Lens edging system based on mirror frame physical deformation test and elastic modulus calibration

Abstract

The application provides a lens edging system based on mirror frame physical deformation test and elastic modulus calibration, relates to the technical field of spectacle lens processing, and reduces the fitting error rate of lens and mirror frame assembly by introducing a mirror frame physical deformation test and elastic modulus calibration mechanism. A micro-force test unit is used to apply controlled external force to the left and right directions and the up and down directions of the mirror frame, the actual elastic modulus of the mirror frame is calculated by measuring the deformation, a data fusion and prediction module compares the measured elastic modulus with the reference elastic modulus, obtains a material coefficient calibration amount, and generates an angle-differentiated radial compensation curve based on the calibration amount, so that the compensation parameters accurately reflect the elastic deformation characteristics of the individual mirror frame. Through the synergistic effect of dynamic threshold judgment and scale-invariant feature transformation matching algorithm of light compensation, adaptive and highly reliable edging quality inspection is realized, the accuracy and robustness of the edging quality inspection are improved, and reliable quality closed-loop control is provided for lens remote fitting.

Description

Technical Field

[0001] This invention relates to the field of spectacle lens processing technology, specifically to a lens edging system based on frame physical deformation testing and elastic modulus calibration. Background Technology

[0002] In the process of fitting eyeglasses, after the optical shop selects a frame based on the eye exam results, it needs to send the frame shape data to the lens grinding point. The grinding machine then grinds the lens blank to a precise profile that matches the frame so that the customer can assemble it themselves. However, even if different frames use the same material label (such as TR90, titanium alloy, etc.), their actual elastic modulus still varies individually. The elastic modulus of frames in the same batch can fluctuate by ±0.2GPa. This individual difference in material elasticity leads to two types of assembly problems between the ground lens and the frame: when the radial dimension of the lens is too large, the lens pressing into the frame will generate an assembly stress exceeding 0.5MPa, which can easily lead to frame deformation or even breakage after long-term use; when the radial dimension of the lens is too small, a gap greater than 0.1mm will appear between the lens and the frame, causing the lens to loosen, fall off, or shift its optical center. The existing technology has the following technical problems in use: Problem 1: Existing lens edging systems rely solely on the material label of the frame to predict edging compensation parameters, failing to provide differentiated calibration for individual elastic modulus differences between frames of the same batch and material. Because frame materials exhibit elasticity fluctuations between batches and individuals during actual production, the use of a uniform material coefficient in existing technologies leads to a mismatch between compensation parameters and the actual elastic deformation characteristics of the frame, severely impacting the safety and comfort of wearing eyeglasses. Problem 2: The existing lens edging system uses a fixed threshold to determine the fitting error in its quality inspection process. This makes it impossible to dynamically adjust the judgment criteria based on the actual elastic modulus of each individual frame, leading to misjudgments or omissions for frames with excessively large or small elastic moduli. At the same time, the lens edges are easily affected by reflections from ring light sources during optical inspection. Traditional scale-invariant feature transformation matching algorithms have a low success rate in feature point extraction and matching under reflective conditions, resulting in inaccurate fitting error calculations and reduced reliability of the inspection results. Summary of the Invention

[0003] To achieve the above objectives, the present invention provides the following technical solution: a lens edging system based on physical deformation testing and elastic modulus calibration of eyeglass frames, comprising an optical shop terminal system and a lens grinding point subsystem. The terminal system for optical shops includes: Intelligent scanning unit is used to scan the frame insert to obtain geometric parameters; The micro-force testing unit is used to apply controlled external force to the frame and measure the deformation. The measured elastic modulus is calculated based on the controlled external force, geometric parameters and deformation. The data fusion and prediction module is used to compare the measured elastic modulus with the reference elastic modulus to obtain the material coefficient calibration value, generate compensation parameters based on the material coefficient calibration value, generate target processing data based on geometric parameters and compensation parameters, and set dynamic threshold parameters based on the target processing data. The data encoding and transmission module is used to package and transmit the target processing data and dynamic threshold parameters to the grinding wheel processing point. The grinding wheel processing point system includes: The grinding machine control module controls the grinding machine to process lenses based on target processing data; The adaptive verification module is used to acquire actual images of the processed lens, perform feature matching with template images generated based on target processing data, and calculate the adaptation error. The dynamic threshold determination unit calculates the dynamic threshold based on the material coefficient calibration amount and determines whether the adaptation error meets the requirements of the dynamic threshold. If it does, the lens can be sent.

[0004] Furthermore, the scanning of the lens frame liner to obtain geometric parameters includes: The obtained geometric parameters include the radial distance of each sampling angle in the circumferential direction of the frame liner, the groove depth corresponding to each sampling angle, and the radius of curvature corresponding to each sampling angle, resulting in a radial distance sequence, a groove depth sequence, and a radius of curvature sequence. Continuous sampling is performed at equal angular intervals along the circumference of the frame insert. The radial distance from the center of the frame groove to the geometric center of the frame is recorded at each sampling angle. The vertical distance from the bottom to the top of the groove is measured at multiple feature points as the groove depth. The radius of curvature is obtained by fitting an arc based on the radial distance between adjacent sampling points.

