Search

CN-121998953-A - Method and system for detecting real-time quality defects of OCA optical adhesive

CN121998953ACN 121998953 ACN121998953 ACN 121998953ACN-121998953-A

Abstract

The invention relates to the technical field of optical material quality detection, and discloses a real-time quality defect detection method and system for OCA optical cement. The method comprises the steps of obtaining infrared thermal imaging data and colloid thickness data of an optical adhesive sample, establishing an original thermal imaging data set, solving temperature values of points in the colloid according to the original thermal imaging data set to obtain temperature gradient field data, drawing to obtain a temperature distribution model in the colloid, calculating to obtain a temperature average change rate reflecting heat conduction characteristics of each area, screening candidate defect areas, constructing multidimensional feature vectors, normalizing to generate standard defect feature vectors, carrying out similarity matching, extracting abnormal bubble areas, carrying out refinement classification on the abnormal bubble areas by adopting a decision tree classifier to generate detailed defect classification results, and calculating geometric parameters and severity ratings of each defect area to obtain an optical colloid quality detection report. The method improves the reliability of the detection result of the internal defects of the optical cement.

Inventors

  • Bo Jiakai
  • SUN HUIFANG

Assignees

  • 深圳市领跃信科技有限公司

Dates

Publication Date
20260508
Application Date
20260128

Claims (10)

