Self-adaptive compensation method and system for optical lens coating

Abstract

The invention relates to the technical field of adaptive control, and particularly discloses an adaptive compensation method and system for optical lens coating, and the method comprises the steps: obtaining technological parameters and evaporation dynamic data through a two-dimensional collection system, and carrying out the cleaning and integration of a moving average method into a data set; then precipitation harmonic variation features and resonance pit depth features are extracted through two-stage feature analysis, and a judgment matrix is constructed to recognize the evaporative respiration phenomenon; if the phenomenon occurs, a development steady-state index is obtained based on a pit depth variation coefficient and bombardment dispersion, and the starting time of the multiple pits is predicted; matching historical data through a Manhattan distance to determine an actually formed time window, and judging whether an optimized duration is met or not; and finally, gradual parameter optimization or emergency shutdown measures are respectively adopted in an autonomous tolerance scene and an emergency compensation scene. The system comprises a data acquisition module, a phenomenon identification module, a prediction analysis module, a window judgment module and a hierarchical optimization module. And the uniformity of the film layer can be optimized.

Description

Technical Field

[0001] This invention relates to the field of adaptive control technology, and specifically to an adaptive compensation method and system for optical lens coating. Background Technology

[0002] As a core imaging device in minimally invasive surgery, medical endoscopes require precision optical coating as a key technological step to improve the light transmittance, contrast, and anti-reflection performance of endoscope lenses. Among these methods, electron beam evaporation coating has become the mainstream coating method for high refractive index materials such as TiO2 and Ta2O5 due to its advantages of high film purity, strong adhesion, and good process controllability. However, the evaporation breathing phenomenon has become a core technical bottleneck restricting the uniformity of the film layer.

[0003] Specifically, during electron beam evaporation coating, the high-energy electron beam bombards the film surface, causing localized areas to form "evaporation pits" due to instantaneous high temperatures. The dynamic evolution of these pits causes fluctuations of ±3-5% in the geometric distance between the evaporation source and the endoscope lens substrate, directly leading to instability in the deposition rate (fluctuation range of ±3-5%). Because medical endoscopes have extremely high requirements for the uniformity of film thickness (usually controlled within ±1%), these minute fluctuations in the deposition rate are amplified by the "cosine deposition law," ultimately resulting in uneven film thickness distribution on the lens surface, with the thickness difference between the center and edge areas potentially exceeding 5%.

[0004] Therefore, the present invention provides an adaptive compensation method and system for optical lens coating. Summary of the Invention

[0005] The purpose of this invention is to provide an adaptive compensation method and system for optical lens coating, so as to solve the above-mentioned background problems.

[0006] The objective of this invention can be achieved through the following technical solutions: An adaptive compensation system for optical lens coating includes the following modules: Data acquisition module: Collects process parameter data and vapor deposition dynamic data during the coating process of medical endoscopes, and integrates the process parameter data and vapor deposition dynamic data into a vapor deposition parameter dataset; Phenomenon recognition module: Based on the evaporation parameter dataset, a two-level feature analysis is performed to obtain precipitation harmonic characteristics and resonance pit depth characteristics. Based on the precipitation harmonic characteristics and resonance pit depth characteristics, an evaporation judgment matrix is ​​constructed to identify whether evaporative breathing phenomenon has occurred. Predictive Analysis Module: If evaporative respiration occurs, predictive analysis will be performed to identify the start time of the development of evaporative respiration into a multi-pitted state. Window judgment module: Based on the determined start time, distance measurement analysis is performed to obtain the actual formation time window of multiple pit states, and the actual formation time window is used to determine whether the time required for system optimization is met; Hierarchical optimization module: If the time required for system optimization is met, a progressive parameter optimization strategy is adopted for optimization; if the time required for system optimization is not met, the coating process must be stopped immediately.

[0007] Furthermore, the process of performing the aforementioned two-level feature analysis is as follows: First-level feature analysis: The initial sedimentation rate is obtained, the real-time sedimentation rate at all collection time points is obtained, the absolute difference between the real-time sedimentation rates at adjacent collection time points is calculated to obtain the sedimentation rate difference, and the ratio of the sedimentation rate difference to the initial sedimentation rate is calculated to obtain the instantaneous sedimentation fluctuation rate at that collection time point. Based on the cosine precipitation law, the instantaneous fluctuation rate of precipitation is corrected to obtain the precipitation harmonic variation characteristics; The resonance pit depth characteristics were obtained by performing a second-level feature analysis on the evaporation parameter dataset.