[0005] Furthermore, the step of applying a controlled external force to the frame and measuring the deformation, and calculating the measured elastic modulus based on the controlled external force, geometric parameters, and deformation, includes: Preset external forces were applied to the left and right and up and down directions of the frame, and the measurements were repeated multiple times in each direction and the average value was taken to obtain the deformation in the left and right directions and the deformation in the up and down directions. The combined deformation variable is obtained by weighted averaging of the left-right and up-down directional deformation variables. The straight-line distance between the left and right posts of the frame is taken as the frame length, and the product of the frame groove width and the average groove depth is taken as the cross-sectional area. The measured elastic modulus is obtained by dividing the product of the control force and the frame length by the product of the cross-sectional area and the overall deformation.

[0006] Furthermore, the step of comparing the measured elastic modulus with the reference elastic modulus to obtain the material coefficient calibration value, and generating compensation parameters based on the material coefficient calibration value, includes: Obtain the measured elastic modulus, subtract the reference elastic modulus of the corresponding material label from the measured elastic modulus, and then divide by the reference elastic modulus to obtain the material coefficient calibration value; The basic predicted deformation is obtained based on the material label and pre-stored geometric parameters. The basic predicted deformation is corrected based on the material coefficient calibration amount and the preset global calibration coefficient to obtain the radial compensation curve with angle difference as the compensation parameter.

[0007] Furthermore, the radial compensation curve generation process includes: When the radial compensation curve is generated, an angle weighting factor is introduced. The angle weighting factor is obtained by multiplying the ratio of the radius of curvature of the current angle to the average radius of curvature by the ratio of the groove depth of the current angle to the average groove depth. The radial compensation amount of the current angle is obtained by combining the material coefficient calibration amount, the global calibration coefficient and the angle weight factor, and then multiplying them by the basic predicted deformation. The radial compensation amounts at each angle are combined to form a radial compensation curve.

[0008] Furthermore, the step of generating target machining data based on geometric parameters and compensation parameters, and setting dynamic threshold parameters based on the target machining data, includes: The radial distance sequence in the geometric parameters is superimposed with the radial compensation curve in the compensation parameters angle by angle to obtain the target radial distance sequence as the target machining data; The target processing data also includes frame material information, measured elastic modulus, and groove depth data; The target processing data and dynamic threshold parameters are organized into structured data packets by the data encoding and transmission module, encrypted, and then transmitted to the grinding disc processing point subsystem.

[0009] Furthermore, the process of controlling the grinding machine to process lenses based on target processing data includes: Generate rough grinding path and fine grinding path based on the target radial distance sequence in the target processing data; In the rough grinding stage, the grinding mill is controlled to leave a radial allowance for rapid grinding, and in the fine grinding stage, the grinding mill is controlled to precisely process to the radial dimension specified by the target radial distance sequence. After processing, the actual radial profile of the formed lens is measured and compared with the target radial distance sequence. When the deviation exceeds the preset range, compensation correction processing is triggered.

[0010] Furthermore, the actual image of the processed lens is acquired and matched with a template image generated based on the target processing data to calculate the adaptation error, including: When acquiring the edge image of the processed lens, a ring light source is set up and illuminated with the initial brightness value; Calculate the global average brightness of the acquired image, and trigger illumination compensation processing when the average brightness is higher than a first brightness threshold or lower than a second brightness threshold. The illumination compensation processing includes adaptive histogram equalization. The illumination compensation weight coefficient is determined based on the relationship between the global average brightness and the brightness threshold, and this weight coefficient is used to weight the subsequent feature matching results.

[0011] Furthermore, the step of performing feature matching with the template image generated based on the target processing data and calculating the adaptation error includes: Scale-invariant feature transformation key points are extracted from the lens edge image after illumination compensation processing, and the same type of feature key points are extracted from the template image generated based on the target processing data. Match the two sets of key points and select the best matching points. Calculate the average distance between all the best matching points and multiply the average distance by the camera calibration coefficient to convert it into the average adaptation error in the radial direction. The radial error corresponding to the maximum local matching distance is calculated as the maximum local error.

[0012] Furthermore, the calculation of the dynamic threshold based on the material coefficient calibration amount includes: Obtain a preset baseline threshold and weight coefficient, multiply the absolute value of the material coefficient calibration by the weight coefficient, and add it to the baseline threshold to obtain the dynamic threshold; The average adaptation error and the maximum local error are obtained. It is determined whether the average adaptation error is less than or equal to the dynamic threshold, and whether the maximum local error is less than or equal to a preset multiple of the dynamic threshold. If both conditions are met, the lens is determined to be ready for shipment. This invention provides a lens edging system based on physical deformation testing of the lens frame and calibration of its elastic modulus. It offers the following advantages: 1. This invention reduces the fitting error rate of lens and frame assembly by introducing a physical deformation test and elastic modulus calibration mechanism for the eyeglass frame. A micro-force testing unit applies controlled external forces to the left-right and up-down directions of the frame, measures the deformation, and calculates the actual elastic modulus of the frame. The data fusion and prediction module compares the measured elastic modulus with the benchmark elastic modulus to obtain the material coefficient calibration value. Based on this calibration value, a radial compensation curve with angular differences is generated, ensuring that the compensation parameters accurately reflect the elastic deformation characteristics of the individual frame. For various frames whose elastic modulus deviates from the benchmark value, the fitting error rate is significantly reduced compared to existing technologies. After assembly, all frames meet the standards for stress-free or gap-free assembly. Simultaneously, the fitting accuracy is improved from millimeters to micrometers, solving the assembly quality problem caused by individual elasticity differences in frame materials, and improving the reliability of eyeglass fitting and user comfort.

