision module, and an inertial measurement unit, wherein the pressure sensor array and the 3D vision module are installed at different positions of the glasses frame; characterized in that: The method comprises: S1. Obtaining three-dimensional facial scanning data of the wearer through a 3D vision module, and determining initial fitting parameters based on the three-dimensional scanning data; S2. Obtaining a relative positional relationship between the pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters based on the relative positional relationship to obtain a first adjustment parameter; the relative positional relationship includes: a physical positional relationship between the pressure sensor array and the 3D vision module and a proportional relationship between pressure resolution and 3D reconstruction accuracy; S3. Determine the relative angle between the pressure sensor distribution plane and the frame mounting plane based on the mounting plane of the eyeglass frame, and perform mechanical compensation on the first adjustment parameter according to the relative angle to obtain a second adjustment parameter; S4. Establish a first mapping relationship between contact pressure distribution and facial contour and standard fitting facial contour features based on the wearer's historical data in a standard fitting state, monitor the contact pressure distribution and facial contour between the wearer's face and the eyeglass frame in real time, construct a second mapping relationship and real-time fitting facial contour features between the two, calculate a set of proportional coefficients between the first mapping relationship and the second mapping relationship, dynamically optimize the second adjustment parameter based on the proportional coefficient set to obtain a target adjustment parameter, and control the adjustment mechanism to execute the target adjustment parameter.

2. The method for adjusting the face fit of smart protective glasses according to claim 1, characterized in that: Said S1 comprises: Acquire three-dimensional facial scanning data of the wearer, and establish a three-dimensional facial point cloud model based on the three-dimensional facial scanning data; Extract the nose bridge curvature features and ear pinna positioning features through the facial 3D point cloud model; Calculate the spatial geometric relationship parameters between the nose bridge curvature features and the auricle positioning features; The initial fitting parameters are determined by spatial geometric relationship parameters, and the initial fitting parameters include nose pad fitting parameters, temple opening and closing parameters, frame inclination parameters and pressure pre-distribution parameters; the nose pad fitting parameters characterize the curvature matching degree of the nose bridge contact surface; the temple opening and closing parameters characterize the matching of the auricle distance and the temple length; the frame inclination parameters characterize the angle between the frame plane and the facial coronal plane; and the pressure pre-distribution parameters characterize the expected pressure distribution of each contact area.

3. The method for adjusting the face fit of smart protective glasses according to claim 1, characterized in that: The acquiring of the relative positional relationship between the pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters according to the relative positional relationship to obtain the first adjustment parameters, includes: Obtaining a physical position offset between the pressure sensor array and the 3D vision module, and performing spatial coordinate transformation on the initial fitting parameters based on the physical position offset; A proportional relationship between pressure resolution and three-dimensional reconstruction accuracy is obtained, and initial fitting parameters after spatial coordinate conversion are accurately matched according to the proportional relationship to obtain first adjustment parameters.

4. The method for adjusting the face fit of smart protective glasses according to claim 3, characterized in that: The obtaining of a physical position offset between the pressure sensor array and the 3D vision module, and performing a spatial coordinate transformation on the initial fitting parameters based on the physical position offset, includes: Establish the global coordinate system of the eyeglass frame and the local coordinate system of the pressure sensor; Determine the transformation matrix between the two coordinate systems based on the calibration parameters of the 3D vision module; The initial fitting parameters are transformed from the global coordinate system to the local coordinate system through the transformation matrix.

5. The method for adjusting the face fit of smart protective glasses according to claim 3, characterized in that: The obtaining of the proportional relationship between the pressure resolution and the three-dimensional reconstruction accuracy, and performing precision matching on the initial fitting parameters after the spatial coordinate conversion to obtain the first adjustment parameter according to the proportional relationship, includes: Calculate the ratio of the pressure sensor spatial resolution to the 3D vision module point cloud density; Performing Newton interpolation algorithm compensation on the converted fitting parameters according to the ratio; A Kalman filter is applied to smooth the compensation result of the Newton interpolation algorithm to obtain the first adjustment parameter.

