Shadowless lamp quality control system based on digital processing

Abstract

The application provides a shadowless lamp quality control system based on digital processing, and relates to the technical field of intelligent manufacturing, comprising: a polishing index division module; a polishing area division module; a polishing quality simulation detection module; a quality deviation analysis module; a polishing parameter optimization module, which is used for polishing parameter optimization based on a polishing processing twin model and a predetermined polishing evaluation function to meet the polishing workload of the second stage, generate first optimized polishing parameters; a polishing control module, which is used for executing polishing control of the second stage according to the second starting node and the first optimized polishing parameters. Through the application, the technical problem that the overall polishing precision and quality are poor and the production quality of the surgical shadowless lamp product is poor due to the inability to accurately detect and timely correct the processing defects in the mirror polishing process can be solved, the polishing processing defects can be effectively corrected in a targeted manner in time, and the polishing precision and quality of the mirror are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent manufacturing, and particularly relates to a shadowless lamp quality control system based on digital processing. BACKGROUND

[0002] The whole reflection type surgical shadowless lamp is a surgical lighting device adopting a whole reflection optical system, which can provide uniform and shadowless illumination to help doctors clearly see the surgical site during the operation. The whole reflection mirror is the core component of the surgical shadowless lamp, which is composed of one or more mirrors and is used for reflecting and focusing the light source to provide uniform illumination of the surgical area.

[0003] At present, when the existing whole reflection type surgical shadowless lamp is produced and processed, the machining defects occurring in the polishing process of the reflection mirror cannot be accurately detected and timely corrected, which leads to poor polishing precision and polishing quality of the whole reflection mirror, and finally causes the technical problem of poor production quality of the surgical shadowless lamp. SUMMARY

[0004] The purpose of the present application is to provide a shadowless lamp quality control system based on digital processing, which solves the technical problem that the machining defects occurring in the polishing process of the reflection mirror cannot be accurately detected and timely corrected when the existing whole reflection type surgical shadowless lamp is produced and processed, which leads to poor polishing precision and polishing quality of the whole reflection mirror, and finally causes the technical problem of poor production quality of the surgical shadowless lamp.

[0005] In view of the above problems, the present application provides a shadowless lamp quality control system based on digital processing, wherein the system comprises: a polishing index division module for reading the polishing index and polishing process of a target mirror, determining a plurality of polishing stages based on the polishing process, dividing the polishing index according to the plurality of polishing stages, and obtaining a set of stage polishing indexes; a polishing area division module for dividing the mirror to be polished into a plurality of polishing sub-areas based on the single polishing area of the grinding wheel, wherein each polishing sub-area has a unique serial number; a polishing quality simulation detection module for performing a polishing operation on a first sub-area based on a predetermined polishing path, collecting an image of the first sub-area through an optical microscopic device after the operation is completed, performing polishing quality simulation detection based on the microscopic image of the first sub-area, and obtaining a first stage detection result, wherein the first sub-area is any one of the plurality of polishing sub-areas, and the polishing operation is the first stage polishing operation; a quality deviation analysis module for selecting a second stage polishing index from the set of stage polishing indexes, performing quality deviation analysis on the first stage detection result according to the second stage polishing index, and setting the deviation analysis result as the second stage polishing workload; a polishing parameter optimization module for performing polishing parameter optimization based on a polishing processing twin model and a predetermined polishing evaluation function to meet the second stage polishing workload, and generating first optimized polishing parameters; and a polishing control module for determining a second starting node according to the first serial number of the first sub-area, feeding the second starting node and the first optimized polishing parameters to a polishing control unit, and performing the second stage polishing control of the first sub-area until the entire polishing process is completed, wherein the second starting node is a starting time node of the second stage.