[0008] Furthermore, the cosine precipitation law corrects the instantaneous fluctuation rate of precipitation in the following way: Obtain the baseline evaporation distance and the effective evaporation distance at the collection time point, calculate the absolute difference between the effective evaporation distance and the baseline evaporation distance, and calculate the evaporation distance fluctuation coefficient by comparing the result of the absolute difference calculation with the baseline evaporation distance. The instantaneous fluctuation rate of precipitation is calculated by multiplying the fluctuation coefficient of evaporation distance, and the result of the product calculation is added to the instantaneous fluctuation rate of precipitation to obtain the corrected instantaneous fluctuation rate of precipitation. The corrected instantaneous fluctuation rate of precipitation is labeled as a precipitation harmonic characteristic.

[0009] Furthermore, the process of performing the second-level feature analysis is as follows: Obtain the effective evaporation distance at all collection time points, and calculate the evaporation pit depth at the collection time point by comparing the effective evaporation distance at the collection time point with the reference evaporation distance. The depth of the evaporation pit at adjacent acquisition time points is obtained and the difference is calculated to obtain the characteristic of the resonance pit depth.

[0010] Furthermore, the process of performing the predictive analysis is as follows: Obtain the baseline evaporation distance, obtain the effective evaporation distance at the time points within the sliding time window, and perform pit depth variation analysis based on the baseline evaporation distance and the effective evaporation distance to obtain the pit depth variation coefficient. The electron beam bombardment coordinates at the acquisition time points within the sliding time window are obtained and the bombardment dispersion is obtained by performing dispersion analysis. The steady-state development index is obtained by multiplying the crater depth variation coefficient and the bombardment dispersion after standard normalization.

[0011] Furthermore, the process of performing the pit depth variation analysis is as follows: The depth of the evaporation pit at each collection time point is calculated by comparing the effective evaporation distance with the reference evaporation distance. The difference between the evaporation pit depths at adjacent acquisition time points is calculated to obtain the resonance pit depth feature, and the resonance pit depth feature of each acquisition time point is integrated into a resonance pit depth sequence. The coefficient of variation is calculated from the resonant pit depth sequence, and the coefficient of variation of the resonant pit depth sequence is denoted as the pit depth coefficient of variation.

[0012] Furthermore, the process of performing the aforementioned dispersion analysis is as follows: Obtain the electron beam bombardment coordinates at the acquisition time points, extract the X-axis coordinates of all electron beam bombardment coordinates and calculate the variance, labeled as Var(X), and extract the Y-axis coordinates of all electron beam bombardment coordinates and calculate the variance, labeled as Var(Y). The bombardment dispersion is obtained by summing Var(X) and Var(Y).

[0013] Furthermore, the process of performing the distance metric analysis is as follows: Obtain the historical sedimentation harmonic sequence and historical resonance pit depth sequence corresponding to the starting window in each historical multi-pit event; Obtain the formation time window for each historical multi-pit event; Obtain the precipitation harmonic sequence and resonance pit depth sequence corresponding to the starting window in actual multi-pit events; Based on the historical precipitation harmonic sequence and historical resonance pit depth sequence of each historical multi-pit event, as well as the precipitation harmonic sequence and resonance pit depth sequence of the actual multi-pit event, a multi-dimensional evolution analysis is performed to obtain the actual formation time window of the actual multi-pit event.

[0014] Furthermore, the process of performing the aforementioned multidimensional evolutionary analysis is as follows: Calculate the Manhattan distance between the actual multi-pit event and each historical multi-pit event; Obtain all Manhattan distances and sort them in descending order. Extract the formation time window of the historical multi-pit event corresponding to the smallest Manhattan distance, and use it as the actual formation time window of the actual multi-pit event.

[0015] An adaptive compensation method for optical lens coating includes the following steps: Step 1: Collect process parameter data and vapor deposition dynamic data during the coating process of medical endoscopes, and integrate the process parameter data and vapor deposition dynamic data into a vapor deposition parameter dataset; Step 2: Perform two-level feature analysis based on the evaporation parameter dataset to obtain precipitation harmonic characteristics and resonance pit depth characteristics. Construct an evaporation judgment matrix based on precipitation harmonic characteristics and resonance pit depth characteristics to identify whether evaporative breathing has occurred. Step 3: If evaporative respiration occurs, predict and analyze the evaporative respiration phenomenon to identify the starting time of its development into a multi-pitted state. Step 4: Based on the determined start time, perform distance metric analysis to obtain the actual formation time window of the multiple pit states, and determine whether the time required for system optimization is met based on the actual formation time window; Step 5: If the time required for system optimization is met, then an incremental parameter optimization strategy is adopted for optimization; if the time required for system optimization is not met, then the coating process must be stopped immediately.