[0013] 2. This invention achieves adaptive and highly reliable edge grinding quality inspection through the synergistic effect of dynamic threshold determination and scale-invariant feature transform matching algorithm based on illumination compensation. The dynamic threshold determination unit calculates the dynamic threshold in real time based on the absolute value of the material coefficient calibration and preset weight coefficients, so that the judgment standard is automatically adjusted according to the actual elastic modulus of the lens frame, avoiding misjudgment or omission of the fixed threshold for the lens frame with an elastic modulus that is too high or too low. At the same time, when the adaptive inspection module acquires the lens edge image, it automatically triggers adaptive histogram equalization illumination compensation processing based on the global average brightness, and determines the illumination compensation weight coefficient according to the brightness threshold relationship, and performs weighted correction on the feature matching results. Under the condition of lens edge reflection, the success rate of scale-invariant feature transform feature matching is improved. The dynamic threshold significantly improves the judgment accuracy compared with the fixed threshold, effectively ensuring the accuracy and robustness of edge grinding quality inspection, and providing reliable quality closed-loop control for remote lens fitting. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the lens edging system based on the physical deformation test and elastic modulus calibration of the lens frame according to the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] like Figure 1 As shown, a lens edging system based on frame physical deformation testing and elastic modulus calibration is presented. The system includes an optical shop terminal system and a lens grinding point subsystem. The terminal system for eyewear stores includes an intelligent scanning unit, a micro-force testing unit, a data fusion and prediction module, and a data fusion and prediction module: Intelligent scanning unit is used to scan the frame insert to obtain geometric parameters; The micro-force testing unit is used to apply controlled external force to the frame and measure the deformation. The measured elastic modulus is calculated based on the controlled external force, geometric parameters and deformation. The data fusion and prediction module is used to compare the measured elastic modulus with the reference elastic modulus to obtain the material coefficient calibration value, generate compensation parameters based on the material coefficient calibration value, generate target processing data based on geometric parameters and compensation parameters, and set dynamic threshold parameters based on the target processing data. The data encoding and transmission module is used to package and transmit the target processing data and dynamic threshold parameters to the grinding wheel processing point. The grinding subsystem includes a grinding machine control module, an adaptive inspection module, a dynamic threshold determination unit, and a data receiving and decoding module. The grinding machine control module controls the grinding machine to process lenses based on target processing data; The adaptive verification module is used to acquire actual images of the processed lens, perform feature matching with template images generated based on target processing data, and calculate the adaptation error. The dynamic threshold determination unit calculates the dynamic threshold based on the material coefficient calibration amount and determines whether the adaptation error meets the requirements of the dynamic threshold. If it does, the lens can be sent.

[0017] In the eyewear shop terminal system, the intelligent scanning unit is responsible for collecting three types of basic data from the frame insert, i.e., the plastic or paper template of the frame, representing the actual shape of the frame groove: the radial distance at each angle along the circumference of the frame, the groove depth at the corresponding angle, and the radius of curvature. These data constitute the basic geometric parameters for all subsequent compensation calculations. The micro-force testing unit applies a controlled external force to the frame, measures the deformation displacement of the frame under this force, and combines it with the frame length and cross-sectional area parameters obtained from the intelligent scanning unit to reverse-calculate the actual elastic modulus of the frame material in the current batch. After receiving the above geometric parameters and the measured elastic modulus, the data fusion and prediction module first compares the measured value with the preset benchmark elastic modulus to calculate the material coefficient calibration amount, then generates compensation parameters based on the calibration amount, and finally superimposes the original geometric parameters and the compensation parameters to form the target processing data. The data encoding and transmission module packages and encrypts the target processing data and the dynamic threshold parameters required for subsequent inspection, and sends them to the grinding point subsystem. This effectively solves the assembly quality problem caused by individual elasticity differences in the frame material, and improves the reliability of eyeglass fitting and user wearing comfort.

[0018] The data receiving and decoding module in the lens grinding subsystem is responsible for parsing data packets. The lens grinding machine control module drives the grinding machine to complete the rough and fine grinding of the lens based on the target processing data. The adaptive inspection module performs optical inspection on the processed lens, calculating the fitting error between the actual lens contour and the target contour through image matching. The dynamic threshold judgment module calculates the appropriate acceptance threshold for the current frame in real time based on the material coefficient calibration amount, and determines whether the fitting error is within this threshold. Ultimately, it decides whether the lens can be sent to the customer, effectively ensuring the accuracy and robustness of the edge grinding quality inspection, and providing reliable quality closed-loop control for remote lens fitting.

[0019] In this embodiment, the process of the intelligent scanning unit acquiring geometric parameters from the scanning frame backing includes: The obtained geometric parameters include the radial distance of each sampling angle in the circumferential direction of the frame liner, the groove depth corresponding to each sampling angle, and the radius of curvature corresponding to each sampling angle, resulting in a radial distance sequence, a groove depth sequence, and a radius of curvature sequence. Continuous sampling is performed at equal angular intervals along the circumference of the frame insert. The radial distance from the center of the frame groove to the geometric center of the frame is recorded at each sampling angle. The vertical distance from the bottom to the top of the groove is measured at multiple feature points as the groove depth. The radius of curvature is obtained by fitting an arc based on the radial distance between adjacent sampling points.