  1. 1. The real-time quality defect detection method of the OCA optical adhesive is characterized by comprising the following steps of: acquiring infrared thermal imaging data and colloid thickness data of an optical adhesive sample, and establishing three-dimensional data to obtain an original thermal imaging data set containing complete space information; solving the temperature values of each point in the colloid according to the original thermal imaging data set to obtain temperature gradient field data, and drawing according to the temperature gradient field data to obtain a temperature distribution model in the colloid; Calculating the temperature value in the temperature distribution model to obtain the average change rate of the temperature reflecting the heat conduction characteristics of each region, and screening the region with the average change rate of the temperature exceeding the preset change rate threshold as the candidate defect region; For each position point in the candidate defect area, extracting temperature gradient characteristics, conversion rate characteristics and space distribution characteristics in the surrounding vicinity of the position point to construct a multidimensional characteristic vector, and normalizing the multidimensional characteristic vector by combining the colloid thickness data to generate a standard defect characteristic vector; Performing similarity matching according to the standard defect feature vector and a pre-established defect feature library to obtain a preliminary defect classification result; Extracting an abnormal bubble region according to the preliminary defect classification result, and carrying out refinement classification on the abnormal bubble region by adopting a decision tree classifier to generate a detailed defect classification result; And performing spatial clustering analysis on the detailed defect classification result to obtain a defect adjacent relation data set, combining a plurality of single defects continuously distributed in the defect adjacent relation data set into a composite defect region, calculating the geometric parameter and severity rating of each defect region, and obtaining an optical colloid quality detection report.
  2. 2. The method for detecting real-time quality defects of OCA optical cement according to claim 1, wherein solving the temperature values of each point inside the colloid according to the original thermal imaging data set to obtain temperature gradient field data, and drawing a temperature distribution model inside the colloid according to the temperature gradient field data, comprises: according to the original thermal imaging data set, performing spatial interpolation operation on the temperature distribution of the surface and the interior of the colloid by adopting cubic spline interpolation to obtain an initial temperature field distribution matrix covering the whole colloid area; Performing numerical discretization processing on a three-dimensional heat conduction partial differential equation by adopting a finite difference method according to the initial temperature field distribution matrix to obtain a temperature field numerical solution meeting a heat balance condition; performing space differential operation on the temperature field numerical solution through a gradient operator to obtain temperature gradient field data reflecting the internal heat flow distribution characteristics of the colloid; and drawing a temperature distribution model in the colloid by adopting linear interpolation according to the temperature gradient field data, wherein the temperature distribution model comprises an isothermal surface distribution map.
  3. 3. The method for detecting real-time quality defects of OCA optical cement according to claim 1, wherein calculating the temperature value in the temperature distribution model to obtain a temperature average change rate reflecting the thermal conductivity characteristics of each region, and screening the region with the temperature average change rate exceeding a preset change rate threshold as a candidate defect region comprises: sampling according to the temperature distribution model to obtain a continuous time sequence, and carrying out sectional extraction on temperature values in the continuous time sequence by adopting a sliding window to obtain a standardized time sequence temperature matrix; Carrying out point-by-point subtraction calculation on the temperature values of adjacent time points in the time sequence temperature matrix through differential calculation to obtain a spatial distribution record table containing abnormal temperature difference positions; Smoothing the temperature difference change sequence in each space region in the space distribution record table by adopting a moving average filter, and calculating the average change rate of the temperature in each space region in unit time; And comparing the temperature average change rate with a preset change rate threshold value, and determining the region with the temperature average change rate larger than the preset change rate threshold value as the candidate defect region.
  4. 4. The method for detecting real-time quality defects of OCA optical cement according to claim 1, wherein for each position point in the candidate defect region, extracting a temperature gradient feature, a conversion rate feature and a spatial distribution feature in the surrounding vicinity thereof to construct a multidimensional feature vector, normalizing the multidimensional feature vector in combination with the colloid thickness data, and generating a standard defect feature vector, comprising: Establishing a cube neighborhood window at each position point in the candidate defect area by adopting a three-dimensional domain searching algorithm to obtain a temperature distribution data set containing a neighborhood space structure; According to the temperature distribution data set, calculating the temperature variation amplitude in each direction through a gradient operator to obtain temperature gradient component values in the transverse direction, the longitudinal direction and the depth direction; constructing and obtaining an initial multidimensional feature vector containing gradient features, velocity features and position features according to the temperature gradient component values, the temperature average change rate and the space coordinate information of each position point; Determining a thickness proportionality coefficient by combining the colloid thickness data, and carrying out weight adjustment on the characteristic components in the initial multidimensional characteristic vector according to the thickness proportionality coefficient to obtain an adjusted multidimensional characteristic vector; carrying out normalization mapping according to the adjusted multidimensional feature vector to obtain a normalized feature value; And vector splicing operation is carried out on the normalized feature values, so that standard defect feature vectors combined according to a preset sequence are obtained.
  5. 5. The method for detecting real-time quality defects of OCA optical cement according to claim 1, wherein performing similarity matching according to the standard defect feature vector and a pre-established defect feature library to obtain a preliminary defect classification result comprises: Extracting according to the standard defect feature vector to obtain feature components of each dimension; Calculating Euclidean distance values between the feature components of each dimension and standard feature vectors in a pre-established defect feature library; comparing the Euclidean distance value with a preset distance threshold, and marking the feature vector corresponding to the Euclidean distance value as a candidate matching item if the Euclidean distance value is smaller than the preset distance threshold; according to the candidate matching item, calculating an included angle cosine value of the feature vector by adopting cosine similarity, and carrying out comprehensive scoring calculation according to a weight coefficient of the included angle cosine value in a defect feature library to obtain a final matching degree value; The optimal matching result with the highest matching degree is obtained after descending order arrangement is carried out according to the final matching degree value; And converting the matching numerical value into a confidence score through a confidence coefficient conversion function according to the optimal matching result, extracting a corresponding defect type label from a defect feature library, and determining a preliminary defect classification result.
  6. 6. The method for detecting real-time quality defects of OCA optical cement according to claim 1, wherein extracting an abnormal bubble region according to the preliminary defect classification result, refining and classifying the abnormal bubble region by using a decision tree classifier, and generating a detailed defect classification result, comprises: When the preliminary defect classification result comprises an internal bubble defect, carrying out three-dimensional positioning on the bubble position to obtain the bubble position; when the bubble position exceeds a preset central area, marking the position as an abnormal bubble point, and performing thickness sampling according to the bubble position to obtain colloid thickness distribution, and combining to obtain a first data set containing the bubble position and the thickness distribution; Temperature sampling is carried out according to the abnormal bubble point to obtain a temperature difference value of the abnormal bubble point, if the temperature difference value exceeds a preset temperature fluctuation threshold value, an abnormal temperature region is marked, and a corresponding relation between temperature difference abnormality and structural defect is determined in the first data set to obtain a second data set containing the abnormal temperature region; and carrying out refined classification on the second data set through a decision tree classifier to obtain a detailed defect classification result.
  7. 7. The method for detecting real-time quality defects of OCA optical cement according to claim 1, wherein performing spatial cluster analysis on the detailed defect classification result to obtain a defect proximity relation data set, merging a plurality of single defects continuously distributed in the defect proximity relation data set into composite defect regions, calculating geometric parameters and severity ratings of each defect region, and obtaining an optical cement quality detection report, comprising: Measuring the space adjacent distance between each defect point according to the detailed defect classification result to obtain the distance value between each adjacent defect; If the distance value between two defect points is smaller than a preset proximity threshold value, marking the two defect points as associated defect pairs, and counting all the associated defect pairs to obtain a defect proximity relation data set; Carrying out distribution continuity judgment according to the associated defect pairs in the defect adjacent relation data set, and merging a plurality of continuously distributed associated defect pairs into a unified composite defect area if three or more associated defect pairs form a continuous path in space; Extracting geometric parameters of the composite defect region, wherein the geometric parameters comprise normalized defect region area, shape complexity and distribution density; Calculating comprehensive scores according to preset weights and the geometric parameters, and grading according to the comprehensive scores to obtain severity ratings; and outputting the defect position, the geometric parameter and the severity rating of the composite defect area as an optical colloid quality detection report.
  8. 8. The real-time quality defect detection system of OCA optical cement is characterized by comprising: The data acquisition module is used for acquiring infrared thermal imaging data and colloid thickness data of the optical adhesive sample, and establishing three-dimensional data to obtain an original thermal imaging data set containing complete space information; The temperature analysis module is used for solving the temperature values of each point in the colloid according to the original thermal imaging data set to obtain temperature gradient field data, and drawing a temperature distribution model in the colloid according to the temperature gradient field data; The region screening module is used for calculating the temperature values in the temperature distribution model to obtain the average change rate of the temperature reflecting the heat conduction characteristics of each region, and screening the regions with the average change rate of the temperature exceeding the preset change rate threshold as candidate defect regions; The feature extraction module is used for extracting temperature gradient features, conversion rate features and space distribution features in the surrounding vicinity of each position point in the candidate defect region to construct a multidimensional feature vector, and normalizing the multidimensional feature vector by combining the colloid thickness data to generate a standard defect feature vector; the defect matching module is used for carrying out similarity matching according to the standard defect feature vector and a pre-established defect feature library to obtain a preliminary defect classification result; The refinement classification module is used for extracting an abnormal bubble region according to the preliminary defect classification result, and performing refinement classification on the abnormal bubble region by adopting a decision tree classifier to generate a detailed defect classification result; And the report generation module is used for carrying out spatial cluster analysis on the detailed defect classification result to obtain a defect adjacent relation data set, combining a plurality of single defects continuously distributed in the defect adjacent relation data set into a composite defect area, and calculating the geometric parameter and severity rating of each defect area to obtain an optical colloid quality detection report.
  9. 9. An electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, the processor implementing the method for detecting real-time quality defects of OCA optical cement according to any one of claims 1 to 7 when the computer program is executed by the processor.
  10. 10. A computer readable storage medium, characterized in that the computer readable storage medium comprises a stored computer program, wherein the computer program, when run, controls a device in which the computer readable storage medium is located to perform the method for detecting real-time quality defects of OCA optical cement according to any one of claims 1 to 7.