6. The method for adjusting the face fit of smart protective glasses according to claim 1, characterized in that: The determining of the relative angle between the pressure sensor distribution plane and the frame mounting plane based on the mounting plane of the glasses frame, and performing mechanical compensation on the first adjustment parameter according to the relative angle to obtain the second adjustment parameter, includes: The spatial posture of the glasses frame is obtained through the inertial measurement unit to generate posture data; Calculate the actual angle between the normal vector of the pressure sensor plane and the normal vector of the facial contact plane through the posture data; performing vector decomposition on the pressure measurement value according to the actual angle; Based on the decomposition result, mechanical compensation is performed on the first adjustment parameter to obtain the second adjustment parameter.

7. The method for adjusting the face fit of smart protective glasses according to claim 1, characterized in that: The establishing of a first mapping relationship between contact pressure distribution and facial contour based on historical data of the wearer in a standard fitting state includes: In a standard fitting state, the fit of the smart protective glasses is controlled to gradually change from loose to tight according to a preset adjustment step length, an adjustment change point is generated according to the change, and a plurality of facial contour features corresponding to a plurality of contact pressure distributions at the adjustment change points when the wearer wears the smart protective glasses are simultaneously obtained; Generate a sequence by corresponding the plurality of contact pressure distributions to the plurality of facial contour features one by one, and establish a first mapping relationship between standard fitting facial contour features, contact pressure, and facial contour based on the sequence; A multivariate regression analysis is performed on the contact pressure distributions and the corresponding facial contour features to obtain a sensitivity parameter matrix of the facial contour features as the contact pressure changes.

8. The method for adjusting the face fit of smart protective glasses according to claim 7, characterized in that: The calculating of the proportional coefficient set of the first mapping relationship and the second mapping relationship includes: The deviation between the first mapping relationship and the second mapping relationship is multiplied by the sensitivity parameter matrix to obtain a set of proportional coefficients, and the set of proportional coefficients is input into the facial comfort mapping model to obtain a final set of proportional coefficients. The set of proportional coefficients includes proportional coefficients corresponding to nose pad adaptation parameters, proportional coefficients corresponding to temple opening and closing parameters, proportional coefficients corresponding to frame inclination parameters, and proportional coefficients corresponding to pressure pre-distribution parameters.

9. The method for adjusting the face fit of smart protective glasses according to claim 8, characterized in that: The facial comfort mapping model includes: Collect historical data on the target users' proportional coefficient sets and subjective comfort evaluations; The historical data of subjective comfort evaluation is divided by the first algorithm. The division result is a dataset with strong correlation with subjective comfort evaluation and a dataset with weak correlation with subjective comfort evaluation. Only the dataset with strong correlation with subjective comfort evaluation is retained. Perform feature extraction on historical data through feature engineering to generate proportional coefficient set features and subjective comfort evaluation features, with part of the proportional coefficient set features and subjective comfort evaluation features used as a training set and part as a test set; A deep learning network model is used to train a training set, wherein the input of the deep learning network model is the training set, and the output of the deep learning network model is a set of proportional coefficients; Regularly update the parameters of the deep learning network model through user feedback data.

10. The method for adjusting the face fit of smart protective glasses according to claim 9, characterized in that: The first algorithm includes: The subjective comfort evaluation strong correlation dataset and the subjective comfort evaluation weak correlation dataset are determined according to the following formula: G(u,i)=∑ v∈S(u,K)∩N(i) t vi , Where G(u,i) represents the subjective comfort evaluation degree of target user u on subjective comfort evaluation history data i; S(u,K) represents the K related evaluation contents that are most similar to the target user's subjective comfort evaluation; N(i) represents the subjective comfort evaluation degree of target user that is most similar to the target user's subjective comfort evaluation history data i; t vi It represents the subjective comfort evaluation degree of the target user v on the subjective comfort evaluation history data i, that is, the score of the user target v on the visual comfort evaluation history data i; and when G(u,i)∈Q, it is determined that the visual comfort evaluation history data i belongs to the subjective comfort evaluation strong correlation data set; when When , it is determined that the visual comfort evaluation historical data i belongs to the subjective comfort evaluation weak correlation data set; wherein Q is the subjective comfort evaluation standard related to the threshold range of the first algorithm.

Images