[0016] The beneficial effects of this invention are as follows: 1. Collect process parameter data and vapor deposition dynamic data during the coating process of medical endoscopes, and integrate the process parameter data and vapor deposition dynamic data into a vapor deposition parameter dataset; provide comprehensive data support for tracing the causes of evaporative breathing phenomenon in medical endoscope coatings, adjusting process parameters, and improving coating stability; perform two-level feature analysis based on the vapor deposition parameter dataset to obtain precipitation harmonic characteristics and resonance pit depth characteristics, and construct an evaporation judgment matrix based on precipitation harmonic characteristics and resonance pit depth characteristics to identify whether evaporative breathing phenomenon has occurred; through two-level feature extraction and judgment matrix construction, achieve rapid identification of evaporative breathing phenomenon, and provide targeted basis for optimizing medical endoscope coating process and suppressing phenomenon.

[0017] 2. If evaporative respiration occurs, predictive analysis is performed to identify the onset time of its development into a multi-pitted state; key nodes in the evolution of evaporative respiration into a multi-pitted state are identified, providing a time window for subsequent intervention and optimization; distance metric analysis is performed based on the determined onset time to obtain the actual formation time window of the multi-pitted state, and it is determined whether the time required for system optimization is met based on the actual formation time window; the time feasibility of system optimization is verified, preparing for the timely initiation of targeted suppression strategies; if the time required for system optimization is met, a progressive parameter optimization strategy is adopted for optimization; if the time required for system optimization is not met, the coating process must be stopped immediately; through time-adaptive differentiated treatment, the risk of multi-pitted defects expanding is prevented, and the optical performance and production rationality of medical endoscope coating are improved. Attached Figure Description

[0018] The invention will now be further described with reference to the accompanying drawings.

[0019] Figure 1 This is a functional block diagram of an adaptive compensation system for optical lens coating in this invention; Figure 2 This is a logic diagram for identifying whether evaporative respiration has occurred in this invention; Figure 3 This is a flowchart of the steps of an adaptive compensation method for optical lens coating in this invention. Detailed Implementation