[0006] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0007] By reading the polishing index and polishing process of the target mirror, a plurality of polishing stages are determined based on the polishing process, and the polishing index is divided according to the plurality of polishing stages to obtain a set of stage polishing indexes; then, taking the single polishing area of the grinding wheel as the reference, the area of the mirror to be polished is divided to determine a plurality of polishing sub-areas, wherein each polishing sub-area has a unique serial number mark; further based on the predetermined polishing path, a one-time polishing job is performed on the first sub-area, and after the job is completed, the first sub-area is imaged by an optical microscopic device, the polishing quality simulation detection is performed based on the first sub-area microscopic image, and the first stage detection result is obtained, wherein the first sub-area is any one of the plurality of polishing sub-areas, and the one-time polishing job is the first stage polishing job; then, the second stage polishing index is selected in the set of stage polishing indexes, the quality deviation analysis is performed on the first stage detection result according to the second stage polishing index, and the deviation analysis result is set as the second stage polishing workload; then, based on the polishing processing twin model and the predetermined polishing evaluation function, the polishing parameter optimization is performed to generate the first optimized polishing parameter, with the purpose of meeting the second stage polishing workload; finally, the second starting node is determined according to the first serial number of the first sub-area, the second starting node and the first optimized polishing parameter are fed back to the polishing control unit, and the first sub-area second stage polishing control is performed until the entire polishing process is completed, wherein the second starting node is the starting time node of the second stage. That is, by dividing the polishing index according to a plurality of polishing stages to obtain a set of stage polishing indexes, on the other hand, taking the single polishing area of the grinding wheel as the reference, the area of the mirror to be polished is divided to determine a plurality of polishing sub-areas; then, after the one-time polishing job is completed, the polishing quality simulation detection is performed based on the first sub-area microscopic image to obtain the first stage detection result, then the quality deviation analysis is performed on the first stage detection result according to the second stage polishing index, and the deviation analysis result is set as the second stage polishing workload; further based on the polishing processing twin model and the predetermined polishing evaluation function, the polishing parameter optimization is performed to generate the first optimized polishing parameter, with the purpose of meeting the second stage polishing workload; finally, the second starting node is determined based on the first serial number of the first sub-area, and the first sub-area second stage polishing control is performed according to the second starting node and the first optimized polishing parameter until the entire polishing process is completed. The precision and accuracy of the polishing defect detection in the mirror polishing process can be improved, so that the processing defects can be timely and effectively corrected, the mirror polishing precision and quality can be improved, and the technical effect of meeting the production requirements of the surgical shadowless lamp product can be achieved.

[0008] The above description is only a summary of the technical solutions of the present application. In order to enable the technical means of the present application to be more clearly understood, and to be implemented according to the content of the description, and in order to enable the above and other purposes, characteristics and advantages of the present application to be more apparent and easy to understand, the following specific embodiments of the present application are described. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only exemplary, and other drawings can be obtained by the provided drawings without creative labor for those skilled in the art.

[0010] Figure 1 A structure schematic diagram of a shadowless lamp quality control system based on digital processing according to the present application;

[0011] Figure 2 A flowchart for obtaining a first stage detection result in a shadowless lamp quality control system based on digital processing according to the present application.

[0012] BRIEF DESCRIPTION OF DRAWINGS

[0013] Polishing index division module 11, polishing area division module 12, polishing quality simulation detection module 13, quality deviation analysis module 14, polishing parameter optimization module 15, polishing control module 16. DETAILED DESCRIPTION

[0014] The present application provides a shadowless lamp quality control system based on digital processing, which solves the technical problem that the overall reflection type surgical shadowless lamp production and processing cannot accurately detect and timely correct the processing defects in the mirror polishing process, resulting in poor mirror overall polishing precision and polishing quality, and ultimately causing poor production quality of surgical shadowless lamp products. The present application can improve the fineness and accuracy of polishing defect detection in the mirror polishing process, so that the processing defects can be effectively corrected in a targeted manner in a timely manner, thereby improving the mirror polishing precision and polishing quality, and meeting the production requirements of surgical shadowless lamp products.

[0015] Below, the technical solutions in the present application will be described clearly and completely with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application. In addition, it should be noted that, for the convenience of description, only parts related to the present application are shown in the drawings, not all.

[0016] Embodiments, please refer to the attached Figure 1 The present application provides a shadowless lamp quality control system based on digital processing, comprising:

[0017] The polishing index division module 11 is configured to read the polishing index and polishing process of the target mirror, determine a plurality of polishing stages based on the polishing process, divide the polishing index according to the plurality of polishing stages, and obtain a set of stage polishing indexes.

[0018] Specifically, the overall reflection type surgical shadowless lamp is a surgical lighting device using an overall reflection optical system, which can provide uniform and shadowless illumination effect. The overall reflection mirror is the core component of the surgical shadowless lamp. The present application is used to optimize the polishing process of the overall reflection mirror to improve the polishing precision and quality of the mirror. First, the polishing index and polishing process of the target mirror are read. The polishing index refers to the expected processing index of the target mirror, including surface roughness, reflectivity, and microscopic defect constraints, which can be set according to actual conditions. The polishing process refers to the entire polishing process of the polishing technology of the target mirror. Then, according to the characteristics and experience of the polishing technology, the key steps and stages in the polishing process are determined, which include rough polishing, medium polishing, fine polishing, etc.

[0019] Further, the polishing index is divided according to the plurality of polishing stages, i.e. according to the expected target of each polishing stage, the polishing index is divided to obtain the stage polishing index corresponding to each polishing stage, i.e. each polishing stage has a specific set of polishing indexes, which collectively define the polishing target of the stage, and a set of stage polishing indexes is obtained.

[0020] The polishing area division module 12 is configured to divide the to-be-polished mirror into a plurality of polishing sub-areas based on the single polishing area of the grinding wheel, wherein each polishing sub-area has a unique serial number.