[0020] During the scanning process, the intelligent scanning unit first fixes the eyeglass frame insert to a rotatable fixture, which rotates the insert at a constant speed for one revolution. For each fixed small angle rotation (e.g., every degree or half-degree), a laser displacement sensor emits a laser beam that illuminates the center of the insert groove. By measuring the time of the reflected light or using triangulation, the straight-line distance from the center point of the groove to the geometric center of the frame (usually the centroid of the innermost closed curve of the insert) at that angle is calculated. This distance is the radial distance for that angle. The radial distances for all angles are arranged in angular order to form a radial distance sequence. Simultaneously, during scanning, the sensor can also measure the vertical distance from the top to the bottom of the groove at several preset feature points (e.g., both sides of the bridge of the nose, near the temple posts, and the midpoint of the upper and lower edges of the frame). This value is the groove depth. Since the groove depth may differ at different angles, a groove depth sequence corresponding to each angle is obtained. For the radius of curvature, the system takes the radial distance between three adjacent sampling points, assuming these three points lie on the same arc. The radius of the arc is calculated using geometric methods (e.g., finding the intersection of the perpendicular bisectors of two chords), and this radius is used as the radius of curvature at the intermediate sampling point. This process is repeated for all sampling points to obtain a sequence of radius of curvature. These three sequences together constitute the complete geometric parameters of the frame and are stored in the cache of the data fusion and prediction module for subsequent calculations.

[0021] In this embodiment, the process of applying a controlled external force to the mirror frame using a micro-force testing unit and measuring the deformation, and then calculating the measured elastic modulus based on the controlled external force, geometric parameters, and deformation includes: Preset external forces were applied to the left and right and up and down directions of the frame, and the measurements were repeated multiple times in each direction and the average value was taken to obtain the deformation in the left and right directions and the deformation in the up and down directions. The combined deformation variable is obtained by weighted averaging of the left-right and up-down directional deformation variables. The straight-line distance between the left and right posts of the frame is taken as the frame length, and the product of the frame groove width and the average groove depth is taken as the cross-sectional area. The measured elastic modulus is obtained by dividing the product of the control force and the frame length by the product of the cross-sectional area and the overall deformation.

[0022] In the micro-force testing unit, the entire mirror frame is first fixed to a rigid platform during testing. In the left-right direction test, a miniature electric push rod applies a force to the right from the inside of the left end of the mirror frame, while a fixed stop is placed on the inside of the right end. A force sensor installed at the front end of the push rod measures the force in real time. When the force reaches a preset value (e.g., 0.5 Newtons), the push stops. At this point, a displacement sensor on the push rod records the distance the front end of the push rod has moved; this distance is the left-right deformation for that measurement. In the up-down direction test, a force is applied downwards from the midpoint of the upper edge, with a stop placed at the midpoint of the lower edge, and the deformation is measured similarly. Each direction is measured multiple times (e.g., three times), and the arithmetic mean is taken as the final deformation for that direction. Then, the system assigns a higher weight (e.g., 0.6) to the left-right deformation and a lower weight (e.g., 0.4) to the up-down deformation, and the weighted sum is the comprehensive deformation. The length of the mirror frame is obtained by the straight-line distance between the left and right ends acquired by the intelligent scanning unit. The cross-sectional area is calculated by multiplying the average groove depth (the average of all angular groove depths) measured by the intelligent scanning unit by the width of the frame groove (usually a standard value, but also obtainable through scanning). Finally, the applied thrust is multiplied by the frame length and then divided by the product of the cross-sectional area and the overall deformation to obtain the measured elastic modulus of the frame material. This elastic modulus is then transmitted to the data fusion and prediction module.

[0023] In this embodiment, the measured elastic modulus is compared with the reference elastic modulus using a data fusion and prediction module to obtain the material coefficient calibration value. The process of generating compensation parameters based on the material coefficient calibration value includes: Obtain the measured elastic modulus, subtract the reference elastic modulus of the corresponding material label from the measured elastic modulus, and then divide by the reference elastic modulus to obtain the material coefficient calibration value; The basic predicted deformation is obtained based on the material label and pre-stored geometric parameters. The basic predicted deformation is corrected based on the material coefficient calibration amount and the preset global calibration coefficient to obtain the radial compensation curve with angle difference as the compensation parameter.

[0024] The data fusion and prediction module first reads the baseline elastic modulus corresponding to the current frame material label (e.g., TR 90) from its internal database. This baseline elastic modulus is a standard value obtained through statistical testing of a large number of frames of the same material. The measured elastic modulus from the micro-force testing unit is subtracted from the baseline elastic modulus, and the difference is divided by the baseline elastic modulus; the resulting ratio is the material coefficient calibration value. This calibration value can be positive (indicating the measured elastic modulus is higher than the baseline, the frame is stiffer), negative (lower, the frame is softer), or zero (consistent with the baseline). Simultaneously, the data fusion and prediction module pre-stores a basic predicted deformation variable based on the material label and geometric parameters. This basic predicted deformation variable is the expected deformation variable for each angle calculated using traditional empirical formulas or statistical models without considering individual elasticity differences. The material coefficient calibration value is multiplied by a preset global calibration coefficient (obtained by fitting a large amount of experimental data and used to control the compensation intensity), then incremented by one, and multiplied again by the basic predicted deformation variable to obtain a new deformation variable sequence. This sequence is the radial compensation curve for angle differences. This curve reflects the additional radial machining dimensions that need to be added or reduced at each angle to compensate for individual differences in the elasticity of the frames.