[0020] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments. Example 1

[0021] like Figures 1-2 As shown, an adaptive compensation system for optical lens coating includes: Data acquisition module: Collects process parameter data and vapor deposition dynamic data during the coating process of medical endoscopes, and integrates the process parameter data and vapor deposition dynamic data into a vapor deposition parameter dataset; The process of collecting process parameter data and vapor deposition dynamic data during the coating process is as follows: Based on the coating process of medical endoscopes, a two-dimensional data acquisition system is constructed at the level of process parameters and the level of vapor deposition dynamics to collect raw process parameter data and raw vapor deposition dynamic data. It should be noted that the process parameter data are the core coating settings preset before and during the coating process of medical endoscopes, which are the benchmark for the coating process; the vapor deposition dynamic data are the process feedback values ​​dynamically generated during the coating process, which are real-time data reflecting the evaporation state and film formation. Process parameters: The raw process parameter data is collected by the industrial automatic control system device, and the acquisition frequency is set to 100Hz. The acquisition sequence of process parameter data is completely synchronized with the coating process sequence. The process parameter data includes electron beam power, electron beam bombardment coordinates, and effective area of ​​film loading; Dynamic aspects of vapor deposition: Deploy laser measuring instruments and intelligent measuring instruments to collect raw vapor deposition dynamic data, and set the acquisition frequency and acquisition sequence to be the same as the process parameters between layers; The dynamic data of vapor deposition includes the reference vapor distance, the effective vapor distance, and the real-time deposition rate. It should be noted that the reference evaporation distance is the original distance between the evaporation source and the endoscope lens substrate, the effective evaporation distance is the real-time distance between the evaporation source and the endoscope lens substrate, the initial precipitation rate is the speed at which the film material forms a film layer on the endoscope lens surface during the initial stage of evaporation, and the real-time precipitation rate is the real-time speed at which the film material forms a film layer on the endoscope lens surface during the evaporation process. Cleaning is performed based on the original process parameter data and the original vapor deposition dynamic data to obtain the process parameter data and vapor deposition dynamic data. The cleaning process is as follows: The moving average method was used to process the original process parameter data and the original vapor deposition dynamic data. A fixed sliding time window of 50ms was set, and the sliding step of the window was 10ms. The arithmetic mean of the original data within the window was calculated, and the obtained mean was used as the smoothed data value at the center of the window. The obtained process parameter data and vapor deposition dynamic data are integrated into a vapor deposition parameter dataset. Phenomenon recognition module: Based on the evaporation parameter dataset, a two-level feature analysis is performed to obtain precipitation harmonic characteristics and resonance pit depth characteristics. Based on the precipitation harmonic characteristics and resonance pit depth characteristics, an evaporation judgment matrix is ​​constructed to identify whether evaporative breathing phenomenon has occurred. The process of performing two-level feature analysis based on the vapor deposition parameter dataset is as follows: First-level feature analysis: The real-time precipitation rate at the beginning of evaporation was obtained and the average value was calculated as the initial precipitation rate. It should be noted that the evaporation initiation stage is the transition stage in medical endoscope optical coating, from the start of the electron beam bombarding the surface of the coating material to the time before the coating material enters a stable evaporation state. The sedimentation rate at all collection time points is obtained. The sedimentation rate difference is calculated by the absolute difference between the sedimentation rates at adjacent collection time points. The instantaneous fluctuation rate of sedimentation at that collection time point is calculated by the ratio of the sedimentation rate difference to the initial sedimentation rate. The instantaneous fluctuation rate of precipitation is corrected based on the cosine precipitation law; The cosine precipitation law states that in vacuum evaporation coating, the deposition rate of the film on the substrate surface is inversely proportional to the square of the evaporation distance and directly proportional to the cosine of the angle between the evaporation direction and the normal to the substrate surface. Specifically, the process of correcting the instantaneous fluctuation rate of precipitation is as follows: Obtain the baseline evaporation distance and the effective evaporation distance at the collection time point, calculate the absolute difference between the effective evaporation distance and the baseline evaporation distance, and calculate the evaporation distance fluctuation coefficient by comparing the result of the absolute difference calculation with the baseline evaporation distance. The instantaneous fluctuation rate of precipitation is calculated by multiplying the fluctuation coefficient of evaporation distance, and the result of the product calculation is added to the instantaneous fluctuation rate of precipitation to obtain the corrected instantaneous fluctuation rate of precipitation. The corrected instantaneous fluctuation rate of precipitation is labeled as a precipitation harmonic characteristic; Second-level feature analysis: Obtain the effective evaporation distance at all collection time points, and calculate the evaporation pit depth at the collection time point by comparing the effective evaporation distance at the collection time point with the reference evaporation distance. The depth of the evaporation pit at adjacent acquisition time points is obtained and the difference is calculated to obtain the resonance pit depth characteristic; It should be noted that the physical significance of obtaining precipitation harmonic variation characteristics and resonance pit depth characteristics is as follows: by characterizing the fluctuation state of the precipitation rate and the real-time change of the evaporation pit depth during the coating process, the specific situation of the evaporation process can be monitored; precipitation harmonic variation characteristics reflect the degree of fluctuation of the film material precipitation rate caused by factors such as the change of evaporation distance, which helps to assess the stability of the precipitation process; resonance pit depth characteristics directly reflect the changing trend of the evaporation pit over time, which can be used to infer the consumption uniformity and stability of the evaporation surface. The process of constructing an evaporation determination matrix based on precipitation harmonic characteristics and resonance pit depth characteristics to identify whether evaporative respiration has occurred is as follows: Obtain complete coating process data without evaporation breathing phenomenon from multiple sets of historical evaporation databases, and extract the normal precipitation harmonic characteristics and normal resonance pit depth characteristics at each acquisition time point during the coating process; The normal precipitation harmonic variation characteristics and normal resonance pit depth characteristics of each collection time point in each group are integrated to obtain the normal precipitation harmonic variation dataset and the normal resonance pit depth dataset. The normal precipitation harmonic dataset and the normal resonance pit depth dataset are respectively averaged to obtain the precipitation harmonic threshold and the resonance pit depth threshold. Specifically, the method for constructing the evaporation determination matrix is ​​as follows: The precipitation harmonic variation characteristics and the resonance pit depth characteristics are compared with the precipitation harmonic variation threshold and the resonance pit depth threshold, respectively. Condition 1: If the precipitation harmonic characteristic is less than or equal to the precipitation harmonic threshold and the resonance pit depth characteristic is less than or equal to the resonance pit depth threshold, then it is determined that no evaporative respiration has occurred. Condition 2: If the precipitation harmonic characteristic is less than or equal to the precipitation harmonic threshold and the resonance pit depth characteristic is greater than the resonance pit depth threshold, then it is judged as a suspected occurrence of evaporative respiration. Condition 3: If the precipitation harmonic variation characteristic is greater than the precipitation harmonic variation threshold and the resonance pit depth characteristic is less than or equal to the resonance pit depth threshold, then it is judged as a suspected occurrence of evaporative respiration. Condition 4: If the precipitation harmonic variation characteristic is greater than the precipitation harmonic variation threshold and the resonance pit depth characteristic is greater than the resonance pit depth threshold, then it is determined that evaporative respiration has occurred. The technical solution of this embodiment is as follows: Process parameter data and vapor deposition dynamic data are collected during the coating process of medical endoscopes, and the process parameter data and vapor deposition dynamic data are integrated into a vapor deposition parameter dataset; this provides comprehensive data support for tracing the causes of evaporative breathing in medical endoscope coatings, adjusting process parameters, and improving coating stability; based on the vapor deposition parameter dataset, a two-level feature analysis is performed to obtain precipitation harmonic characteristics and resonance pit depth characteristics; based on the precipitation harmonic characteristics and resonance pit depth characteristics, an evaporation judgment matrix is ​​constructed to identify whether evaporative breathing has occurred; through two-level feature extraction and judgment matrix construction, rapid identification of evaporative breathing is achieved, providing targeted basis for optimizing the medical endoscope coating process and suppressing the phenomenon. Example 2