[0021] Specifically, first, according to the size of the grinding wheel and the polishing process, the area of the mirror that can be covered by the grinding wheel in one polishing operation is determined, which can be circular, square or other geometric shapes, depending on the design of the grinding wheel and the polishing requirements. Then, based on the single polishing area of the grinding wheel, the area of the mirror to be polished is divided into multiple equal polishing sub-areas to ensure that each sub-area can be fully covered by the grinding wheel, and the size and shape of each polishing sub-area should match the single polishing area of the grinding wheel. Finally, each polishing sub-area is assigned a unique serial number or label for identification and management during the polishing process, which can be numbers, letters or symbols to ensure that each sub-area has a clear identification. Through this method, the uniformity and efficiency of the polishing process can be ensured, where each sub-area is divided according to the single polishing area of the grinding wheel, so that each area can be properly polished to improve the polishing quality of the mirror.

[0022] The polishing quality simulation detection module 13 is configured to perform a first polishing operation on the first sub-area based on the predetermined polishing path, and after the operation is completed, an image of the first sub-area is collected by an optical microscopic device, and a polishing quality simulation detection is performed based on the microscopic image of the first sub-area to obtain a first stage detection result, wherein the first sub-area is any one of the plurality of polishing sub-areas, and the first polishing operation is the first stage polishing operation.

[0023] Specifically, first, according to the predetermined polishing path and process parameters, a first polishing operation is performed on the first sub-area, which is any one of the plurality of polishing sub-areas, and the first polishing operation is the first stage polishing operation, and the first stage is the first processing stage in the plurality of polishing stages. Then, after the first polishing operation of the first sub-area is completed, an image of the polished first sub-area is collected using an optical microscopic device, wherein by selecting appropriate magnification and illumination conditions, the details of the polished area can be clearly observed to obtain a microscopic image of the first sub-area.

[0024] Further, the first sub-region microscopic image is detected for polishing quality simulation by artificial intelligence methods such as convolutional neural network, BP neural network and digital twin, for example, a convolutional neural network model is trained for surface feature extraction; a polishing state is predicted based on a digital twin model; and a first-stage detection result is obtained, which includes a predicted surface roughness, a predicted reflectivity and a predicted microscopic defect feature after the first-stage processing. By using artificial intelligence methods such as convolutional neural network, BP neural network and digital twin to detect the first sub-region microscopic image for polishing quality simulation, the intelligence and automation of polishing quality detection can be improved, and the efficiency and accuracy of detection can be improved, and defects in the polishing process can be found in time to provide a basis for subsequent polishing optimization control.

[0025] The quality deviation analysis module 14 is configured to select a second-stage polishing indicator from the stage polishing indicators, perform quality deviation analysis on the first-stage detection result according to the second-stage polishing indicator, and set the deviation analysis result as the second-stage polishing workload.

[0026] Specifically, first, a second-stage polishing indicator is selected from the stage polishing indicators, where the second stage is the next polishing stage adjacent to the first stage; then, quality deviation is calculated, for example, surface roughness deviation calculation, on the first-stage detection result according to the second-stage polishing indicator; a deviation calculation result is obtained, where the deviation calculation result is the difference between the second-stage polishing indicator and the first-stage detection result; and then, the deviation analysis result is set as the second-stage polishing workload. By setting the polishing workload of the next stage based on the deviation calculation and analysis result, the processing defects of the previous polishing stage can be corrected in time and effectively, so as to realize accurate control and optimization of the overall polishing process and improve the accuracy and polishing quality of the overall polishing.

[0027] The polishing parameter optimization module 15 is configured to perform polishing parameter optimization based on a polishing processing twin model and a predetermined polishing evaluation function to meet the second-stage polishing workload, and generate first optimized polishing parameters.

[0028] Specifically, digital twin technology is a virtual model based on a physical entity or system, which realizes state monitoring, performance evaluation and optimization control of the entity or system through real-time data synchronization and simulation. Based on the digital twin technology, a simulation model is established according to the attribute information of the polishing equipment and the attribute of the mirror material, and a polishing processing twin model is built. Through this polishing processing twin model based on the digital twin technology, virtual simulation and prediction of the polishing process can be realized. An evaluation function is created, which will be used to evaluate the polishing quality under different polishing parameter combinations. The evaluation function includes key indicators such as surface roughness, reflectivity and number of microscopic defects.

[0029] Then, based on the polishing processing twin model and a predetermined polishing evaluation function, polishing parameter optimization is performed to meet the second stage polishing workload as the optimization purpose, wherein the polishing parameters include polishing speed, polishing pressure and polishing time length, such as: the polishing parameters can be optimized by using an optimization algorithm, the goal of the optimization algorithm is to maximize the polishing evaluation function to meet the requirement of the second stage polishing workload, and the first optimized polishing parameter is obtained according to the optimization result, wherein the first optimized polishing parameter is used to guide the second stage polishing operation to improve the polishing quality and production efficiency.

[0030] The polishing control module 16 is used to determine the second starting node according to the first sequence number of the first sub-region, feed the second starting node and the first optimized polishing parameter to the polishing control unit, and perform the second stage polishing control of the first sub-region until the entire polishing process is completed, wherein the second starting node is the starting time node of the second stage.