[0025] In this embodiment, the radial compensation curve generation process includes: When generating the radial compensation curve, an angle weighting factor is introduced. The angle weighting factor is obtained by multiplying the ratio of the radius of curvature of the current angle to the average radius of curvature by the ratio of the groove depth of the current angle to the average groove depth. The radial compensation amount of the current angle is obtained by combining the material coefficient calibration amount, the global calibration coefficient and the angle weight factor, and then multiplying them by the basic predicted deformation. The radial compensation amounts at each angle are combined to form a radial compensation curve.

[0026] To improve the accuracy of compensation, the data fusion and prediction module introduces an angle weighting factor when calculating the radial compensation for each angle. This weighting factor is calculated as follows: First, the average radius of curvature (AGM) and the average groove depth (AGM) of all angles are calculated. For the specific angle currently being calculated, the radius of curvature is divided by the AGM, yielding a ratio; the groove depth is divided by the AGM, yielding another ratio. These two ratios are then multiplied to obtain the weighting factor for that angle. A weighting factor greater than one indicates a larger curvature or deeper groove at that angle, contributing more to elastic deformation and requiring stronger compensation; a weighting factor less than one indicates the opposite. Next, the previously calculated material coefficient calibration value, global calibration coefficient, and the weighting factor for that angle are multiplied to obtain a comprehensive coefficient. This comprehensive coefficient is then incremented by one, and finally multiplied by the basic predicted deformation for that angle to obtain the final radial compensation amount. This calculation is repeated for all angles to obtain the radial compensation amount for each angle. These compensation amounts are then arranged in angular order to form a complete radial compensation curve. This curve was then used to correct the original radial distance sequence.

[0027] In this embodiment, the data fusion and prediction module generates target processing data based on geometric parameters and compensation parameters. The process of setting dynamic threshold parameters based on the target processing data includes: The radial distance sequence from the geometric parameters is superimposed angle by angle with the radial compensation curve from the compensation parameters to obtain the target radial distance sequence as the target machining data. The target processing data also includes information on the frame material, measured elastic modulus, and groove depth. The target processing data and dynamic threshold parameters are organized into structured data packets by the data encoding and transmission module, encrypted, and then transmitted to the grinding disc processing point subsystem.

[0028] The data fusion and prediction module obtains the original radial distance sequence (i.e., the uncompensated ideal profile) from the intelligent scanning unit and the compensation amount (which can be positive or negative) for each angle from the radial compensation curve generated in the fifth segment. For each identical angle, the original radial distance is algebraically added to the compensation amount for that angle to obtain the final radial dimension that should be processed at that angle. The final radial dimensions of all angles are arranged in angular order to form the target radial distance sequence. This sequence is the basis for the actual processing of the grinding machine. In addition, the frame material information (such as material name), the measured elastic modulus value obtained by the micro-force testing unit, and the groove depth data obtained by the intelligent scanning unit are also included to form the target processing data package. Subsequently, the data encoding and transmission module organizes these target processing data and the dynamic threshold parameters (including the baseline threshold and weighting coefficients) required for subsequent inspection into a structured data package according to a predefined data structure. The data packet is processed by an encryption algorithm (such as symmetric encryption) and an appended checksum (such as a hash value) for verifying data integrity. It is then sent to the grinding wheel processing point in one of three optional ways: first, by transmitting directly via a Universal Serial Bus cable; second, by encoding the data into a high-density QR code, which is then scanned and read by the grinding wheel processing point; and third, by uploading it to a cloud server, where the grinding wheel processing point downloads it using the order number.

[0029] In this embodiment, the process of using the grinding machine control module to control the grinding machine to process lenses based on target processing data includes: Generate rough grinding and fine grinding paths based on the target radial distance sequence in the target processing data; In the rough grinding stage, the grinding mill is controlled to leave a radial allowance for rapid grinding, and in the fine grinding stage, the grinding mill is controlled to precisely process to the radial dimension specified by the target radial distance sequence. After processing, the actual radial profile of the formed lens is measured and compared with the target radial distance sequence. When the deviation exceeds the preset range, compensation correction processing is triggered.

[0030] The grinding machine control module, upon receiving the target radial distance sequence, first converts it into a processing path instruction. Considering processing efficiency and accuracy, the process is divided into two stages: rough grinding and fine grinding. In the rough grinding stage, a preset radial allowance (e.g., 0.1 mm) is added to the target radial distance, generating a rough grinding path slightly larger than the final profile. The grinding machine uses a coarser-grit wheel and a high feed rate to rapidly grind the lens from a rough shape to a state close to the final size but with a margin. In the fine grinding stage, the target radial distance sequence is directly used to generate the fine grinding path. The grinding machine switches to a finer-grit wheel and grinds precisely to the target radial dimension at a lower feed rate. After processing, an online thickness gauge (a non-contact or contact displacement sensor) scans along the circumference of the lens to measure the actual processed radial profile, obtaining a measured radial profile sequence. The measured radial profile is then compared point-by-point with the target radial distance sequence at the same angle, and the deviation at each angle is calculated. If the deviation of all angles is within a preset allowable range (e.g., ±0.005 mm), the processing is deemed qualified and proceeds to the next inspection stage; if the deviation of some angles exceeds the allowable range, but the deviation is small, a correction value will be calculated and a compensation correction processing will be automatically triggered, that is, fine grinding will be performed again to eliminate the deviation; if the deviation is too large, it will be judged as a defective product and an alarm will be triggered.