[0022] Please see Figure 1 As shown, an adaptive compensation system for optical lens coating includes: Predictive Analysis Module: If evaporative respiration occurs, predictive analysis will be performed to identify the start time of the development of evaporative respiration into a multi-pitted state. After determining that evaporative respiration has occurred, the effective evaporation distance of the time points collected within the sliding time window is extracted. Preferably, the length of the sliding time window is 50ms, and the sliding step of the sliding time window is 10ms; The depth of the evaporation pit at each collection time point is calculated by comparing the effective evaporation distance with the reference evaporation distance. The difference between the evaporation pit depths at adjacent acquisition time points is calculated to obtain the resonance pit depth feature, and the resonance pit depth feature of each acquisition time point is integrated into a resonance pit depth sequence. For example, the resonant pit depth sequence is composed of data within a sliding time window, and each sliding time window corresponds to a resonant pit depth sequence; The coefficient of variation is calculated from the resonant pit depth sequence, and the coefficient of variation of the resonant pit depth sequence is denoted as the pit depth coefficient of variation. It should be noted that the coefficient of variation of pit depth reflects the uniformity and stability of the change in the depth of evaporation pit over time. After determining that evaporative respiration has occurred, obtain the electron beam bombardment coordinates of the acquisition time points within the sliding time window; Obtain the electron beam bombardment coordinates at all acquisition time points, extract the X-axis coordinates of all electron beam bombardment coordinates and calculate the variance, labeled as Var(X), and extract the Y-axis coordinates of all electron beam bombardment coordinates and calculate the variance, labeled as Var(Y). The bombardment dispersion is obtained by summing Var(X) and Var(Y); It should be noted that the physical meaning of bombardment dispersion is to reflect the spatial concentration of the positions of electron beam bombardment on the surface of the film material; the smaller the bombardment dispersion, the more the electron beam bombardment position is continuously focused on the central area of ​​the film material; the larger the bombardment dispersion, the more the bombardment position diffuses towards the edge of the film material or dispersed bombardment points appear. The steady-state development index is obtained by multiplying the pit depth variation coefficient and the bombardment dispersion after standard normalization. It should be noted that the purpose of obtaining the steady-state development index is to comprehensively assess and predict the development trend of pits after the evaporation breathing phenomenon during the vapor deposition process by integrating the evaporation pit depth fluctuation index and the electron beam bombardment spatial dispersion index. Specifically, the pit depth variation coefficient reflects the uniformity and stability of the evaporation pit depth change over time, while the bombardment dispersion reflects the spatial concentration of the electron beam bombardment position on the film surface. The steady-state development index obtained by multiplying the two after standard normalization can quantify the potential risk of the vapor deposition process evolving from a single pit state to a multi-pit state. The steady-state development index within the sliding time window is obtained according to the time sequence, and the steady-state development index obtained each time is compared with the steady-state development index threshold. If the steady-state development index is greater than the steady-state development index threshold, it indicates that the evaporation pit is developing into a multi-pit state and is marked as an actual multi-pit event. It should be noted that the first sliding time window in which the development steady-state index is greater than the development steady-state index threshold is judged as the starting window for the development of evaporation pits into a multi-pit state, and the first collection time point within the starting window is marked as the starting time. Window judgment module: Based on the determined start time, distance measurement analysis is performed to obtain the actual formation time window of multiple pit states, and the actual formation time window is used to determine whether the time required for system optimization is met; Based on the historical evaporation database, all marked historical multi-pit events were obtained, and feature extraction was performed on all historical multi-pit events, as follows: The historical precipitation harmonic variation features and historical resonance pit depth features within the corresponding starting window of each historical multi-pit event are obtained and integrated to obtain the historical precipitation harmonic variation sequence and the historical resonance pit depth sequence. Obtain the start time and complete formation time of each historical multi-pit event, and calculate the formation time window of each historical multi-pit event by the difference between the complete formation time and the start time. For example, the complete formation time is from the start time of the multi-pit state until the surface of the film material shows a stable central pit and multiple peripheral secondary pits, and the core indicators of pit depth variation coefficient and bombardment dispersion no longer fluctuate significantly. The precipitation harmonic variation characteristics and resonance pit depth characteristics within the initial window of the actual multi-pit event are obtained and integrated into precipitation harmonic variation sequence and resonance pit depth sequence. Calculate the Manhattan distance between the actual multi-pit event and each historical multi-pit event; Understandably, the Manhattan distance is calculated as follows: the absolute difference between the historical sedimentation harmonic sequence and the historical resonance pit depth sequence and the sedimentation harmonic sequence and the resonance pit depth sequence at the corresponding acquisition time points is calculated one by one, and then all the absolute difference calculation results are accumulated to obtain the Manhattan distance. Obtain all Manhattan distances and sort them in descending order. Extract the formation time window of the historical multi-pit event corresponding to the minimum Manhattan