[0031] Specifically, first, the starting time node analysis of the second stage is performed according to the first sequence number of the first sub-region, such as: the second starting node can be obtained by superimposing calculation according to the current time node and the polishing time length of the remaining sub-regions before the first sub-region, wherein the second starting node is the starting time node of the second stage. Then, the second starting node and the first optimized polishing parameter are sent to the polishing control unit, wherein the polishing control unit is an integrated system for controlling the working state and parameter setting of the polishing equipment. Then, the polishing control unit performs the second stage polishing control of the first sub-region according to the feedback parameter and starting node, such as: the control unit adjusts the parameter setting of the polishing equipment, such as polishing speed, polishing pressure, polishing time length, etc., to achieve the expected polishing effect. Repeat the above steps to perform polishing optimization control for each polishing sub-region until the entire polishing process is completed.

[0032] The application discloses a shadowless lamp quality control system based on digital processing, which comprises a polishing index division module, a polishing area division module, a polishing quality simulation detection module, a quality deviation analysis module, a polishing parameter optimization module and a polishing control module. The polishing index division module is used for reading polishing indexes and a polishing process of a target mirror, determining a plurality of polishing stages based on the polishing process, and dividing the polishing indexes according to the polishing stages to obtain a stage polishing index set. The polishing area division module is used for dividing a polishing mirror into a plurality of polishing sub-areas according to a single polishing area of a grinding wheel, wherein each polishing sub-area has a unique serial number. The polishing quality simulation detection module is used for performing a single polishing operation on a first sub-area based on a predetermined polishing path, collecting an image of the first sub-area by using an optical microscopic device after the operation is completed, and performing polishing quality simulation detection based on the microscopic image of the first sub-area to obtain a first stage detection result, wherein the first sub-area is any one of the plurality of polishing sub-areas, and the single polishing operation is a first stage polishing operation. The quality deviation analysis module is used for selecting a second stage polishing index from the stage polishing index set, performing quality deviation analysis on the first stage detection result according to the second stage polishing index, and setting a deviation analysis result as a second stage polishing workload. The polishing parameter optimization module is used for performing polishing parameter optimization based on a polishing processing twin model and a predetermined polishing evaluation function to meet the second stage polishing workload, and generating first optimized polishing parameters. The polishing control module is used for determining a second starting node according to a first serial number of the first sub-area, feeding back the second starting node and the first optimized polishing parameters to a polishing control unit, and performing second stage polishing control of the first sub-area until the whole polishing process is completed, wherein the second starting node is a starting time node of the second stage. The application can solve the technical problem that the polishing precision and quality of a mirror are poor and the production quality of a surgical shadowless lamp is poor due to the fact that a processing defect in a mirror polishing process cannot be accurately detected and timely corrected.The accuracy and precision of the polishing defect detection in the mirror polishing process can be improved, so that the processing defects can be timely and effectively corrected, the polishing precision and quality of the mirror are improved, and the technical effects of meeting the production requirements of the surgical shadowless lamp are achieved.

[0033] Further, the polishing index division module 11 is also used for:

[0034] The polishing index includes surface roughness, reflectivity, and microscopic defect constraints.

[0035] Specifically, the polishing index of the target mirror includes surface roughness, reflectivity, and microscopic defect constraints, which can be set by a person skilled in the art according to actual processing requirements. The surface roughness refers to the degree of microscopic unevenness of the surface of the optical element, which directly affects the reflection and scattering of light and is an important indicator for evaluating the polishing quality. The reflectivity refers to the ability of the optical element to reflect light, and the higher the reflectivity, the better the performance of the optical element. It is usually used to evaluate the optical performance of the optical element. Microscopic defects include scratches, pits, etc., which can affect the optical performance and mechanical strength of the optical element, and therefore need to be strictly controlled.

[0036] Further, as shown in the accompanying Figure 2 The polishing quality simulation detection module 13 is also used for:

[0037] The first sub-region microscopic image is input into a feature extraction channel, and a polishing feature set is output, wherein the polishing feature set includes a particle feature set, a ripple feature set, a texture feature set, a scratch feature set, and a pit feature set. The feature extraction channel is constructed based on a convolutional neural network. The particle feature set, the ripple feature set, and the texture feature set are input into a surface evaluation model, and a predicted roughness is output. Reflectivity prediction is performed based on the scratch feature set, the pit feature set, and the predicted roughness, and a predicted reflectivity is output. The scratch feature set, the pit feature set, the predicted roughness, and the predicted reflectivity are taken as the first-stage detection result.

[0038] Specifically, the convolutional neural network is a deep learning model that can automatically learn and extract features from images. First, a feature extraction channel is constructed based on a convolutional neural network, wherein the extracted features of the feature extraction channel include particle features, ripple features, texture features, scratch features, and pit features. Then, sample data is obtained based on the extracted features, and the feature extraction channel is supervised trained to obtain a feature extraction channel that meets the expected index. Next, the first sub-region microscopic image is input into the feature extraction channel to automatically identify and classify microscopic defects in the polishing process, and a polishing feature set is obtained, wherein the polishing feature set includes a particle feature set, a ripple feature set, a texture feature set, a scratch feature set, and a pit feature set.