[0031] In this embodiment, the process of acquiring the actual image of the processed lens through the adaptive inspection module, performing feature matching with the template image generated based on the target processing data, and calculating the adaptation error includes: When acquiring the edge image of the processed lens, a ring light source is set up and illuminated with the initial brightness value; Calculate the global average brightness of the acquired image, and trigger illumination compensation processing when the average brightness is higher than a first brightness threshold or lower than a second brightness threshold. Illumination compensation processing includes adaptive histogram equalization.

[0032] The illumination compensation weight coefficient is determined based on the relationship between the global average brightness and the brightness threshold, and this weight coefficient is used to weight the subsequent feature matching results.

[0033] In the adaptive inspection module, during optical inspection, the finished lens is first placed on a rotatable platform with its edge facing a high-resolution industrial camera. A ring-shaped LED light source is installed around the camera, with its brightness set to an initial value (e.g., 150 levels). The camera captures an image of the lens edge and then calculates the arithmetic mean of the brightness of all pixels in the image to obtain the global average brightness. Two brightness thresholds are preset internally: a first threshold (e.g., 150) indicates overexposure, and a second threshold (e.g., 50) indicates underexposure. If the global average brightness is higher than the first threshold, the image is overexposed due to reflection; if it is lower than the second threshold, the lighting is insufficient. In both cases, lighting compensation processing is automatically triggered. This processing uses an adaptive histogram equalization method: first, the frequency of each brightness level in the image is statistically analyzed; then, the brightness distribution is stretched and smoothed, enhancing the overall image contrast, brightening dark details, and darkening bright details, thereby improving edge discernibility. Simultaneously, based on the specific value of the global average brightness, a lighting compensation weight coefficient is determined: if the image is too bright, this weight coefficient is set to less than one (e.g., 0.3) to reduce the matching confidence of overexposed areas; if the image is too dark, this weight coefficient is set to greater than one (e.g., 1.2) to enhance the importance of matching in dark areas; if the brightness is normal, the weight coefficient is set to one. This weight coefficient will be used to weight and correct the matching distance in subsequent feature matching steps.

[0034] In this embodiment, the process of performing feature matching with the template image generated based on the target processing data and calculating the adaptation error in the adaptive verification module includes: Scale-invariant feature transformation key points are extracted from the lens edge image after illumination compensation processing, and the same type of feature key points are extracted from the template image generated based on the target processing data. Match the two sets of key points and select the best matching points. Calculate the average distance between all the best matching points and multiply the average distance by the camera calibration coefficient to convert it into the average adaptation error in the radial direction. The radial error corresponding to the maximum local matching distance is calculated as the maximum local error.

[0035] After completing the illumination compensation process, the adaptive verification module performs scale-invariant feature transformation (SIN) feature extraction on both the lens edge image and the template image. This process first detects extreme points in the image at different scales, then selects stable keypoints and assigns an orientation (based on a local gradient orientation histogram) to each keypoint. Finally, it generates a 128-dimensional feature descriptor vector to uniquely identify the local texture information of the image surrounding the keypoint. For the lens edge image, all extracted keypoints and their descriptors constitute the actual feature point set; for the template image (which is an ideal lens edge contour map generated in advance using computer graphics methods based on the target radial distance sequence), a template feature point set is also extracted. A fast nearest neighbor matching library (an approximate nearest neighbor search algorithm) is used to match the two feature point sets: for each feature point in the actual image, the two points in the template feature point set with the closest Euclidean distance to its descriptor are found. If the ratio of the closest distance to the second closest distance is less than a preset ratio threshold (e.g., 0.75), it is considered a good match. After selecting all excellent matching point pairs, the Euclidean distance between each pair of matching points in the image coordinate system is calculated. Then, the average distance of all matching point pairs is calculated to obtain the average matching distance. Since the camera and lens have been pre-calibrated (i.e., it is known how many millimeters one pixel in the image represents in the real world), this average matching distance is multiplied by a calibration coefficient to convert it into the average radial adaptation error in actual physical space. Simultaneously, the pair with the largest distance among all matching point pairs is identified, and this largest distance is also converted into physical dimensions to obtain the maximum local error. The average adaptation error reflects the degree of fit between the overall lens contour and the target contour, while the maximum local error reflects whether there are any serious local deviations.

[0036] In this embodiment, the process of calculating the dynamic threshold based on the material coefficient calibration using the dynamic threshold determination unit includes: Obtain the preset baseline threshold and weight coefficient, multiply the absolute value of the material coefficient calibration by the weight coefficient, and add it to the baseline threshold to obtain the dynamic threshold; The system obtains the average adaptation error and the maximum local error, determines whether the average adaptation error is less than or equal to the dynamic threshold, and whether the maximum local error is less than or equal to a preset multiple of the dynamic threshold. If both conditions are met, the lens is deemed ready for shipment.