distance, and use it as the actual formation time window of the actual multi-pit event. It should be noted that the physical meaning of Manhattan distance is a similarity measurement tool for the evolutionary trajectory of multidimensional features. It essentially captures the synchronicity of changes in the two dimensions of precipitation harmonic characteristics and resonance pit depth characteristics between historical multi-pit events and actual events by accumulating the absolute differences point by point. If the two events show similar changes and trends at each time point of the feature sequence, the Manhattan distance is small, indicating that the two have highly similar evolutionary patterns. Conversely, if there is a significant time lag or difference in feature amplitude, the accumulated value will increase, and the similarity between the two is poor. In some embodiments, the optimal control duration is compared with the actual formation time window; It should be noted that the optimal control duration is calculated by statistically analyzing all optimal control events that successfully controlled multiple pitting events into single pitting events within the formation time window in the past L months, and taking the average time of all optimal control events as the optimal control duration. Preferably, L is 6; If the optimization control time is greater than or equal to the actual formation time window, it means that the formation time does not meet the time required for optimization. If the optimization control time is less than the actual formation time window, it means that the formation time meets the time required for optimization. Hierarchical optimization module: If the time required for system optimization is met, a progressive parameter optimization strategy is adopted for optimization; if the time required for system optimization is not met, the coating process must be stopped immediately. If the system optimization time is satisfied, the optimization process using an incremental parameter optimization strategy is as follows: If the actual formation time window meets the system optimization time, it is marked as an autonomous tolerance scenario; if the actual formation time window does not meet the system optimization time, it is marked as an emergency compensation scenario. Optimization decisions were formulated for autonomous tolerance scenarios and emergency compensation scenarios respectively; Autonomous tolerance scenario: Obtain the current precipitation harmonic variation characteristics and precipitation harmonic variation threshold, calculate the difference between the current precipitation harmonic variation characteristics and precipitation harmonic variation threshold, and calculate the ratio between the difference calculation result and the precipitation harmonic variation threshold to obtain the precipitation correction coefficient; Obtain the current electron beam power, multiply the current electron beam power with the precipitation correction coefficient to calculate the power correction value, and then perform a difference processing between the current electron beam power and the power correction value to obtain the target electron beam power. The target electron beam power is input into the industrial automatic control system device to adjust the electron beam power; It should be noted that the relationship between electron beam power and precipitation rate is that electron beam power directly affects the evaporation rate of the film material: the higher the power, the greater the energy transferred by the electron beam to the film material, the faster the film material temperature rises, the faster the evaporation (sublimation) rate, and therefore the faster the precipitation rate. Obtain the current bombardment dispersion and the reference dispersion, and calculate the dispersion compensation amount by calculating the difference between the current bombardment dispersion and the reference dispersion; Obtain the current electron beam bombardment coordinates, calibrate the electron beam bombardment coordinates based on discrete compensation, improve the spatial concentration of electron beam bombardment, and reduce the probability of edge secondary pit formation. The process where the coating process must be stopped immediately if the system optimization time is not met is as follows: Emergency compensation scenarios: In the case of an emergency compensation scenario, the medical endoscope coating process should be stopped immediately. It should be noted that when the evaporative breathing phenomenon evolves into a multi-pitted morphology, the coating process must be stopped immediately. This is because multi-pitted morphology can cause precipitation rate fluctuations that exceed the threshold required for endoscope film uniformity, directly resulting in thickness deviations in the center and edge regions, and the product will fail quality inspection. At this point, process optimization measures can no longer compensate for the nonlinear geometric distance fluctuations and the complexity of the spatial precipitation field. The closed-loop control system also has difficulty distinguishing the evaporation contribution of multi-pitted morphology. Continuing to coat the film will lead to the scrapping of the entire batch of products, increasing cost losses, and may cause medical risks due to imaging quality defects. The technical solution of this embodiment is as follows: If evaporative breathing occurs, predictive analysis is performed to identify the start time of the evaporative breathing phenomenon developing into a multi-pitted state; the key nodes in the evolution of evaporative breathing from the evaporative breathing phenomenon to the multi-pitted state are locked to provide a time window for subsequent intervention and optimization; based on the determined start time, distance measurement analysis is performed to obtain the actual formation time window of the multi-pitted state, and it is determined whether the time required for system optimization is met based on the actual formation time window; the time feasibility of system optimization is verified to prepare for the timely initiation of targeted suppression strategies; if the time required for system optimization is met, a progressive parameter optimization strategy is adopted for optimization; if the time required for system optimization is not met, the coating process must be stopped immediately; through differentiated treatment based on time adaptability, the risk of multi-pitted defects expanding is prevented, and the optical performance and production rationality of medical endoscope coating are improved. Example 3