[0039] The BP neural network is a kind of multi-layer feedforward neural network, which is trained by error back propagation algorithm. A surface evaluation model is constructed based on the BP neural network, which is a neural network model that can be iteratively optimized in machine learning, and is used for predicting surface roughness according to polishing characteristics. Then, the particle feature set, the ripple feature set and the texture feature set are input into the surface evaluation model, and the predicted roughness is output. On the other hand, the reflectivity simulation prediction is carried out based on the target light source information, the mirror material attributes, the scratch feature set, the indentation feature set and the predicted roughness, and the predicted reflectivity is obtained. Finally, the scratch feature set, the indentation feature set, the predicted roughness and the predicted reflectivity are taken as the first stage detection results.

[0040] Further, the polishing quality simulation detection module 13 is also used for:

[0041] The mirror material attributes are acquired, and the sample data set is acquired by network retrieval with the mirror material attributes as constraints. The sample data includes sample particle feature set, sample ripple feature set, sample texture feature set and sample surface roughness. The sample data set is equally divided into M parts, and M times are selected with replacement in the M data sets. A first data set is constructed, and M times are iteratively selected to obtain M data sets. The BP neural network is supervised trained and cross-validated by using the M data sets, M surface evaluation units meeting the expected convergence constraints are obtained, and a surface evaluation model is built based on the M surface evaluation units. The output of the surface evaluation model is the mode of the output results of the M surface evaluation units.

[0042] Specifically, first, the mirror material attributes are acquired, wherein the mirror material attributes include material type, size information, surface structure and other characteristics. Then, the sample data set is acquired by network retrieval based on big data technology with the mirror material attributes as constraints. The sample data includes sample particle feature set, sample ripple feature set, sample texture feature set and sample surface roughness, and the sample particle feature set, sample ripple feature set, sample texture feature set and sample surface roughness have a corresponding relationship. Then, the sample data set is equally divided into M parts to ensure that the size and structure of each data set are as consistent as possible to ensure the representativeness of each data set, and M data sets are obtained. Further, M times are selected with replacement in the M data sets, that is, each time, a data set is randomly selected from the M data sets, and then put back, to construct a first data set. Then, the above selection process is repeated, and M times are selected each time to obtain M different data sets.

[0043] The initial surface evaluation unit is constructed based on a BP neural network, wherein the input of the initial surface evaluation unit is the particle feature set, the ripple feature set and the texture feature set, and the output is the surface roughness; then the BP neural network is supervised trained and cross-validated by using M data sets until M surface evaluation units meeting the expected convergence constraint are obtained, wherein the expected convergence constraint is the model training requirement, which can be set according to the actual situation, wherein the supervised training and cross-validation process is to use the data set for supervised training, to minimize the prediction error by adjusting the weight and bias of the network until the preset training times are met; then the data set is exchanged, and the verification training is performed by using different data sets until the expected convergence constraint is met. Finally, the surface evaluation model is built based on the M surface evaluation units, wherein the output of the surface evaluation model is the mode of the output results of the M surface evaluation units, and the mode refers to the numerical value with the highest frequency in a group of data, which can reflect the common prediction results of most surface evaluation units. By this method, a surface evaluation model based on M surface evaluation units can be constructed, which can integrate the information of multiple evaluation units, thereby improving the prediction accuracy and robustness.

[0044] Further, the polishing quality simulation detection module 13 is also used for:

[0045] obtaining target light source information, wherein the target light source information includes light source brightness, light source color temperature, spectral distribution and luminous flux; based on the target light source information, the mirror material attribute, the scratch feature set, the dent feature set and the predicted roughness, N times of simulated reflectivity tests are performed in a simulation platform to obtain N reflectivities, and a reflectivity mean value is calculated; the deviation values of the N reflectivities and the reflectivity mean value are calculated in turn, the reflectivity with a deviation value less than a predetermined deviation threshold is selected as a standard reflectivity, the mean value of multiple standard reflectivities is calculated, and the standard reflectivity mean value is set as the predicted reflectivity.

[0046] Specifically, target light source information of a target surgical shadowless lamp product is acquired, wherein the target light source information includes light source brightness, light source color temperature, spectral distribution, and luminous flux. Then, N times of simulated reflectivity tests are performed in a simulation platform based on the target light source information, the mirror material properties, the scratch feature set, the indentation feature set, and the predicted roughness. First, a digital twin model is created on the simulation platform, which accurately replicates the optical element in the actual polishing process. The digital twin model should include key parameters such as the size, shape, material properties, surface roughness, and microscopic defects of the element. Then, according to the light source characteristics in actual application, the target light source information, mirror material property information, scratch feature set, indentation feature set, and predicted roughness are input into the simulation platform. Further, N times of simulated reflectivity tests are performed in the simulation platform. In each test, according to the input parameters, the interaction between the incident light and the mirror is simulated, the distribution and intensity of the reflected light are calculated, N reflectivities are obtained, and the mean of the N reflectivities is calculated to obtain the reflectivity mean.