Description

Method and system for detecting real-time quality defects of OCA optical adhesive Technical Field The invention relates to the technical field of optical material quality detection, in particular to a real-time quality defect detection method and system for OCA optical cement. Background Optical Clear Adhesive (OCA) is used as a key material in display technology, and is widely applied to touch screen and liquid crystal display manufacturing, and the quality of the optical clear adhesive directly influences the optical performance and user experience of products. With the continuous improvement of the requirements of consumer electronics on display quality, the quality of the optical adhesive directly influences the optical performance and user experience of the final product, so that the quality detection becomes a non-negligible link in the manufacturing process. In one prior art, conventional inspection methods such as optical microscopy and machine vision techniques are directed to inspecting surface defects. However, for microscopic defects such as micro bubbles, impurities, uneven thickness, etc. inside the colloid, the conventional optical method based on the surface morphology is difficult to directly detect. These internal defects can change the local heat conduction performance of the material, but the prior art lacks means for accurately inverting the internal temperature field from the measurable surface or the whole heat signal and establishing quantitative correspondence between the internal temperature field and the defect type and severity, so that the reliability of the detection result is low, and the quality control requirement of high-precision manufacturing cannot be met. Disclosure of Invention In order to solve the technical problems, the invention provides a real-time quality defect detection method and a real-time quality defect detection system for OCA optical cement, which are used for realizing quantitative analysis on defect types and improving the reliability of detection results. In order to solve the above technical problems, the present invention provides a method for detecting real-time quality defects of OCA optical cement, comprising: acquiring infrared thermal imaging data and colloid thickness data of an optical adhesive sample, and establishing three-dimensional data to obtain an original thermal imaging data set containing complete space information; solving the temperature values of each point in the colloid according to the original thermal imaging data set to obtain temperature gradient field data, and drawing according to the temperature gradient field data to obtain a temperature distribution model in the colloid; Calculating the temperature value in the temperature distribution model to obtain the average change rate of the temperature reflecting the heat conduction characteristics of each region, and screening the region with the average change rate of the temperature exceeding the preset change rate threshold as the candidate defect region; For each position point in the candidate defect area, extracting temperature gradient characteristics, conversion rate characteristics and space distribution characteristics in the surrounding vicinity of the position point to construct a multidimensional characteristic vector, and normalizing the multidimensional characteristic vector by combining the colloid thickness data to generate a standard defect characteristic vector; Performing similarity matching according to the standard defect feature vector and a pre-established defect feature library to obtain a preliminary defect classification result; Extracting an abnormal bubble region according to the preliminary defect classification result, and carrying out refinement classification on the abnormal bubble region by adopting a decision tree classifier to generate a detailed defect classification result; And performing spatial clustering analysis on the detailed defect classification result to obtain a defect adjacent relation data set, combining a plurality of single defects continuously distributed in the defect adjacent relation data set into a composite defect region, calculating the geometric parameter and severity rating of each defect region, and obtaining an optical colloid quality detection report. In an alternative embodiment, the acquiring the infrared thermal imaging data and the colloid thickness data of the optical cement sample, and performing three-dimensional data establishment to obtain an original thermal imaging data set containing complete spatial information, includes: Obtaining radiation values of each point of the colloid by carrying out line-by-line infrared scanning on the surface of the optical cement sample, and recording coordinates of the surface of the colloid to obtain infrared thermal imaging data; according to the radiation value, the radiation value is converted through a Spanish-Boltzmann law to obtain a temperature value of each point