[0023] Please see Figure 3 As shown, an adaptive compensation method for optical lens coating includes: Step 1: Collect process parameter data and vapor deposition dynamic data during the coating process of medical endoscopes, and integrate the process parameter data and vapor deposition dynamic data into a vapor deposition parameter dataset; Step 2: Perform two-level feature analysis based on the evaporation parameter dataset to obtain precipitation harmonic characteristics and resonance pit depth characteristics. Construct an evaporation judgment matrix based on precipitation harmonic characteristics and resonance pit depth characteristics to identify whether evaporative breathing has occurred. Step 3: If evaporative respiration occurs, predict and analyze the evaporative respiration phenomenon to identify the starting time of its development into a multi-pitted state. Step 4: Based on the determined start time, perform distance metric analysis to obtain the actual formation time window of the multiple pit states, and determine whether the time required for system optimization is met based on the actual formation time window; Step 5: If the time required for system optimization is met, then an incremental parameter optimization strategy is adopted for optimization; if the time required for system optimization is not met, then the coating process must be stopped immediately.

[0024] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. An adaptive compensation system for optical lens coating, characterized in that: Includes the following modules: Data acquisition module: Collects process parameter data and vapor deposition dynamic data during the coating process of medical endoscopes, and integrates the process parameter data and vapor deposition dynamic data into a vapor deposition parameter dataset; Phenomenon recognition module: Based on the evaporation parameter dataset, a two-level feature analysis is performed to obtain precipitation harmonic characteristics and resonance pit depth characteristics. Based on the precipitation harmonic characteristics and resonance pit depth characteristics, an evaporation judgment matrix is ​​constructed to identify whether evaporative breathing phenomenon has occurred. Predictive Analysis Module: If evaporative respiration occurs, predictive analysis will be performed to identify the start time of the development of evaporative respiration into a multi-pitted state. Window judgment module: Based on the determined start time, distance measurement analysis is performed to obtain the actual formation time window of multiple pit states, and the actual formation time window is used to determine whether the time required for system optimization is met; Hierarchical optimization module: If the time required for system optimization is met, a progressive parameter optimization strategy is adopted for optimization; if the time required for system optimization is not met, the coating process must be stopped immediately.

2. The adaptive compensation system for optical lens coating according to claim 1, characterized in that: The process of performing the aforementioned two-level feature analysis is as follows: First-level feature analysis: The initial sedimentation rate is obtained, the real-time sedimentation rate at all collection time points is obtained, the absolute difference between the real-time sedimentation rates at adjacent collection time points is calculated to obtain the sedimentation rate difference, and the ratio of the sedimentation rate difference to the initial sedimentation rate is calculated to obtain the instantaneous sedimentation fluctuation rate at that collection time point. Based on the cosine precipitation law, the instantaneous fluctuation rate of precipitation is corrected to obtain the precipitation harmonic variation characteristics; The resonance pit depth characteristics were obtained by performing a second-level feature analysis on the evaporation parameter dataset.

3. The adaptive compensation system for optical lens coating according to claim 2, characterized in that: The cosine precipitation law corrects the instantaneous fluctuation rate of precipitation in the following way: Obtain the baseline evaporation distance and the effective evaporation distance at the collection time point, calculate the absolute difference between the effective evaporation distance and the baseline evaporation distance, and calculate the evaporation distance fluctuation coefficient by comparing the result of the absolute difference calculation with the baseline evaporation distance. The instantaneous fluctuation rate of precipitation is calculated by multiplying the fluctuation coefficient of evaporation distance, and the result of the product calculation is added to the instantaneous fluctuation rate of precipitation to obtain the corrected instantaneous fluctuation rate of precipitation. The corrected instantaneous fluctuation rate of precipitation is labeled as a precipitation harmonic characteristic.