[0047] Further, the N reflectivities and the reflectivity mean are sequentially calculated to obtain N deviation values, and the N deviation values are judged according to a predetermined deviation threshold value, wherein the predetermined deviation threshold value can be set according to actual conditions to remove outliers in the simulation test process, improve the accuracy of the test reflectivity data, select reflectivities with deviation values less than the predetermined deviation threshold value as standard reflectivities, and then calculate the mean of the multiple standard reflectivities, set the standard reflectivity mean as the predicted reflectivity, thereby improving the accuracy and reliability of the predicted reflectivity setting.

[0048] Further, the polishing parameter optimization module 15 is also used for:

[0049] Based on digital twin technology, a simulation model is built based on the attribute information of the polishing equipment and the mirror material properties to build a polishing processing twin model. Based on the polishing processing twin model and a predetermined polishing evaluation function, polishing parameter optimization is performed to obtain first optimized polishing parameters, wherein the polishing parameters include polishing speed, polishing pressure, and polishing duration.

[0050] Specifically, based on the digital twin technology, a virtual model of the polishing equipment is created using computer-aided design (CAD) software according to attribute information of the polishing equipment; a virtual model of the mirror material is created using computer-aided engineering (CAE) software according to the attribute of the mirror material; then the virtual model of the polishing equipment and the virtual model of the mirror material are combined to form a twin model of the polishing process. Further based on the polishing process twin model, simulated polishing is performed, polishing parameter fitness evaluation is performed based on a predetermined polishing evaluation function, and polishing parameter optimization is performed using an optimization algorithm, wherein the polishing parameters include polishing speed, polishing pressure, and polishing time length, and first optimized polishing parameters are obtained according to the optimization results.

[0051] Further, the polishing parameter optimization module 15 is also used for:

[0052] The expression of the predetermined polishing evaluation function is: ;

[0053] Wherein, is the fitness of the i-th polishing parameter, is the first weight, is the second weight, is the third weight, is the fourth weight, is the fifth weight, is the simulated surface roughness after processing of the i-th polishing parameter, is the surface roughness of the second stage polishing workload, is the simulated reflectivity after processing of the i-th polishing parameter, is the reflectivity of the second stage polishing workload, is the polishing time length of the i-th polishing parameter, is the total amount of defect features after processing of the i-th polishing parameter, is the energy consumption of the i-th polishing parameter.

[0054] Specifically, in the predetermined polishing evaluation function, is the fitness of the i-th polishing parameter, and the i-th polishing parameter is any one of the polishing parameters, wherein the greater the fitness, the better the overall polishing quality is represented; is the first weight, is the second weight, is the third weight, is the fourth weight, is the fifth weight, wherein the sum of the first weight, the second weight, the third weight, the fourth weight, and the fifth weight is 1, and the specific weight value is set based on the influence degree of the index on the overall polishing quality, and the greater the influence degree, the greater the corresponding weight is; is the simulated surface roughness after processing of the i-th polishing parameter, a surface roughness of a polishing workload for a second stage, a simulated reflectivity after processing with an i-th polishing parameter, a reflectivity of a polishing workload for a second stage, a polishing duration of an i-th polishing parameter, a total amount of defect features after processing with an i-th polishing parameter, an energy consumption of an i-th polishing parameter.

[0055] Further, the polishing parameter optimization module 15 is further configured to:

[0056] obtain polishing parameter rated thresholds, wherein the polishing parameter rated thresholds comprise a polishing speed threshold, a polishing pressure threshold, and a polishing duration threshold; randomly select one parameter in the polishing speed threshold, the polishing pressure threshold, and the polishing duration threshold for combination to generate a plurality of initial polishing parameters, and simulate polishing on the plurality of initial polishing parameters by using the polishing processing twin model to obtain a plurality of simulated polishing results; evaluate the plurality of simulated polishing results based on the predetermined polishing evaluation function to obtain a plurality of fitnesses, set the initial polishing parameters as initial solutions, arrange the plurality of initial solutions in descending order of the fitnesses to generate an initial solution sequence; set the first X initial solutions in the initial solution sequence as first solutions and the last Y initial solutions as tail solutions, and randomly cluster the Y tail solutions with the X first solutions as the center to obtain X fields, wherein Y is an integer multiple of X, and the number of tail solutions in each field is the same; in the X fields, take the first solution in the field as an optimization direction, and perform one-time optimization adjustment on the tail solutions in the field according to a preset optimization step to obtain X updated fields, wherein if the adjusted tail solution does not satisfy the polishing parameter rated threshold, no adjustment is performed; identify the X updated fields, and if the fitness of the tail solution in the updated field is greater than the fitness of the first solution, update and replace the first solution with the tail solution; continue to perform iterative optimization until a predetermined optimization times constraint is satisfied, output X current fields, and select the first solution of the optimal current field as the first optimized polishing parameter, wherein the optimal current field is the field with the largest sum of fitnesses in the X current fields.