[0037] The dynamic threshold determination unit first obtains the preset reference threshold and weight coefficient from the data packet. The reference threshold is an empirical value determined based on the upper limit of the allowable error of the material frame under ideal conditions (for example, 0.02 mm), and the weight coefficient is an adjustment factor (for example, 0.01) used to control the influence degree of the calibration amount of the material coefficient on the threshold. The dynamic threshold determination unit multiplies the absolute value of the previously calculated calibration amount of the material coefficient by the weight coefficient, and then adds the product to the reference threshold to obtain the dynamic threshold specific to the current frame. This dynamic threshold increases or remains unchanged as the deviation degree between the measured elastic modulus and the reference value increases, so that frames with greater elastic differences obtain a more lenient judgment criterion, thus avoiding misjudging qualified products as unqualified due to excessive strictness. Next, the calculated average adaptation error and the maximum local error are obtained, and it is judged whether the average adaptation error is less than or equal to the dynamic threshold, and then it is judged whether the maximum local error is less than or equal to 1.5 times the dynamic threshold (this multiple is also preset, allowing slightly larger local deviations). Only when both of these conditions are met, the lens is judged to be a qualified product, and the result output unit is triggered to print a qualified label, which includes the order number, inspection time, and error value, and is sent to the customer together with the lens. If either condition is not met, it is judged as unqualified, and the unit sends the error distribution information to the display screen of the grinding machine, and displays the deviation magnitude of each angle in the form of a heat map, prompting the operator to adjust the compensation parameters according to the deviation situation and reprocess the lens.

[0038] In this embodiment, by introducing a frame physical deformation test and elastic modulus calibration mechanism, the adaptation error rate of the lens and frame assembly is reduced. The micro-force test unit applies controlled external forces to the frame in the left-right direction and the up-down direction respectively, measures the actual deformation amount, and calculates the actual elastic modulus of the frame by inverse calculation; the data fusion and prediction module compares the measured elastic modulus with the reference elastic modulus to obtain the calibration amount of the material coefficient, and generates a radially compensated curve with angle differentiation based on this calibration amount, so that the compensation parameters accurately reflect the elastic deformation characteristics of the individual frame. For various frames with elastic modulus deviating from the reference value, the adaptation error rate is greatly reduced compared with the prior art, and all stress-free or gap-free after assembly meet the standards. At the same time, the adaptation accuracy is improved from the millimeter level to the micron level, effectively solving the assembly quality problem caused by the individual elastic differences of the frame materials, and improving the reliability of glasses dispensing and the wearing comfort of users.

[0039] By combining dynamic threshold judgment with a scale-invariant feature transform matching algorithm based on illumination compensation, an adaptive and highly reliable edge grinding quality inspection is achieved. The dynamic threshold judgment unit calculates the dynamic threshold in real time based on the absolute value of the material coefficient calibration and preset weighting coefficients. This allows the judgment standard to automatically adjust according to the actual elastic modulus of the lens frame, avoiding misjudgments or omissions caused by fixed thresholds for frames with excessively high or low elastic moduli. Simultaneously, the adaptive inspection module automatically triggers adaptive histogram equalization illumination compensation processing based on the global average brightness when acquiring lens edge images. It also determines the illumination compensation weighting coefficients based on the brightness threshold relationship, weighting and correcting the feature matching results. Under reflective conditions at the lens edge, the success rate of scale-invariant feature transform matching is improved. The dynamic threshold significantly enhances the judgment accuracy compared to a fixed threshold, effectively ensuring the accuracy and robustness of edge grinding quality inspection and providing reliable quality closed-loop control for remote lens fitting.

[0040] This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable code that, when executed by the one or more processors, can perform the lens edging system described above, based on frame physical deformation testing and elastic modulus calibration.

[0041] The methods or systems according to embodiments of this application can also be implemented using the architecture of the electronic device shown in this application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. The storage device in the electronic device, such as a ROM or hard disk, may store the lens edging system based on frame physical deformation testing and elastic modulus calibration provided in this application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in this application is merely exemplary; when implementing different devices, one or more components in the electronic device shown in this application may be omitted according to actual needs.

[0042] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0043] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various 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 lens edging system based on physical deformation testing and elastic modulus calibration of the lens frame, characterized in that, The system includes an optical shop terminal system and a lens grinding point subsystem: The optical shop terminal system includes: Intelligent scanning unit is used to scan the frame insert to obtain geometric parameters; The micro-force testing unit is used to apply controlled external force to the frame and measure the deformation. The measured elastic modulus is calculated based on the controlled external force, geometric parameters and deformation. The data fusion and prediction module is used to compare the measured elastic modulus with the reference elastic modulus to obtain the material coefficient calibration value, generate compensation parameters based on the material coefficient calibration value, generate target processing data based on geometric parameters and compensation parameters, and set dynamic threshold parameters based on the target processing data. The data encoding and transmission module is used to package and transmit the target processing data and dynamic threshold parameters to the grinding wheel processing point. The grinding wheel processing point subsystem includes: The grinding machine control module controls the grinding machine to process lenses based on target processing data; The adaptive verification module is used to acquire the actual image of the processed lens, perform feature matching with the template image generated based on the target processing data, and calculate the adaptation error. The dynamic threshold determination unit calculates the dynamic threshold based on the material coefficient calibration amount and determines whether the adaptation error meets the requirements of the dynamic threshold. If it does, the lens can be sent.

2. The lens edging system based on frame physical deformation testing and elastic modulus calibration according to claim 1, characterized in that, The scanning of the lens frame liner to obtain geometric parameters includes: The obtained geometric parameters include the radial distance of each sampling angle in the circumferential direction of the frame liner, the groove depth corresponding to each sampling angle, and the radius of curvature corresponding to each sampling angle, resulting in a radial distance sequence, a groove depth sequence, and a radius of curvature sequence. Continuous sampling is performed at equal angular intervals along the circumference of the frame insert. The radial distance from the center of the frame groove to the geometric center of the frame is recorded at each sampling angle. The vertical distance from the bottom to the top of the groove is measured at multiple feature points as the groove depth. The radius of curvature is obtained by fitting an arc based on the radial distance between adjacent sampling points.