4. The adaptive compensation system for optical lens coating according to claim 2, characterized in that: The process of performing the second-level feature analysis is as follows: Obtain the effective evaporation distance at all collection time points, and calculate the evaporation pit depth at the collection time point by comparing the effective evaporation distance at the collection time point with the reference evaporation distance. The depth of the evaporation pit at adjacent acquisition time points is obtained and the difference is calculated to obtain the characteristic of the resonance pit depth.

5. The adaptive compensation system for optical lens coating according to claim 1, characterized in that: The process of performing the predictive analysis is as follows: Obtain the baseline evaporation distance, obtain the effective evaporation distance at the time points within the sliding time window, and perform pit depth variation analysis based on the baseline evaporation distance and the effective evaporation distance to obtain the pit depth variation coefficient. The electron beam bombardment coordinates at the acquisition time points within the sliding time window are obtained and the bombardment dispersion is obtained by performing dispersion analysis. The steady-state development index is obtained by multiplying the crater depth variation coefficient and the bombardment dispersion after standard normalization.

6. The adaptive compensation system for optical lens coating according to claim 5, characterized in that: The process of performing the pit depth variation analysis is as follows: The depth of the evaporation pit at each collection time point is calculated by comparing the effective evaporation distance with the reference evaporation distance. The difference between the evaporation pit depths at adjacent acquisition time points is calculated to obtain the resonance pit depth feature, and the resonance pit depth feature of each acquisition time point is integrated into a resonance pit depth sequence. The coefficient of variation is calculated from the resonant pit depth sequence, and the coefficient of variation of the resonant pit depth sequence is denoted as the pit depth coefficient of variation.

7. The adaptive compensation system for optical lens coating according to claim 5, characterized in that: The process of performing the aforementioned dispersion analysis is as follows: Obtain the electron beam bombardment coordinates at the acquisition time points, extract the X-axis coordinates of all electron beam bombardment coordinates and calculate the variance, labeled as Var(X), and extract the Y-axis coordinates of all electron beam bombardment coordinates and calculate the variance, labeled as Var(Y). The bombardment dispersion is obtained by summing Var(X) and Var(Y).

8. The adaptive compensation system for optical lens coating according to claim 1, characterized in that: The process of performing the distance metric analysis is as follows: Obtain the historical sedimentation harmonic sequence and historical resonance pit depth sequence corresponding to the starting window in each historical multi-pit event; Obtain the formation time window for each historical multi-pit event; Obtain the precipitation harmonic sequence and resonance pit depth sequence corresponding to the starting window in actual multi-pit events; Based on the historical precipitation harmonic sequence and historical resonance pit depth sequence of each historical multi-pit event, as well as the precipitation harmonic sequence and resonance pit depth sequence of the actual multi-pit event, a multi-dimensional evolution analysis is performed to obtain the actual formation time window of the actual multi-pit event.

9. An adaptive compensation system for optical lens coating according to claim 8, characterized in that: The process of performing the aforementioned multidimensional evolutionary analysis is as follows: Calculate the Manhattan distance between the actual multi-pit event and each historical multi-pit event; Obtain all Manhattan distances and sort them in descending order. Extract the formation time window of the historical multi-pit event corresponding to the smallest Manhattan distance, and use it as the actual formation time window of the actual multi-pit event.

10. An adaptive compensation method for optical lens coating according to claim 1, used to implement the adaptive compensation system for optical lens coating according to any one of claims 1-9, characterized in that: Includes the following steps: Step 1: Collect process parameter data and vapor deposition dynamic data during the coating process of medical endoscopes, and integrate the process parameter data and vapor deposition dynamic data into a vapor deposition parameter dataset; Step 2: Perform two-level feature analysis based on the evaporation parameter dataset to obtain precipitation harmonic characteristics and resonance pit depth characteristics. Construct an evaporation judgment matrix based on precipitation harmonic characteristics and resonance pit depth characteristics to identify whether evaporative breathing has occurred. Step 3: If evaporative respiration occurs, predict and analyze the evaporative respiration phenomenon to identify the starting time of its development into a multi-pitted state. Step 4: Based on the determined start time, perform distance metric analysis to obtain the actual formation time window of the multiple pit states, and determine whether the time required for system optimization is met based on the actual formation time window; Step 5: If the time required for system optimization is met, then an incremental parameter optimization strategy is adopted for optimization; if the time required for system optimization is not met, then the coating process must be stopped immediately.

Images