[0057] Specifically, obtain polishing parameter rated thresholds, wherein the polishing parameter rated thresholds comprise a polishing speed threshold, a polishing pressure threshold, and a polishing duration threshold; then randomly select one parameter in the polishing speed threshold, the polishing pressure threshold, and the polishing duration threshold for combination to obtain a first initial polishing parameter, and generate a plurality of initial polishing parameters by using the same method; then simulate polishing on the plurality of initial polishing parameters by using the polishing processing twin model to obtain a plurality of simulated polishing results. Further, evaluate the plurality of simulated polishing results based on the predetermined polishing evaluation function to obtain a plurality of fitnesses.

[0058] The initial polishing parameters are set as initial solutions, and then the initial solutions are arranged in descending order of fitness to generate an initial solution sequence; further, the first X initial solutions in the initial solution sequence are set as head solutions, and the last Y initial solutions are set as tail solutions, where the sum of X and Y is the number of initial solutions, and Y is an integer multiple of X; then, the X head solutions are taken as the center to randomly cluster the Y tail solutions to obtain X fields, where the number of tail solutions in each field is the same.

[0059] In the X fields, the head solution in each field is taken as the optimization direction, and the tail solutions in each field are adjusted once according to a preset optimization step, where the preset optimization step includes a polishing speed adjustment step, a polishing pressure adjustment step and a polishing time adjustment step, and during the once optimization adjustment, if the adjusted tail solution does not satisfy the polishing parameter rated threshold, the tail solution is not adjusted, and X updated fields are obtained. Further, the X updated fields are identified, and if there is a tail solution in the updated field whose fitness is greater than that of the head solution, the tail solution is used to update and replace the head solution. Iterative optimization is continued until a predetermined optimization number constraint is satisfied, the predetermined optimization number constraint can be set according to optimization requirements, where the higher the required accuracy, the greater the predetermined optimization number, X current fields are output, and the field with the maximum sum of fitness in the X current fields is selected as the optimal current field, and the head solution of the optimal current field is selected as the first optimized polishing parameter.

[0060] By using the above optimization algorithm for polishing parameter optimization, since the algorithm has strong global search ability, it can avoid falling into local optimum, and improve the accuracy and comprehensiveness of the first optimized polishing parameter.

[0061] In summary, the shadowless lamp quality control system based on digital processing provided by the present application has the following technical effects:

[0062] 1. By dividing the polishing index according to multiple polishing stages to obtain a set of stage polishing indexes; on the other hand, based on the single polishing area of the grinding wheel, the mirror to be polished is divided into multiple polishing sub-areas; then after the completion of the first polishing operation, the polishing quality simulation detection is carried out based on the first sub-area microscopic image to obtain the first stage detection result, then the quality deviation analysis is carried out on the first stage detection result according to the second stage polishing index, and the deviation analysis result is set as the second stage polishing workload; further based on the polishing processing twin model and the predetermined polishing evaluation function, the polishing parameter optimization is carried out to meet the second stage polishing workload, and the first optimized polishing parameter is generated; finally, the second starting node is determined based on the first serial number of the first sub-area, and the second stage polishing control of the first sub-area is carried out according to the second starting node and the first optimized polishing parameter, until the entire polishing process is completed. The accuracy and precision of the polishing defect detection in the mirror polishing process can be improved, so that the processing defects can be timely and effectively corrected, the polishing precision and quality of the mirror can be improved, and the technical effect of meeting the production requirements of the surgical shadowless lamp product can be achieved.

[0063] 2. By using artificial intelligence methods such as convolutional neural network, BP neural network and digital twin to carry out polishing quality simulation detection on the first sub-area microscopic image, the intelligence and automation of polishing quality detection can be improved, and the efficiency and accuracy of detection can be improved, and defects in the polishing process can be found in time to provide basis for subsequent polishing optimization control.

[0064] 3. By setting the polishing workload of the next stage based on the deviation calculation and analysis result, the processing defects of the last polishing stage can be timely and effectively corrected, so as to realize the accurate control and optimization of the whole polishing process, and improve the precision and polishing quality of the whole polishing.

[0065] The above description of disclosed embodiments enables those skilled in the art to carry out or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0066] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application also intends to include these modifications and variations.