3. The lens edging system based on frame physical deformation testing and elastic modulus calibration according to claim 1, characterized in that, The process of applying a controlled external force to the frame and measuring its deformation, and calculating the measured elastic modulus based on the controlled external force, geometric parameters, and deformation, includes: Preset external forces were applied to the left and right and up and down directions of the frame, and the measurements were repeated multiple times in each direction and the average value was taken to obtain the deformation in the left and right directions and the deformation in the up and down directions. The combined deformation variable is obtained by weighted averaging of the left-right and up-down directional deformation variables. The straight-line distance between the left and right posts of the frame is taken as the frame length, and the product of the frame groove width and the average groove depth is taken as the cross-sectional area. The measured elastic modulus is obtained by dividing the product of the control force and the frame length by the product of the cross-sectional area and the overall deformation.

4. The lens edging system based on frame physical deformation testing and elastic modulus calibration according to claim 3, characterized in that, The step of comparing the measured elastic modulus with the reference elastic modulus to obtain the material coefficient calibration value, and generating compensation parameters based on the material coefficient calibration value, includes: Obtain the measured elastic modulus, subtract the reference elastic modulus of the corresponding material label from the measured elastic modulus, and then divide by the reference elastic modulus to obtain the material coefficient calibration value; The basic predicted deformation is obtained based on the material label and pre-stored geometric parameters. The basic predicted deformation is corrected based on the material coefficient calibration amount and the preset global calibration coefficient to obtain the radial compensation curve with angle difference as the compensation parameter.

5. The lens edging system based on frame physical deformation testing and elastic modulus calibration according to claim 4, characterized in that, The radial compensation curve generation process includes: When the radial compensation curve is generated, an angle weighting factor is introduced. The angle weighting factor is obtained by multiplying the ratio of the radius of curvature of the current angle to the average radius of curvature by the ratio of the groove depth of the current angle to the average groove depth. The radial compensation amount of the current angle is obtained by combining the material coefficient calibration amount, the global calibration coefficient and the angle weight factor, and then multiplying them by the basic predicted deformation. The radial compensation amounts at each angle are combined to form a radial compensation curve.

6. The lens edging system based on frame physical deformation testing and elastic modulus calibration according to claim 5, characterized in that, The step of generating target machining data based on geometric parameters and compensation parameters, and setting dynamic threshold parameters based on the target machining data, includes: The radial distance sequence in the geometric parameters is superimposed with the radial compensation curve in the compensation parameters angle by angle to obtain the target radial distance sequence as the target machining data; The target processing data also includes frame material information, measured elastic modulus, and groove depth data; The target processing data and dynamic threshold parameters are organized into structured data packets by the data encoding and transmission module, encrypted, and then transmitted to the grinding disc processing point subsystem.

7. The lens edging system based on frame physical deformation testing and elastic modulus calibration according to claim 1, characterized in that, The process of controlling the grinding machine to process lenses based on target processing data includes: Generate rough grinding path and fine grinding path based on the target radial distance sequence in the target processing data; In the rough grinding stage, the grinding mill is controlled to leave a radial allowance for rapid grinding, and in the fine grinding stage, the grinding mill is controlled to precisely process to the radial dimension specified by the target radial distance sequence. After processing, the actual radial profile of the formed lens is measured and compared with the target radial distance sequence. When the deviation exceeds the preset range, compensation correction processing is triggered.

8. The lens edging system based on frame physical deformation testing and elastic modulus calibration according to claim 7, characterized in that, The actual image of the processed lens is acquired and matched with a template image generated based on the target processing data to calculate the adaptation error, including: When acquiring the edge image of the processed lens, a ring light source is set up and illuminated with the initial brightness value; Calculate the global average brightness of the acquired image, and trigger illumination compensation processing when the average brightness is higher than a first brightness threshold or lower than a second brightness threshold. The illumination compensation processing includes adaptive histogram equalization. The illumination compensation weight coefficient is determined based on the relationship between the global average brightness and the brightness threshold, and this weight coefficient is used to weight the subsequent feature matching results.

9. The lens edging system based on frame physical deformation testing and elastic modulus calibration according to claim 8, characterized in that, The step of performing feature matching with the template image generated based on the target processing data and calculating the adaptation error includes: Scale-invariant feature transformation key points are extracted from the lens edge image after illumination compensation processing, and the same type of feature key points are extracted from the template image generated based on the target processing data. Match the two sets of key points and select the best matching points. Calculate the average distance between all the best matching points and multiply the average distance by the camera calibration coefficient to convert it into the average adaptation error in the radial direction. The radial error corresponding to the maximum local matching distance is calculated as the maximum local error.

10. The lens edging system based on frame physical deformation testing and elastic modulus calibration according to claim 9, characterized in that, The calculation of the dynamic threshold based on the material coefficient calibration includes: Obtain a preset baseline threshold and weight coefficient, multiply the absolute value of the material coefficient calibration by the weight coefficient, and add it to the baseline threshold to obtain the dynamic threshold; The average adaptation error and the maximum local error are obtained. It is determined whether the average adaptation error is less than or equal to the dynamic threshold and whether the maximum local error is less than or equal to a preset multiple of the dynamic threshold. When both conditions are met, the lens is determined to be ready for shipment.

Images