Claims

1. A shadowless lamp quality control system based on digitized processing, characterized by, The polishing index includes surface roughness, reflectivity, and microscopic defect constraints. The polishing quality simulation detection module is further configured to: input the first sub-area microscopic image into a feature extraction channel, and output a polishing feature set, wherein the polishing feature set includes a particle feature set, a ripple feature set, a texture feature set, a scratch feature set, and a indentation feature set, and the feature extraction channel is constructed based on a convolutional neural network; input the particle feature set, the ripple feature set, and the texture feature set into a surface evaluation model, and output a predicted roughness; perform reflectivity prediction based on the scratch feature set, the indentation feature set, and the predicted roughness, and output a predicted reflectivity; the scratch feature set, the indentation feature set, the predicted roughness, and the predicted reflectivity are taken as the first stage detection result. The polishing quality simulation detection module is further configured to: obtain a mirror material attribute, and network to retrieve a sample data set based on the mirror material attribute as a constraint, wherein the sample data includes a sample particle feature set, a sample ripple feature set, a sample texture feature set, and a sample surface roughness; divide the sample data set into M parts, and select M times with replacement in the M data sets, to construct a first data set, and continue to select M times to obtain M data sets; ​ ​ ​ ​ 2. The shadowless lamp quality control system based on digitalized machining according to claim 1, wherein, ​ ​ ​ The BP neural network is supervised trained and cross-validated by using M data sets, M surface evaluation units meeting the expected convergence constraint are obtained, and a surface evaluation model is built based on the M surface evaluation units, wherein the output of the surface evaluation model is the mode of the output results of the M surface evaluation units.

3. The shadowless lamp quality control system based on digitalized machining according to claim 2, wherein, The polishing quality simulation detection module is further configured to: Obtain target light source information, wherein the target light source information includes light source brightness, light source color temperature, spectral distribution, and luminous flux; Based on the target light source information, the mirror material attributes, the scratch feature set, the indentation feature set, and the predicted roughness, perform N times of simulated reflectivity tests in a simulation platform to obtain N reflectivities, and calculate a reflectivity mean value; In turn, calculate deviation values of the N reflectivities and the reflectivity mean value, select a reflectivity with a deviation value less than a predetermined deviation threshold as a standard reflectivity, calculate a mean value of a plurality of standard reflectivities, and set the standard reflectivity mean value as the predicted reflectivity.

4. The shadowless lamp quality control system based on digitalized machining according to claim 2, wherein, The polishing parameter optimization module is further configured to: Based on digital twinning technology, perform simulation modeling according to attribute information of the polishing equipment and the mirror material attributes, and build a polishing processing twinning model; Based on the polishing processing twinning model and a predetermined polishing evaluation function, perform polishing parameter optimization to obtain first optimized polishing parameters, with the constraint condition being to meet the second stage polishing workload, wherein the polishing parameters include polishing speed, polishing pressure, and polishing duration.

5. The shadowless lamp quality control system based on digitalized machining according to claim 4, wherein, The polishing parameter optimization module is further configured to: The expression of the predetermined polishing evaluation function is: ； wherein, fitness of the i-th polishing parameter, a first weight, a second weight, a third weight, a fourth weight, a fifth weight, simulated surface roughness after processing of the i-th polishing parameter, surface roughness of the second stage polishing workload, simulated reflectivity after processing of the i-th polishing parameter, reflectivity of the second stage polishing workload, polishing time of the i-th polishing parameter, total amount of defect features after processing of the i-th polishing parameter, energy consumption of the i-th polishing parameter.

6. The shadowless lamp quality control system based on digitalized machining according to claim 5, wherein, The polishing parameter optimization module is further configured to: Obtain polishing parameter rated thresholds, wherein the polishing parameter rated thresholds include polishing speed thresholds, polishing pressure thresholds, and polishing duration thresholds; Randomly select one parameter from the polishing speed thresholds, the polishing pressure thresholds, and the polishing duration thresholds to form a plurality of initial polishing parameters, and perform simulation polishing on the plurality of initial polishing parameters by using the polishing processing twinning model to obtain a plurality of simulation polishing results; Evaluate the plurality of simulation polishing results based on the predetermined polishing evaluation function to obtain a plurality of fitness values, set the initial polishing parameters as initial solutions, arrange the plurality of initial solutions in descending order of the fitness values to generate an initial solution sequence; Set the first X initial solutions of the initial solution sequence as first solutions and the last Y initial solutions as tail solutions, and randomly cluster the Y tail solutions around the X first solutions to obtain X fields, wherein Y is an integer multiple of X, and the number of tail solutions in each field is the same; In the X fields, perform one optimization adjustment on the tail solutions in each field in the optimization direction of the first solution in the field according to a preset optimization step to obtain X updated fields, wherein if the adjusted tail solution does not meet the polishing parameter rated thresholds, no adjustment is performed; Identify the X updated fields, and if the fitness value of the tail solution in an updated field is greater than the fitness value of the first solution in the field, update and replace the first solution with the tail solution; Continue the iterative optimization until a predetermined optimization number constraint is met, output X current fields, and select the first solution of the optimal current field as the first optimized polishing parameter, wherein the optimal current field is the field with the largest sum of fitness values among the X current fields.

Images